diff --git a/substitutivity_original_Estonian/Qwen3.5-2B-Base_substitutivity_splits_original_features_train_substitutivity_splits_original_features_test1/README.md b/substitutivity_original_Estonian/Qwen3.5-2B-Base_substitutivity_splits_original_features_train_substitutivity_splits_original_features_test1/README.md deleted file mode 100644 index b2f25c5f0135fbd643695ff8de834c523dc703c7..0000000000000000000000000000000000000000 --- a/substitutivity_original_Estonian/Qwen3.5-2B-Base_substitutivity_splits_original_features_train_substitutivity_splits_original_features_test1/README.md +++ /dev/null @@ -1,58 +0,0 @@ ---- -base_model: Qwen/Qwen3.5-2B-Base -library_name: transformers -model_name: Qwen3.5-2B-Base_substitutivity_splits_original_features_train_substitutivity_splits_original_features_test1 -tags: -- generated_from_trainer -- sft -- trl -licence: license ---- - -# Model Card for Qwen3.5-2B-Base_substitutivity_splits_original_features_train_substitutivity_splits_original_features_test1 - -This model is a fine-tuned version of [Qwen/Qwen3.5-2B-Base](https://huggingface.co/Qwen/Qwen3.5-2B-Base). -It has been trained using [TRL](https://github.com/huggingface/trl). - -## Quick start - -```python -from transformers import pipeline - -question = "If you had a time machine, but could only go to the past or the future once and never return, which would you choose and why?" -generator = pipeline("text-generation", model="None", device="cuda") -output = generator([{"role": "user", "content": question}], max_new_tokens=128, return_full_text=False)[0] -print(output["generated_text"]) -``` - -## Training procedure - -[Visualize in Weights & Biases](https://wandb.ai/katriin-kukk/Cross_lingual_morphological_generalization/runs/x2lpl94k) - - - -This model was trained with SFT. - -### Framework versions - -- TRL: 0.29.0 -- Transformers: 5.5.4 -- Pytorch: 2.10.0 -- Datasets: 4.6.1 -- Tokenizers: 0.22.2 - -## Citations - - - -Cite TRL as: - -```bibtex -@software{vonwerra2020trl, - title = {{TRL: Transformers Reinforcement Learning}}, - author = {von Werra, Leandro and Belkada, Younes and Tunstall, Lewis and Beeching, Edward and Thrush, Tristan and Lambert, Nathan and Huang, Shengyi and Rasul, Kashif and Gallouédec, Quentin}, - license = {Apache-2.0}, - url = {https://github.com/huggingface/trl}, - year = {2020} -} -``` \ No newline at end of file diff --git a/substitutivity_original_Estonian/Qwen3.5-2B-Base_substitutivity_splits_original_features_train_substitutivity_splits_original_features_test1/checkpoint-1245/README.md b/substitutivity_original_Estonian/Qwen3.5-2B-Base_substitutivity_splits_original_features_train_substitutivity_splits_original_features_test1/checkpoint-1245/README.md deleted file mode 100644 index ece204d005f7bf161d0e448efdd56fd297b7c8a2..0000000000000000000000000000000000000000 --- a/substitutivity_original_Estonian/Qwen3.5-2B-Base_substitutivity_splits_original_features_train_substitutivity_splits_original_features_test1/checkpoint-1245/README.md +++ /dev/null @@ -1,209 +0,0 @@ ---- -base_model: Qwen/Qwen3.5-2B-Base -library_name: peft -pipeline_tag: text-generation -tags: -- base_model:adapter:Qwen/Qwen3.5-2B-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/substitutivity_original_Estonian/Qwen3.5-2B-Base_substitutivity_splits_original_features_train_substitutivity_splits_original_features_test1/checkpoint-1245/adapter_config.json b/substitutivity_original_Estonian/Qwen3.5-2B-Base_substitutivity_splits_original_features_train_substitutivity_splits_original_features_test1/checkpoint-1245/adapter_config.json deleted file mode 100644 index a64145ce23adcbdbe246679087a13f4cb63ff664..0000000000000000000000000000000000000000 --- a/substitutivity_original_Estonian/Qwen3.5-2B-Base_substitutivity_splits_original_features_train_substitutivity_splits_original_features_test1/checkpoint-1245/adapter_config.json +++ /dev/null @@ -1,46 +0,0 @@ -{ - "alora_invocation_tokens": null, - "alpha_pattern": {}, - "arrow_config": null, - "auto_mapping": null, - "base_model_name_or_path": "Qwen/Qwen3.5-2B-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": 256, - "lora_bias": false, - "lora_dropout": 0.027050353915701653, - "megatron_config": null, - "megatron_core": "megatron.core", - "modules_to_save": null, - "peft_type": "LORA", - "peft_version": "0.18.1", - "qalora_group_size": 16, - "r": 128, - "rank_pattern": {}, - "revision": null, - "target_modules": [ - "o_proj", - "q_proj", - "k_proj", - "down_proj", - "up_proj", - "v_proj", - "gate_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/substitutivity_original_Estonian/Qwen3.5-2B-Base_substitutivity_splits_original_features_train_substitutivity_splits_original_features_test1/checkpoint-1245/adapter_model.safetensors b/substitutivity_original_Estonian/Qwen3.5-2B-Base_substitutivity_splits_original_features_train_substitutivity_splits_original_features_test1/checkpoint-1245/adapter_model.safetensors deleted file mode 100644 index 8613c74cffaedfc126bc20926f401123708c2c11..0000000000000000000000000000000000000000 --- a/substitutivity_original_Estonian/Qwen3.5-2B-Base_substitutivity_splits_original_features_train_substitutivity_splits_original_features_test1/checkpoint-1245/adapter_model.safetensors +++ /dev/null @@ -1,3 +0,0 @@ -version https://git-lfs.github.com/spec/v1 -oid sha256:9cafaedfbb8f3542e770b5e614799acc29b354bf5a88410c70c9298fdd7e383a -size 349201608 diff --git a/substitutivity_original_Estonian/Qwen3.5-2B-Base_substitutivity_splits_original_features_train_substitutivity_splits_original_features_test1/checkpoint-1245/chat_template.jinja b/substitutivity_original_Estonian/Qwen3.5-2B-Base_substitutivity_splits_original_features_train_substitutivity_splits_original_features_test1/checkpoint-1245/chat_template.jinja deleted file mode 100644 index 0ef09f214eaa6d9bca297988afc1454b5827b2c7..0000000000000000000000000000000000000000 --- a/substitutivity_original_Estonian/Qwen3.5-2B-Base_substitutivity_splits_original_features_train_substitutivity_splits_original_features_test1/checkpoint-1245/chat_template.jinja +++ /dev/null @@ -1,154 +0,0 @@ -{%- set image_count = namespace(value=0) %} -{%- set video_count = namespace(value=0) %} -{%- macro render_content(content, do_vision_count, is_system_content=false) %} - {%- if content is string %} - {{- content }} - {%- elif content is iterable and content is not mapping %} - {%- for item in content %} - {%- if 'image' in item or 'image_url' in item or item.type == 'image' %} - {%- if is_system_content %} - {{- raise_exception('System message cannot contain images.') }} - {%- endif %} - {%- if do_vision_count %} - {%- set image_count.value = image_count.value + 1 %} - {%- endif %} - {%- if add_vision_id %} - {{- 'Picture ' ~ image_count.value ~ ': ' }} - {%- endif %} - {{- '<|vision_start|><|image_pad|><|vision_end|>' }} - {%- elif 'video' in item or item.type == 'video' %} - {%- if is_system_content %} - {{- raise_exception('System message cannot contain videos.') }} - {%- endif %} - {%- if do_vision_count %} - {%- set video_count.value = video_count.value + 1 %} - {%- endif %} - {%- if add_vision_id %} - {{- 'Video ' ~ video_count.value ~ ': ' }} - {%- endif %} - {{- '<|vision_start|><|video_pad|><|vision_end|>' }} - {%- elif 'text' in item %} - {{- item.text }} - {%- else %} - {{- raise_exception('Unexpected item type in content.') }} - {%- endif %} - {%- endfor %} - {%- elif content is none or content is undefined %} - {{- '' }} - {%- else %} - {{- raise_exception('Unexpected content type.') }} - {%- endif %} -{%- endmacro %} -{%- if not messages %} - {{- raise_exception('No messages provided.') }} -{%- endif %} -{%- if tools and tools is iterable and tools is not mapping %} - {{- '<|im_start|>system\n' }} - {{- "# Tools\n\nYou have access to the following functions:\n\n" }} - {%- for tool in tools %} - {{- "\n" }} - {{- tool | tojson }} - {%- endfor %} - {{- "\n" }} - {{- '\n\nIf you choose to call a function ONLY reply in the following format with NO suffix:\n\n\n\n\nvalue_1\n\n\nThis is the value for the second parameter\nthat can span\nmultiple lines\n\n\n\n\n\nReminder:\n- Function calls MUST follow the specified format: an inner block must be nested within XML tags\n- Required parameters MUST be specified\n- You may provide optional reasoning for your function call in natural language BEFORE the function call, but NOT after\n- If there is no function call available, answer the question like normal with your current knowledge and do not tell the user about function calls\n' }} - {%- if messages[0].role == 'system' %} - {%- set content = render_content(messages[0].content, false, true)|trim %} - {%- if content %} - {{- '\n\n' + content }} - {%- endif %} - {%- endif %} - {{- '<|im_end|>\n' }} -{%- else %} - {%- if messages[0].role == 'system' %} - {%- set content = render_content(messages[0].content, false, true)|trim %} - {{- '<|im_start|>system\n' + 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" %} - {%- set content = render_content(message.content, false)|trim %} - {%- if not(content.startswith('') and content.endswith('')) %} - {%- set ns.multi_step_tool = false %} - {%- set ns.last_query_index = index %} - {%- endif %} - {%- endif %} -{%- endfor %} -{%- if ns.multi_step_tool %} - {{- raise_exception('No user query found in messages.') }} -{%- endif %} -{%- for message in messages %} - {%- set content = render_content(message.content, true)|trim %} - {%- if message.role == "system" %} - {%- if not loop.first %} - {{- raise_exception('System message must be at the beginning.') }} - {%- endif %} - {%- elif message.role == "user" %} - {{- '<|im_start|>' + message.role + '\n' + content + '<|im_end|>' + '\n' }} - {%- elif message.role == "assistant" %} - {%- set reasoning_content = '' %} - {%- if message.reasoning_content is string %} - {%- set reasoning_content = message.reasoning_content %} - {%- else %} - {%- if '' in content %} - {%- set reasoning_content = content.split('')[0].rstrip('\n').split('')[-1].lstrip('\n') %} - {%- set content = content.split('')[-1].lstrip('\n') %} - {%- endif %} - {%- endif %} - {%- set reasoning_content = reasoning_content|trim %} - {%- if loop.index0 > ns.last_query_index %} - {{- '<|im_start|>' + message.role + '\n\n' + reasoning_content + '\n\n\n' + content }} - {%- else %} - {{- '<|im_start|>' + message.role + '\n' + content }} - {%- endif %} - {%- if message.tool_calls and message.tool_calls is iterable and message.tool_calls is not mapping %} - {%- for tool_call in message.tool_calls %} - {%- if tool_call.function is defined %} - {%- set tool_call = tool_call.function %} - {%- endif %} - {%- if loop.first %} - {%- if content|trim %} - {{- '\n\n\n\n' }} - {%- else %} - {{- '\n\n' }} - {%- endif %} - {%- else %} - {{- '\n\n\n' }} - {%- endif %} - {%- if tool_call.arguments is defined %} - {%- for args_name, args_value in tool_call.arguments|items %} - {{- '\n' }} - {%- set args_value = args_value | tojson | safe if args_value is mapping or (args_value is sequence and args_value is not string) else args_value | string %} - {{- args_value }} - {{- '\n\n' }} - {%- endfor %} - {%- endif %} - {{- '\n' }} - {%- endfor %} - {%- endif %} - {{- '<|im_end|>\n' }} - {%- elif message.role == "tool" %} - {%- if loop.previtem and loop.previtem.role != "tool" %} - {{- '<|im_start|>user' }} - {%- endif %} - {{- '\n\n' }} - {{- content }} - {{- '\n' }} - {%- if not loop.last and loop.nextitem.role != "tool" %} - {{- '<|im_end|>\n' }} - {%- elif loop.last %} - {{- '<|im_end|>\n' }} - {%- endif %} - {%- else %} - {{- raise_exception('Unexpected message role.') }} - {%- endif %} -{%- endfor %} -{%- if add_generation_prompt %} - {{- '<|im_start|>assistant\n' }} - {%- if enable_thinking is defined and enable_thinking is true %} - {{- '\n' }} - {%- else %} - {{- '\n\n\n\n' }} - {%- endif %} -{%- endif %} \ No newline at end of file diff --git a/substitutivity_original_Estonian/Qwen3.5-2B-Base_substitutivity_splits_original_features_train_substitutivity_splits_original_features_test1/checkpoint-1245/tokenizer.json b/substitutivity_original_Estonian/Qwen3.5-2B-Base_substitutivity_splits_original_features_train_substitutivity_splits_original_features_test1/checkpoint-1245/tokenizer.json deleted file mode 100644 index 67741b04f23bfdb46501f748ce27865ec82eccfb..0000000000000000000000000000000000000000 --- a/substitutivity_original_Estonian/Qwen3.5-2B-Base_substitutivity_splits_original_features_train_substitutivity_splits_original_features_test1/checkpoint-1245/tokenizer.json +++ /dev/null @@ -1,3 +0,0 @@ -version https://git-lfs.github.com/spec/v1 -oid sha256:87a7830d63fcf43bf241c3c5242e96e62dd3fdc29224ca26fed8ea333db72de4 -size 19989343 diff --git a/substitutivity_original_Estonian/Qwen3.5-2B-Base_substitutivity_splits_original_features_train_substitutivity_splits_original_features_test1/checkpoint-1245/tokenizer_config.json b/substitutivity_original_Estonian/Qwen3.5-2B-Base_substitutivity_splits_original_features_train_substitutivity_splits_original_features_test1/checkpoint-1245/tokenizer_config.json deleted file mode 100644 index b4a37b2a6fd3ab3317cd7bac72855be1a843b2bb..0000000000000000000000000000000000000000 --- a/substitutivity_original_Estonian/Qwen3.5-2B-Base_substitutivity_splits_original_features_train_substitutivity_splits_original_features_test1/checkpoint-1245/tokenizer_config.json +++ /dev/null @@ -1,31 +0,0 @@ -{ - "add_prefix_space": false, - "audio_bos_token": "<|audio_start|>", - "audio_eos_token": "<|audio_end|>", - "audio_token": "<|audio_pad|>", - "backend": "tokenizers", - "bos_token": null, - "clean_up_tokenization_spaces": false, - "eos_token": "<|endoftext|>", - "errors": "replace", - "image_token": "<|image_pad|>", - "is_local": false, - "model_max_length": 262144, - "model_specific_special_tokens": { - "audio_bos_token": "<|audio_start|>", - "audio_eos_token": "<|audio_end|>", - "audio_token": "<|audio_pad|>", - "image_token": "<|image_pad|>", - "video_token": "<|video_pad|>", - "vision_bos_token": "<|vision_start|>", - "vision_eos_token": "<|vision_end|>" - }, - "pad_token": "<|endoftext|>", - "pretokenize_regex": "(?i:'s|'t|'re|'ve|'m|'ll|'d)|[^\\r\\n\\p{L}\\p{N}]?[\\p{L}\\p{M}]+|\\p{N}| ?[^\\s\\p{L}\\p{M}\\p{N}]+[\\r\\n]*|\\s*[\\r\\n]+|\\s+(?!\\S)|\\s+", - "split_special_tokens": false, - "tokenizer_class": "TokenizersBackend", - "unk_token": null, - "video_token": "<|video_pad|>", - "vision_bos_token": "<|vision_start|>", - "vision_eos_token": "<|vision_end|>" -} diff --git a/substitutivity_original_Estonian/Qwen3.5-2B-Base_substitutivity_splits_original_features_train_substitutivity_splits_original_features_test1/checkpoint-1245/trainer_state.json b/substitutivity_original_Estonian/Qwen3.5-2B-Base_substitutivity_splits_original_features_train_substitutivity_splits_original_features_test1/checkpoint-1245/trainer_state.json deleted file mode 100644 index b8b601807279ad7ff920876a62f799a54142facf..0000000000000000000000000000000000000000 --- a/substitutivity_original_Estonian/Qwen3.5-2B-Base_substitutivity_splits_original_features_train_substitutivity_splits_original_features_test1/checkpoint-1245/trainer_state.json +++ /dev/null @@ -1,340 +0,0 @@ -{ - "best_global_step": null, - "best_metric": null, - "best_model_checkpoint": null, - "epoch": 3.0, - "eval_steps": 500, - "global_step": 1245, - "is_hyper_param_search": false, - "is_local_process_zero": true, - "is_world_process_zero": true, - "log_history": [ - { - "entropy": 2.0473356211185454, - "epoch": 0.12062726176115803, - "grad_norm": 3.0554118156433105, - "learning_rate": 1.5682344549790923e-05, - "loss": 1.910550079345703, - "mean_token_accuracy": 0.6053373458981514, - "num_tokens": 121428.0, - "step": 50 - }, - { - "entropy": 1.0371007251739501, - "epoch": 0.24125452352231605, - "grad_norm": 2.109868049621582, - "learning_rate": 3.168473694753677e-05, - "loss": 0.9585584259033203, - "mean_token_accuracy": 0.7352996730804443, - "num_tokens": 243966.0, - "step": 100 - }, - { - "entropy": 0.9105490958690643, - "epoch": 0.3618817852834741, - "grad_norm": 1.7675701379776, - "learning_rate": 4.7687129345282605e-05, - "loss": 0.8393524932861328, - "mean_token_accuracy": 0.7588736474514007, - "num_tokens": 366649.0, - "step": 150 - }, - { - "entropy": 0.8660378754138947, - "epoch": 0.4825090470446321, - "grad_norm": 1.6502987146377563, - "learning_rate": 6.368952174302845e-05, - "loss": 0.8005844116210937, - "mean_token_accuracy": 0.7663125538825989, - "num_tokens": 488367.0, - "step": 200 - }, - { - "entropy": 0.8439123743772506, - "epoch": 0.6031363088057901, - "grad_norm": 1.538365364074707, - "learning_rate": 7.969191414077429e-05, - "loss": 0.7773619842529297, - "mean_token_accuracy": 0.7718527519702911, - "num_tokens": 603131.0, - "step": 250 - }, - { - "entropy": 0.8223806774616241, - "epoch": 0.7237635705669482, - "grad_norm": 1.1559664011001587, - "learning_rate": 9.569430653852014e-05, - "loss": 0.7587767791748047, - "mean_token_accuracy": 0.7767325991392136, - "num_tokens": 726412.0, - "step": 300 - }, - { - "entropy": 0.8256935775279999, - "epoch": 0.8443908323281062, - "grad_norm": 1.2322715520858765, - "learning_rate": 0.00011169669893626597, - "loss": 0.769886245727539, - "mean_token_accuracy": 0.773840668797493, - "num_tokens": 846516.0, - "step": 350 - }, - { - "entropy": 0.8065843790769577, - "epoch": 0.9650180940892642, - "grad_norm": 0.9908286333084106, - "learning_rate": 0.00012769909133401183, - "loss": 0.7451084899902344, - "mean_token_accuracy": 0.7794792425632476, - "num_tokens": 958978.0, - "step": 400 - }, - { - "epoch": 1.0, - "eval_entropy": 0.791056221790528, - "eval_mean_token_accuracy": 0.7507295300451557, - "eval_not_syn_loss": 0.7860428094863892, - "eval_not_syn_runtime": 53.5431, - "eval_not_syn_samples_per_second": 26.521, - "eval_not_syn_steps_per_second": 3.324, - "eval_num_tokens": 993806.0, - "step": 415 - }, - { - "epoch": 1.0, - "eval_entropy": 0.7424047408478983, - "eval_mean_token_accuracy": 0.8023804037758474, - "eval_num_tokens": 993806.0, - "eval_syn_loss": 0.7273509502410889, - "eval_syn_runtime": 54.8679, - "eval_syn_samples_per_second": 25.88, - "eval_syn_steps_per_second": 3.244, - "step": 415 - }, - { - "entropy": 0.7607260761838971, - "epoch": 1.0844390832328106, - "grad_norm": 1.0653696060180664, - "learning_rate": 0.0001327927019027069, - "loss": 0.7034613037109375, - "mean_token_accuracy": 0.7869195998317063, - "num_tokens": 1080301.0, - "step": 450 - }, - { - "entropy": 0.7370894801616669, - "epoch": 1.2050663449939687, - "grad_norm": 1.763787865638733, - "learning_rate": 0.00013265416569994628, - "loss": 0.6816973114013671, - "mean_token_accuracy": 0.7914507347345352, - "num_tokens": 1201215.0, - "step": 500 - }, - { - "entropy": 0.742829909324646, - "epoch": 1.3256936067551266, - "grad_norm": 0.97663414478302, - "learning_rate": 0.00013239847940553975, - "loss": 0.6867338562011719, - "mean_token_accuracy": 0.790328289270401, - "num_tokens": 1322895.0, - "step": 550 - }, - { - "entropy": 0.7312483406066894, - "epoch": 1.4463208685162847, - "grad_norm": 1.0874731540679932, - "learning_rate": 0.00013202609518972978, - "loss": 0.676800537109375, - "mean_token_accuracy": 0.7921512228250503, - "num_tokens": 1436749.0, - "step": 600 - }, - { - "entropy": 0.7178869074583054, - "epoch": 1.5669481302774426, - "grad_norm": 1.064300298690796, - "learning_rate": 0.0001315376715980265, - "loss": 0.6635584259033203, - "mean_token_accuracy": 0.7944776183366775, - "num_tokens": 1555400.0, - "step": 650 - }, - { - "entropy": 0.7198166918754577, - "epoch": 1.6875753920386007, - "grad_norm": 0.7779315114021301, - "learning_rate": 0.00013093407238659802, - "loss": 0.6660567474365234, - "mean_token_accuracy": 0.7968300533294678, - "num_tokens": 1676892.0, - "step": 700 - }, - { - "entropy": 0.7170061504840851, - "epoch": 1.8082026537997589, - "grad_norm": 1.089854121208191, - "learning_rate": 0.00013021636499475496, - "loss": 0.6632771301269531, - "mean_token_accuracy": 0.7963915091753005, - "num_tokens": 1802689.0, - "step": 750 - }, - { - "entropy": 0.6953610718250275, - "epoch": 1.9288299155609168, - "grad_norm": 0.7996730208396912, - "learning_rate": 0.0001293858186572295, - "loss": 0.6421419525146485, - "mean_token_accuracy": 0.8002577543258667, - "num_tokens": 1920804.0, - "step": 800 - }, - { - "epoch": 2.0, - "eval_entropy": 0.7369416523515508, - "eval_mean_token_accuracy": 0.7872022366925572, - "eval_not_syn_loss": 0.713637113571167, - "eval_not_syn_runtime": 53.2265, - "eval_not_syn_samples_per_second": 26.678, - "eval_not_syn_steps_per_second": 3.344, - "eval_num_tokens": 1987612.0, - "step": 830 - }, - { - "epoch": 2.0, - "eval_entropy": 0.6877032756135705, - "eval_mean_token_accuracy": 0.7997466372639945, - "eval_num_tokens": 1987612.0, - "eval_syn_loss": 0.6614266633987427, - "eval_syn_runtime": 54.5178, - "eval_syn_samples_per_second": 26.047, - "eval_syn_steps_per_second": 3.265, - "step": 830 - }, - { - "entropy": 0.6644553433765065, - "epoch": 2.0482509047044632, - "grad_norm": 0.7562468647956848, - "learning_rate": 0.00012844390215958852, - "loss": 0.6112185287475586, - "mean_token_accuracy": 0.806649933559726, - "num_tokens": 2039660.0, - "step": 850 - }, - { - "entropy": 0.6146580249071121, - "epoch": 2.168878166465621, - "grad_norm": 1.098607063293457, - "learning_rate": 0.00012739228124074913, - "loss": 0.5553749084472657, - "mean_token_accuracy": 0.8187588346004486, - "num_tokens": 2161772.0, - "step": 900 - }, - { - "entropy": 0.6187627577781677, - "epoch": 2.2895054282267795, - "grad_norm": 1.0338040590286255, - "learning_rate": 0.00012623281564719095, - "loss": 0.5596194076538086, - "mean_token_accuracy": 0.8182566928863525, - "num_tokens": 2280833.0, - "step": 950 - }, - { - "entropy": 0.6232365870475769, - "epoch": 2.4101326899879374, - "grad_norm": 1.0010799169540405, - "learning_rate": 0.00012496755584407422, - "loss": 0.5667143249511719, - "mean_token_accuracy": 0.8175061982870102, - "num_tokens": 2398206.0, - "step": 1000 - }, - { - "entropy": 0.6151268655061721, - "epoch": 2.5307599517490953, - "grad_norm": 1.0065034627914429, - "learning_rate": 0.0001235987393890801, - "loss": 0.5598890686035156, - "mean_token_accuracy": 0.8193449324369431, - "num_tokens": 2518888.0, - "step": 1050 - }, - { - "entropy": 0.6173626267910004, - "epoch": 2.651387213510253, - "grad_norm": 0.7958805561065674, - "learning_rate": 0.00012212878697538592, - "loss": 0.5649292755126953, - "mean_token_accuracy": 0.8172902607917786, - "num_tokens": 2639763.0, - "step": 1100 - }, - { - "entropy": 0.6264930808544159, - "epoch": 2.772014475271411, - "grad_norm": 0.7718372344970703, - "learning_rate": 0.00012056029815077323, - "loss": 0.5716455078125, - "mean_token_accuracy": 0.8170186239480972, - "num_tokens": 2755540.0, - "step": 1150 - }, - { - "entropy": 0.6203387409448624, - "epoch": 2.8926417370325694, - "grad_norm": 0.7914765477180481, - "learning_rate": 0.000118896046720439, - "loss": 0.5653961944580078, - "mean_token_accuracy": 0.8170280295610428, - "num_tokens": 2874118.0, - "step": 1200 - }, - { - "epoch": 3.0, - "eval_entropy": 0.6751241854737314, - "eval_mean_token_accuracy": 0.792226163858778, - "eval_not_syn_loss": 0.6902772784233093, - "eval_not_syn_runtime": 53.2557, - "eval_not_syn_samples_per_second": 26.664, - "eval_not_syn_steps_per_second": 3.342, - "eval_num_tokens": 2981418.0, - "step": 1245 - }, - { - "epoch": 3.0, - "eval_entropy": 0.6308753868837035, - "eval_mean_token_accuracy": 0.8131611759073278, - "eval_num_tokens": 2981418.0, - "eval_syn_loss": 0.6440441012382507, - "eval_syn_runtime": 54.7064, - "eval_syn_samples_per_second": 25.957, - "eval_syn_steps_per_second": 3.254, - "step": 1245 - } - ], - "logging_steps": 50, - "max_steps": 4150, - "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": 6.34152082269024e+16, - "train_batch_size": 8, - "trial_name": null, - "trial_params": null -} diff --git a/substitutivity_original_Estonian/Qwen3.5-2B-Base_substitutivity_splits_original_features_train_substitutivity_splits_original_features_test1/checkpoint-1245/training_args.bin b/substitutivity_original_Estonian/Qwen3.5-2B-Base_substitutivity_splits_original_features_train_substitutivity_splits_original_features_test1/checkpoint-1245/training_args.bin deleted file mode 100644 index 48901efb30506830185bfaafecbe236e7095a006..0000000000000000000000000000000000000000 --- a/substitutivity_original_Estonian/Qwen3.5-2B-Base_substitutivity_splits_original_features_train_substitutivity_splits_original_features_test1/checkpoint-1245/training_args.bin +++ /dev/null @@ -1,3 +0,0 @@ -version https://git-lfs.github.com/spec/v1 -oid sha256:770fc15ebc27a54d780af597d32c57ca14b3b528f14788d142bdfce09b555bef -size 6033 diff --git a/substitutivity_original_Estonian/Qwen3.5-2B-Base_substitutivity_splits_original_features_train_substitutivity_splits_original_features_test1/checkpoint-1660/README.md b/substitutivity_original_Estonian/Qwen3.5-2B-Base_substitutivity_splits_original_features_train_substitutivity_splits_original_features_test1/checkpoint-1660/README.md deleted file mode 100644 index ece204d005f7bf161d0e448efdd56fd297b7c8a2..0000000000000000000000000000000000000000 --- a/substitutivity_original_Estonian/Qwen3.5-2B-Base_substitutivity_splits_original_features_train_substitutivity_splits_original_features_test1/checkpoint-1660/README.md +++ /dev/null @@ -1,209 +0,0 @@ ---- -base_model: Qwen/Qwen3.5-2B-Base -library_name: peft -pipeline_tag: text-generation -tags: -- base_model:adapter:Qwen/Qwen3.5-2B-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/substitutivity_original_Estonian/Qwen3.5-2B-Base_substitutivity_splits_original_features_train_substitutivity_splits_original_features_test1/checkpoint-1660/adapter_config.json b/substitutivity_original_Estonian/Qwen3.5-2B-Base_substitutivity_splits_original_features_train_substitutivity_splits_original_features_test1/checkpoint-1660/adapter_config.json deleted file mode 100644 index a64145ce23adcbdbe246679087a13f4cb63ff664..0000000000000000000000000000000000000000 --- a/substitutivity_original_Estonian/Qwen3.5-2B-Base_substitutivity_splits_original_features_train_substitutivity_splits_original_features_test1/checkpoint-1660/adapter_config.json +++ /dev/null @@ -1,46 +0,0 @@ -{ - "alora_invocation_tokens": null, - "alpha_pattern": {}, - "arrow_config": null, - "auto_mapping": null, - "base_model_name_or_path": "Qwen/Qwen3.5-2B-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": 256, - "lora_bias": false, - "lora_dropout": 0.027050353915701653, - "megatron_config": null, - "megatron_core": "megatron.core", - "modules_to_save": null, - "peft_type": "LORA", - "peft_version": "0.18.1", - "qalora_group_size": 16, - "r": 128, - "rank_pattern": {}, - "revision": null, - "target_modules": [ - "o_proj", - "q_proj", - "k_proj", - "down_proj", - "up_proj", - "v_proj", - "gate_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/substitutivity_original_Estonian/Qwen3.5-2B-Base_substitutivity_splits_original_features_train_substitutivity_splits_original_features_test1/checkpoint-1660/adapter_model.safetensors b/substitutivity_original_Estonian/Qwen3.5-2B-Base_substitutivity_splits_original_features_train_substitutivity_splits_original_features_test1/checkpoint-1660/adapter_model.safetensors deleted file mode 100644 index fc19c15dd8d5df731b5c4ca3419ff66b6cecb9da..0000000000000000000000000000000000000000 --- a/substitutivity_original_Estonian/Qwen3.5-2B-Base_substitutivity_splits_original_features_train_substitutivity_splits_original_features_test1/checkpoint-1660/adapter_model.safetensors +++ /dev/null @@ -1,3 +0,0 @@ -version https://git-lfs.github.com/spec/v1 -oid sha256:83d8bc1e85309f49a0aa7ffc2f3abfdb68bfe02366a8bcb68a350ceeb1adc7e2 -size 349201608 diff --git a/substitutivity_original_Estonian/Qwen3.5-2B-Base_substitutivity_splits_original_features_train_substitutivity_splits_original_features_test1/checkpoint-1660/chat_template.jinja b/substitutivity_original_Estonian/Qwen3.5-2B-Base_substitutivity_splits_original_features_train_substitutivity_splits_original_features_test1/checkpoint-1660/chat_template.jinja deleted file mode 100644 index 0ef09f214eaa6d9bca297988afc1454b5827b2c7..0000000000000000000000000000000000000000 --- a/substitutivity_original_Estonian/Qwen3.5-2B-Base_substitutivity_splits_original_features_train_substitutivity_splits_original_features_test1/checkpoint-1660/chat_template.jinja +++ /dev/null @@ -1,154 +0,0 @@ -{%- set image_count = namespace(value=0) %} -{%- set video_count = namespace(value=0) %} -{%- macro render_content(content, do_vision_count, is_system_content=false) %} - {%- if content is string %} - {{- content }} - {%- elif content is iterable and content is not mapping %} - {%- for item in content %} - {%- if 'image' in item or 'image_url' in item or item.type == 'image' %} - {%- if is_system_content %} - {{- raise_exception('System message cannot contain images.') }} - {%- endif %} - {%- if do_vision_count %} - {%- set image_count.value = image_count.value + 1 %} - {%- endif %} - {%- if add_vision_id %} - {{- 'Picture ' ~ image_count.value ~ ': ' }} - {%- endif %} - {{- '<|vision_start|><|image_pad|><|vision_end|>' }} - {%- elif 'video' in item or item.type == 'video' %} - {%- if is_system_content %} - {{- raise_exception('System message cannot contain videos.') }} - {%- endif %} - {%- if do_vision_count %} - {%- set video_count.value = video_count.value + 1 %} - {%- endif %} - {%- if add_vision_id %} - {{- 'Video ' ~ video_count.value ~ ': ' }} - {%- endif %} - {{- '<|vision_start|><|video_pad|><|vision_end|>' }} - {%- elif 'text' in item %} - {{- item.text }} - {%- else %} - {{- raise_exception('Unexpected item type in content.') }} - {%- endif %} - {%- endfor %} - {%- elif content is none or content is undefined %} - {{- '' }} - {%- else %} - {{- raise_exception('Unexpected content type.') }} - {%- endif %} -{%- endmacro %} -{%- if not messages %} - {{- raise_exception('No messages provided.') }} -{%- endif %} -{%- if tools and tools is iterable and tools is not mapping %} - {{- '<|im_start|>system\n' }} - {{- "# Tools\n\nYou have access to the following functions:\n\n" }} - {%- for tool in tools %} - {{- "\n" }} - {{- tool | tojson }} - {%- endfor %} - {{- "\n" }} - {{- '\n\nIf you choose to call a function ONLY reply in the following format with NO suffix:\n\n\n\n\nvalue_1\n\n\nThis is the value for the second parameter\nthat can span\nmultiple lines\n\n\n\n\n\nReminder:\n- Function calls MUST follow the specified format: an inner block must be nested within XML tags\n- Required parameters MUST be specified\n- You may provide optional reasoning for your function call in natural language BEFORE the function call, but NOT after\n- If there is no function call available, answer the question like normal with your current knowledge and do not tell the user about function calls\n' }} - {%- if messages[0].role == 'system' %} - {%- set content = render_content(messages[0].content, false, true)|trim %} - {%- if content %} - {{- '\n\n' + content }} - {%- endif %} - {%- endif %} - {{- '<|im_end|>\n' }} -{%- else %} - {%- if messages[0].role == 'system' %} - {%- set content = render_content(messages[0].content, false, true)|trim %} - {{- '<|im_start|>system\n' + 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" %} - {%- set content = render_content(message.content, false)|trim %} - {%- if not(content.startswith('') and content.endswith('')) %} - {%- set ns.multi_step_tool = false %} - {%- set ns.last_query_index = index %} - {%- endif %} - {%- endif %} -{%- endfor %} -{%- if ns.multi_step_tool %} - {{- raise_exception('No user query found in messages.') }} -{%- endif %} -{%- for message in messages %} - {%- set content = render_content(message.content, true)|trim %} - {%- if message.role == "system" %} - {%- if not loop.first %} - {{- raise_exception('System message must be at the beginning.') }} - {%- endif %} - {%- elif message.role == "user" %} - {{- '<|im_start|>' + message.role + '\n' + content + '<|im_end|>' + '\n' }} - {%- elif message.role == "assistant" %} - {%- set reasoning_content = '' %} - {%- if message.reasoning_content is string %} - {%- set reasoning_content = message.reasoning_content %} - {%- else %} - {%- if '' in content %} - {%- set reasoning_content = content.split('')[0].rstrip('\n').split('')[-1].lstrip('\n') %} - {%- set content = content.split('')[-1].lstrip('\n') %} - {%- endif %} - {%- endif %} - {%- set reasoning_content = reasoning_content|trim %} - {%- if loop.index0 > ns.last_query_index %} - {{- '<|im_start|>' + message.role + '\n\n' + reasoning_content + '\n\n\n' + content }} - {%- else %} - {{- '<|im_start|>' + message.role + '\n' + content }} - {%- endif %} - {%- if message.tool_calls and message.tool_calls is iterable and message.tool_calls is not mapping %} - {%- for tool_call in message.tool_calls %} - {%- if tool_call.function is defined %} - {%- set tool_call = tool_call.function %} - {%- endif %} - {%- if loop.first %} - {%- if content|trim %} - {{- '\n\n\n\n' }} - {%- else %} - {{- '\n\n' }} - {%- endif %} - {%- else %} - {{- '\n\n\n' }} - {%- endif %} - {%- if tool_call.arguments is defined %} - {%- for args_name, args_value in tool_call.arguments|items %} - {{- '\n' }} - {%- set args_value = args_value | tojson | safe if args_value is mapping or (args_value is sequence and args_value is not string) else args_value | string %} - {{- args_value }} - {{- '\n\n' }} - {%- endfor %} - {%- endif %} - {{- '\n' }} - {%- endfor %} - {%- endif %} - {{- '<|im_end|>\n' }} - {%- elif message.role == "tool" %} - {%- if loop.previtem and loop.previtem.role != "tool" %} - {{- '<|im_start|>user' }} - {%- endif %} - {{- '\n\n' }} - {{- content }} - {{- '\n' }} - {%- if not loop.last and loop.nextitem.role != "tool" %} - {{- '<|im_end|>\n' }} - {%- elif loop.last %} - {{- '<|im_end|>\n' }} - {%- endif %} - {%- else %} - {{- raise_exception('Unexpected message role.') }} - {%- endif %} -{%- endfor %} -{%- if add_generation_prompt %} - {{- '<|im_start|>assistant\n' }} - {%- if enable_thinking is defined and enable_thinking is true %} - {{- '\n' }} - {%- else %} - {{- '\n\n\n\n' }} - {%- endif %} -{%- endif %} \ No newline at end of file diff --git a/substitutivity_original_Estonian/Qwen3.5-2B-Base_substitutivity_splits_original_features_train_substitutivity_splits_original_features_test1/checkpoint-1660/tokenizer.json b/substitutivity_original_Estonian/Qwen3.5-2B-Base_substitutivity_splits_original_features_train_substitutivity_splits_original_features_test1/checkpoint-1660/tokenizer.json deleted file mode 100644 index 67741b04f23bfdb46501f748ce27865ec82eccfb..0000000000000000000000000000000000000000 --- a/substitutivity_original_Estonian/Qwen3.5-2B-Base_substitutivity_splits_original_features_train_substitutivity_splits_original_features_test1/checkpoint-1660/tokenizer.json +++ /dev/null @@ -1,3 +0,0 @@ -version https://git-lfs.github.com/spec/v1 -oid sha256:87a7830d63fcf43bf241c3c5242e96e62dd3fdc29224ca26fed8ea333db72de4 -size 19989343 diff --git a/substitutivity_original_Estonian/Qwen3.5-2B-Base_substitutivity_splits_original_features_train_substitutivity_splits_original_features_test1/checkpoint-1660/tokenizer_config.json b/substitutivity_original_Estonian/Qwen3.5-2B-Base_substitutivity_splits_original_features_train_substitutivity_splits_original_features_test1/checkpoint-1660/tokenizer_config.json deleted file mode 100644 index b4a37b2a6fd3ab3317cd7bac72855be1a843b2bb..0000000000000000000000000000000000000000 --- a/substitutivity_original_Estonian/Qwen3.5-2B-Base_substitutivity_splits_original_features_train_substitutivity_splits_original_features_test1/checkpoint-1660/tokenizer_config.json +++ /dev/null @@ -1,31 +0,0 @@ -{ - "add_prefix_space": false, - "audio_bos_token": "<|audio_start|>", - "audio_eos_token": "<|audio_end|>", - "audio_token": "<|audio_pad|>", - "backend": "tokenizers", - "bos_token": null, - "clean_up_tokenization_spaces": false, - "eos_token": "<|endoftext|>", - "errors": "replace", - "image_token": "<|image_pad|>", - "is_local": false, - "model_max_length": 262144, - "model_specific_special_tokens": { - "audio_bos_token": "<|audio_start|>", - "audio_eos_token": "<|audio_end|>", - "audio_token": "<|audio_pad|>", - "image_token": "<|image_pad|>", - "video_token": "<|video_pad|>", - "vision_bos_token": "<|vision_start|>", - "vision_eos_token": "<|vision_end|>" - }, - "pad_token": "<|endoftext|>", - "pretokenize_regex": "(?i:'s|'t|'re|'ve|'m|'ll|'d)|[^\\r\\n\\p{L}\\p{N}]?[\\p{L}\\p{M}]+|\\p{N}| ?[^\\s\\p{L}\\p{M}\\p{N}]+[\\r\\n]*|\\s*[\\r\\n]+|\\s+(?!\\S)|\\s+", - "split_special_tokens": false, - "tokenizer_class": "TokenizersBackend", - "unk_token": null, - "video_token": "<|video_pad|>", - "vision_bos_token": "<|vision_start|>", - "vision_eos_token": "<|vision_end|>" -} diff --git a/substitutivity_original_Estonian/Qwen3.5-2B-Base_substitutivity_splits_original_features_train_substitutivity_splits_original_features_test1/checkpoint-1660/trainer_state.json b/substitutivity_original_Estonian/Qwen3.5-2B-Base_substitutivity_splits_original_features_train_substitutivity_splits_original_features_test1/checkpoint-1660/trainer_state.json deleted file mode 100644 index 4c91a1f39b3484c13b91bcc0934d2f7a01138879..0000000000000000000000000000000000000000 --- a/substitutivity_original_Estonian/Qwen3.5-2B-Base_substitutivity_splits_original_features_train_substitutivity_splits_original_features_test1/checkpoint-1660/trainer_state.json +++ /dev/null @@ -1,452 +0,0 @@ -{ - "best_global_step": null, - "best_metric": null, - "best_model_checkpoint": null, - "epoch": 4.0, - "eval_steps": 500, - "global_step": 1660, - "is_hyper_param_search": false, - "is_local_process_zero": true, - "is_world_process_zero": true, - "log_history": [ - { - "entropy": 2.0473356211185454, - "epoch": 0.12062726176115803, - "grad_norm": 3.0554118156433105, - "learning_rate": 1.5682344549790923e-05, - "loss": 1.910550079345703, - "mean_token_accuracy": 0.6053373458981514, - "num_tokens": 121428.0, - "step": 50 - }, - { - "entropy": 1.0371007251739501, - "epoch": 0.24125452352231605, - "grad_norm": 2.109868049621582, - "learning_rate": 3.168473694753677e-05, - "loss": 0.9585584259033203, - "mean_token_accuracy": 0.7352996730804443, - "num_tokens": 243966.0, - "step": 100 - }, - { - "entropy": 0.9105490958690643, - "epoch": 0.3618817852834741, - "grad_norm": 1.7675701379776, - "learning_rate": 4.7687129345282605e-05, - "loss": 0.8393524932861328, - "mean_token_accuracy": 0.7588736474514007, - "num_tokens": 366649.0, - "step": 150 - }, - { - "entropy": 0.8660378754138947, - "epoch": 0.4825090470446321, - "grad_norm": 1.6502987146377563, - "learning_rate": 6.368952174302845e-05, - "loss": 0.8005844116210937, - "mean_token_accuracy": 0.7663125538825989, - "num_tokens": 488367.0, - "step": 200 - }, - { - "entropy": 0.8439123743772506, - "epoch": 0.6031363088057901, - "grad_norm": 1.538365364074707, - "learning_rate": 7.969191414077429e-05, - "loss": 0.7773619842529297, - "mean_token_accuracy": 0.7718527519702911, - "num_tokens": 603131.0, - "step": 250 - }, - { - "entropy": 0.8223806774616241, - "epoch": 0.7237635705669482, - "grad_norm": 1.1559664011001587, - "learning_rate": 9.569430653852014e-05, - "loss": 0.7587767791748047, - "mean_token_accuracy": 0.7767325991392136, - "num_tokens": 726412.0, - "step": 300 - }, - { - "entropy": 0.8256935775279999, - "epoch": 0.8443908323281062, - "grad_norm": 1.2322715520858765, - "learning_rate": 0.00011169669893626597, - "loss": 0.769886245727539, - "mean_token_accuracy": 0.773840668797493, - "num_tokens": 846516.0, - "step": 350 - }, - { - "entropy": 0.8065843790769577, - "epoch": 0.9650180940892642, - "grad_norm": 0.9908286333084106, - "learning_rate": 0.00012769909133401183, - "loss": 0.7451084899902344, - "mean_token_accuracy": 0.7794792425632476, - "num_tokens": 958978.0, - "step": 400 - }, - { - "epoch": 1.0, - "eval_entropy": 0.791056221790528, - "eval_mean_token_accuracy": 0.7507295300451557, - "eval_not_syn_loss": 0.7860428094863892, - "eval_not_syn_runtime": 53.5431, - "eval_not_syn_samples_per_second": 26.521, - "eval_not_syn_steps_per_second": 3.324, - "eval_num_tokens": 993806.0, - "step": 415 - }, - { - "epoch": 1.0, - "eval_entropy": 0.7424047408478983, - "eval_mean_token_accuracy": 0.8023804037758474, - "eval_num_tokens": 993806.0, - "eval_syn_loss": 0.7273509502410889, - "eval_syn_runtime": 54.8679, - "eval_syn_samples_per_second": 25.88, - "eval_syn_steps_per_second": 3.244, - "step": 415 - }, - { - "entropy": 0.7607260761838971, - "epoch": 1.0844390832328106, - "grad_norm": 1.0653696060180664, - "learning_rate": 0.0001327927019027069, - "loss": 0.7034613037109375, - "mean_token_accuracy": 0.7869195998317063, - "num_tokens": 1080301.0, - "step": 450 - }, - { - "entropy": 0.7370894801616669, - "epoch": 1.2050663449939687, - "grad_norm": 1.763787865638733, - "learning_rate": 0.00013265416569994628, - "loss": 0.6816973114013671, - "mean_token_accuracy": 0.7914507347345352, - "num_tokens": 1201215.0, - "step": 500 - }, - { - "entropy": 0.742829909324646, - "epoch": 1.3256936067551266, - "grad_norm": 0.97663414478302, - "learning_rate": 0.00013239847940553975, - "loss": 0.6867338562011719, - "mean_token_accuracy": 0.790328289270401, - "num_tokens": 1322895.0, - "step": 550 - }, - { - "entropy": 0.7312483406066894, - "epoch": 1.4463208685162847, - "grad_norm": 1.0874731540679932, - "learning_rate": 0.00013202609518972978, - "loss": 0.676800537109375, - "mean_token_accuracy": 0.7921512228250503, - "num_tokens": 1436749.0, - "step": 600 - }, - { - "entropy": 0.7178869074583054, - "epoch": 1.5669481302774426, - "grad_norm": 1.064300298690796, - "learning_rate": 0.0001315376715980265, - "loss": 0.6635584259033203, - "mean_token_accuracy": 0.7944776183366775, - "num_tokens": 1555400.0, - "step": 650 - }, - { - "entropy": 0.7198166918754577, - "epoch": 1.6875753920386007, - "grad_norm": 0.7779315114021301, - "learning_rate": 0.00013093407238659802, - "loss": 0.6660567474365234, - "mean_token_accuracy": 0.7968300533294678, - "num_tokens": 1676892.0, - "step": 700 - }, - { - "entropy": 0.7170061504840851, - "epoch": 1.8082026537997589, - "grad_norm": 1.089854121208191, - "learning_rate": 0.00013021636499475496, - "loss": 0.6632771301269531, - "mean_token_accuracy": 0.7963915091753005, - "num_tokens": 1802689.0, - "step": 750 - }, - { - "entropy": 0.6953610718250275, - "epoch": 1.9288299155609168, - "grad_norm": 0.7996730208396912, - "learning_rate": 0.0001293858186572295, - "loss": 0.6421419525146485, - "mean_token_accuracy": 0.8002577543258667, - "num_tokens": 1920804.0, - "step": 800 - }, - { - "epoch": 2.0, - "eval_entropy": 0.7369416523515508, - "eval_mean_token_accuracy": 0.7872022366925572, - "eval_not_syn_loss": 0.713637113571167, - "eval_not_syn_runtime": 53.2265, - "eval_not_syn_samples_per_second": 26.678, - "eval_not_syn_steps_per_second": 3.344, - "eval_num_tokens": 1987612.0, - "step": 830 - }, - { - "epoch": 2.0, - "eval_entropy": 0.6877032756135705, - "eval_mean_token_accuracy": 0.7997466372639945, - "eval_num_tokens": 1987612.0, - "eval_syn_loss": 0.6614266633987427, - "eval_syn_runtime": 54.5178, - "eval_syn_samples_per_second": 26.047, - "eval_syn_steps_per_second": 3.265, - "step": 830 - }, - { - "entropy": 0.6644553433765065, - "epoch": 2.0482509047044632, - "grad_norm": 0.7562468647956848, - "learning_rate": 0.00012844390215958852, - "loss": 0.6112185287475586, - "mean_token_accuracy": 0.806649933559726, - "num_tokens": 2039660.0, - "step": 850 - }, - { - "entropy": 0.6146580249071121, - "epoch": 2.168878166465621, - "grad_norm": 1.098607063293457, - "learning_rate": 0.00012739228124074913, - "loss": 0.5553749084472657, - "mean_token_accuracy": 0.8187588346004486, - "num_tokens": 2161772.0, - "step": 900 - }, - { - "entropy": 0.6187627577781677, - "epoch": 2.2895054282267795, - "grad_norm": 1.0338040590286255, - "learning_rate": 0.00012623281564719095, - "loss": 0.5596194076538086, - "mean_token_accuracy": 0.8182566928863525, - "num_tokens": 2280833.0, - "step": 950 - }, - { - "entropy": 0.6232365870475769, - "epoch": 2.4101326899879374, - "grad_norm": 1.0010799169540405, - "learning_rate": 0.00012496755584407422, - "loss": 0.5667143249511719, - "mean_token_accuracy": 0.8175061982870102, - "num_tokens": 2398206.0, - "step": 1000 - }, - { - "entropy": 0.6151268655061721, - "epoch": 2.5307599517490953, - "grad_norm": 1.0065034627914429, - "learning_rate": 0.0001235987393890801, - "loss": 0.5598890686035156, - "mean_token_accuracy": 0.8193449324369431, - "num_tokens": 2518888.0, - "step": 1050 - }, - { - "entropy": 0.6173626267910004, - "epoch": 2.651387213510253, - "grad_norm": 0.7958805561065674, - "learning_rate": 0.00012212878697538592, - "loss": 0.5649292755126953, - "mean_token_accuracy": 0.8172902607917786, - "num_tokens": 2639763.0, - "step": 1100 - }, - { - "entropy": 0.6264930808544159, - "epoch": 2.772014475271411, - "grad_norm": 0.7718372344970703, - "learning_rate": 0.00012056029815077323, - "loss": 0.5716455078125, - "mean_token_accuracy": 0.8170186239480972, - "num_tokens": 2755540.0, - "step": 1150 - }, - { - "entropy": 0.6203387409448624, - "epoch": 2.8926417370325694, - "grad_norm": 0.7914765477180481, - "learning_rate": 0.000118896046720439, - "loss": 0.5653961944580078, - "mean_token_accuracy": 0.8170280295610428, - "num_tokens": 2874118.0, - "step": 1200 - }, - { - "epoch": 3.0, - "eval_entropy": 0.6751241854737314, - "eval_mean_token_accuracy": 0.792226163858778, - "eval_not_syn_loss": 0.6902772784233093, - "eval_not_syn_runtime": 53.2557, - "eval_not_syn_samples_per_second": 26.664, - "eval_not_syn_steps_per_second": 3.342, - "eval_num_tokens": 2981418.0, - "step": 1245 - }, - { - "epoch": 3.0, - "eval_entropy": 0.6308753868837035, - "eval_mean_token_accuracy": 0.8131611759073278, - "eval_num_tokens": 2981418.0, - "eval_syn_loss": 0.6440441012382507, - "eval_syn_runtime": 54.7064, - "eval_syn_samples_per_second": 25.957, - "eval_syn_steps_per_second": 3.254, - "step": 1245 - }, - { - "entropy": 0.6077835421369533, - "epoch": 3.012062726176116, - "grad_norm": 1.124720573425293, - "learning_rate": 0.00011713897584164003, - "loss": 0.5557136917114258, - "mean_token_accuracy": 0.8203362726201915, - "num_tokens": 2994602.0, - "step": 1250 - }, - { - "entropy": 0.507642827630043, - "epoch": 3.132689987937274, - "grad_norm": 0.9039916396141052, - "learning_rate": 0.00011529219281884579, - "loss": 0.4467372512817383, - "mean_token_accuracy": 0.8464260965585708, - "num_tokens": 3121094.0, - "step": 1300 - }, - { - "entropy": 0.5137191081047058, - "epoch": 3.2533172496984317, - "grad_norm": 0.8580691814422607, - "learning_rate": 0.00011335896360860367, - "loss": 0.4538694763183594, - "mean_token_accuracy": 0.8448719161748887, - "num_tokens": 3248038.0, - "step": 1350 - }, - { - "entropy": 0.5293738862872124, - "epoch": 3.37394451145959, - "grad_norm": 1.0177273750305176, - "learning_rate": 0.00011134270704383492, - "loss": 0.4639348220825195, - "mean_token_accuracy": 0.8428975892066956, - "num_tokens": 3363487.0, - "step": 1400 - }, - { - "entropy": 0.5218872320652008, - "epoch": 3.494571773220748, - "grad_norm": 1.0530673265457153, - "learning_rate": 0.00010924698878777538, - "loss": 0.4584346389770508, - "mean_token_accuracy": 0.8432288640737533, - "num_tokens": 3480513.0, - "step": 1450 - }, - { - "entropy": 0.5320372316241264, - "epoch": 3.615199034981906, - "grad_norm": 1.0291194915771484, - "learning_rate": 0.0001070755150282529, - "loss": 0.4711898422241211, - "mean_token_accuracy": 0.842066935300827, - "num_tokens": 3597183.0, - "step": 1500 - }, - { - "entropy": 0.5315000694990158, - "epoch": 3.7358262967430638, - "grad_norm": 0.9788838624954224, - "learning_rate": 0.00010483212592345332, - "loss": 0.4729101943969727, - "mean_token_accuracy": 0.8404586082696914, - "num_tokens": 3714173.0, - "step": 1550 - }, - { - "entropy": 0.5331643190979958, - "epoch": 3.856453558504222, - "grad_norm": 1.1248903274536133, - "learning_rate": 0.00010252078881076532, - "loss": 0.47405517578125, - "mean_token_accuracy": 0.8397526168823242, - "num_tokens": 3832652.0, - "step": 1600 - }, - { - "entropy": 0.5343865439295769, - "epoch": 3.97708082026538, - "grad_norm": 0.9008272886276245, - "learning_rate": 0.00010014559119071452, - "loss": 0.47172779083251953, - "mean_token_accuracy": 0.8404951041936874, - "num_tokens": 3955167.0, - "step": 1650 - }, - { - "epoch": 4.0, - "eval_entropy": 0.5736920103598177, - "eval_mean_token_accuracy": 0.7981750027517255, - "eval_not_syn_loss": 0.707631528377533, - "eval_not_syn_runtime": 52.2858, - "eval_not_syn_samples_per_second": 27.158, - "eval_not_syn_steps_per_second": 3.404, - "eval_num_tokens": 3975224.0, - "step": 1660 - }, - { - "epoch": 4.0, - "eval_entropy": 0.5363499959198277, - "eval_mean_token_accuracy": 0.8063885539435269, - "eval_num_tokens": 3975224.0, - "eval_syn_loss": 0.6610181331634521, - "eval_syn_runtime": 54.2167, - "eval_syn_samples_per_second": 26.191, - "eval_syn_steps_per_second": 3.283, - "step": 1660 - } - ], - "logging_steps": 50, - "max_steps": 4150, - "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": 8.481153412421376e+16, - "train_batch_size": 8, - "trial_name": null, - "trial_params": null -} diff --git a/substitutivity_original_Estonian/Qwen3.5-2B-Base_substitutivity_splits_original_features_train_substitutivity_splits_original_features_test1/checkpoint-1660/training_args.bin b/substitutivity_original_Estonian/Qwen3.5-2B-Base_substitutivity_splits_original_features_train_substitutivity_splits_original_features_test1/checkpoint-1660/training_args.bin deleted file mode 100644 index 48901efb30506830185bfaafecbe236e7095a006..0000000000000000000000000000000000000000 --- a/substitutivity_original_Estonian/Qwen3.5-2B-Base_substitutivity_splits_original_features_train_substitutivity_splits_original_features_test1/checkpoint-1660/training_args.bin +++ /dev/null @@ -1,3 +0,0 @@ -version https://git-lfs.github.com/spec/v1 -oid sha256:770fc15ebc27a54d780af597d32c57ca14b3b528f14788d142bdfce09b555bef -size 6033 diff --git a/substitutivity_original_Estonian/Qwen3.5-2B-Base_substitutivity_splits_original_features_train_substitutivity_splits_original_features_test1/checkpoint-2075/README.md b/substitutivity_original_Estonian/Qwen3.5-2B-Base_substitutivity_splits_original_features_train_substitutivity_splits_original_features_test1/checkpoint-2075/README.md deleted file mode 100644 index ece204d005f7bf161d0e448efdd56fd297b7c8a2..0000000000000000000000000000000000000000 --- a/substitutivity_original_Estonian/Qwen3.5-2B-Base_substitutivity_splits_original_features_train_substitutivity_splits_original_features_test1/checkpoint-2075/README.md +++ /dev/null @@ -1,209 +0,0 @@ ---- -base_model: Qwen/Qwen3.5-2B-Base -library_name: peft -pipeline_tag: text-generation -tags: -- base_model:adapter:Qwen/Qwen3.5-2B-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/substitutivity_original_Estonian/Qwen3.5-2B-Base_substitutivity_splits_original_features_train_substitutivity_splits_original_features_test1/checkpoint-2075/adapter_config.json b/substitutivity_original_Estonian/Qwen3.5-2B-Base_substitutivity_splits_original_features_train_substitutivity_splits_original_features_test1/checkpoint-2075/adapter_config.json deleted file mode 100644 index a64145ce23adcbdbe246679087a13f4cb63ff664..0000000000000000000000000000000000000000 --- a/substitutivity_original_Estonian/Qwen3.5-2B-Base_substitutivity_splits_original_features_train_substitutivity_splits_original_features_test1/checkpoint-2075/adapter_config.json +++ /dev/null @@ -1,46 +0,0 @@ -{ - "alora_invocation_tokens": null, - "alpha_pattern": {}, - "arrow_config": null, - "auto_mapping": null, - "base_model_name_or_path": "Qwen/Qwen3.5-2B-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": 256, - "lora_bias": false, - "lora_dropout": 0.027050353915701653, - "megatron_config": null, - "megatron_core": "megatron.core", - "modules_to_save": null, - "peft_type": "LORA", - "peft_version": "0.18.1", - "qalora_group_size": 16, - "r": 128, - "rank_pattern": {}, - "revision": null, - "target_modules": [ - "o_proj", - "q_proj", - "k_proj", - "down_proj", - "up_proj", - "v_proj", - "gate_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/substitutivity_original_Estonian/Qwen3.5-2B-Base_substitutivity_splits_original_features_train_substitutivity_splits_original_features_test1/checkpoint-2075/adapter_model.safetensors b/substitutivity_original_Estonian/Qwen3.5-2B-Base_substitutivity_splits_original_features_train_substitutivity_splits_original_features_test1/checkpoint-2075/adapter_model.safetensors deleted file mode 100644 index 06550741bad9b83f6dbe35b3a38b85c091aea3f7..0000000000000000000000000000000000000000 --- a/substitutivity_original_Estonian/Qwen3.5-2B-Base_substitutivity_splits_original_features_train_substitutivity_splits_original_features_test1/checkpoint-2075/adapter_model.safetensors +++ /dev/null @@ -1,3 +0,0 @@ -version https://git-lfs.github.com/spec/v1 -oid sha256:ae3992049a09e84745f3e9ff29b6582c4efb1c244af17d72f891cb248b6e4593 -size 349201608 diff --git a/substitutivity_original_Estonian/Qwen3.5-2B-Base_substitutivity_splits_original_features_train_substitutivity_splits_original_features_test1/checkpoint-2075/chat_template.jinja b/substitutivity_original_Estonian/Qwen3.5-2B-Base_substitutivity_splits_original_features_train_substitutivity_splits_original_features_test1/checkpoint-2075/chat_template.jinja deleted file mode 100644 index 0ef09f214eaa6d9bca297988afc1454b5827b2c7..0000000000000000000000000000000000000000 --- a/substitutivity_original_Estonian/Qwen3.5-2B-Base_substitutivity_splits_original_features_train_substitutivity_splits_original_features_test1/checkpoint-2075/chat_template.jinja +++ /dev/null @@ -1,154 +0,0 @@ -{%- set image_count = namespace(value=0) %} -{%- set video_count = namespace(value=0) %} -{%- macro render_content(content, do_vision_count, is_system_content=false) %} - {%- if content is string %} - {{- content }} - {%- elif content is iterable and content is not mapping %} - {%- for item in content %} - {%- if 'image' in item or 'image_url' in item or item.type == 'image' %} - {%- if is_system_content %} - {{- raise_exception('System message cannot contain images.') }} - {%- endif %} - {%- if do_vision_count %} - {%- set image_count.value = image_count.value + 1 %} - {%- endif %} - {%- if add_vision_id %} - {{- 'Picture ' ~ image_count.value ~ ': ' }} - {%- endif %} - {{- '<|vision_start|><|image_pad|><|vision_end|>' }} - {%- elif 'video' in item or item.type == 'video' %} - {%- if is_system_content %} - {{- raise_exception('System message cannot contain videos.') }} - {%- endif %} - {%- if do_vision_count %} - {%- set video_count.value = video_count.value + 1 %} - {%- endif %} - {%- if add_vision_id %} - {{- 'Video ' ~ video_count.value ~ ': ' }} - {%- endif %} - {{- '<|vision_start|><|video_pad|><|vision_end|>' }} - {%- elif 'text' in item %} - {{- item.text }} - {%- else %} - {{- raise_exception('Unexpected item type in content.') }} - {%- endif %} - {%- endfor %} - {%- elif content is none or content is undefined %} - {{- '' }} - {%- else %} - {{- raise_exception('Unexpected content type.') }} - {%- endif %} -{%- endmacro %} -{%- if not messages %} - {{- raise_exception('No messages provided.') }} -{%- endif %} -{%- if tools and tools is iterable and tools is not mapping %} - {{- '<|im_start|>system\n' }} - {{- "# Tools\n\nYou have access to the following functions:\n\n" }} - {%- for tool in tools %} - {{- "\n" }} - {{- tool | tojson }} - {%- endfor %} - {{- "\n" }} - {{- '\n\nIf you choose to call a function ONLY reply in the following format with NO suffix:\n\n\n\n\nvalue_1\n\n\nThis is the value for the second parameter\nthat can span\nmultiple lines\n\n\n\n\n\nReminder:\n- Function calls MUST follow the specified format: an inner block must be nested within XML tags\n- Required parameters MUST be specified\n- You may provide optional reasoning for your function call in natural language BEFORE the function call, but NOT after\n- If there is no function call available, answer the question like normal with your current knowledge and do not tell the user about function calls\n' }} - {%- if messages[0].role == 'system' %} - {%- set content = render_content(messages[0].content, false, true)|trim %} - {%- if content %} - {{- '\n\n' + content }} - {%- endif %} - {%- endif %} - {{- '<|im_end|>\n' }} -{%- else %} - {%- if messages[0].role == 'system' %} - {%- set content = render_content(messages[0].content, false, true)|trim %} - {{- '<|im_start|>system\n' + 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" %} - {%- set content = render_content(message.content, false)|trim %} - {%- if not(content.startswith('') and content.endswith('')) %} - {%- set ns.multi_step_tool = false %} - {%- set ns.last_query_index = index %} - {%- endif %} - {%- endif %} -{%- endfor %} -{%- if ns.multi_step_tool %} - {{- raise_exception('No user query found in messages.') }} -{%- endif %} -{%- for message in messages %} - {%- set content = render_content(message.content, true)|trim %} - {%- if message.role == "system" %} - {%- if not loop.first %} - {{- raise_exception('System message must be at the beginning.') }} - {%- endif %} - {%- elif message.role == "user" %} - {{- '<|im_start|>' + message.role + '\n' + content + '<|im_end|>' + '\n' }} - {%- elif message.role == "assistant" %} - {%- set reasoning_content = '' %} - {%- if message.reasoning_content is string %} - {%- set reasoning_content = message.reasoning_content %} - {%- else %} - {%- if '' in content %} - {%- set reasoning_content = content.split('')[0].rstrip('\n').split('')[-1].lstrip('\n') %} - {%- set content = content.split('')[-1].lstrip('\n') %} - {%- endif %} - {%- endif %} - {%- set reasoning_content = reasoning_content|trim %} - {%- if loop.index0 > ns.last_query_index %} - {{- '<|im_start|>' + message.role + '\n\n' + reasoning_content + '\n\n\n' + content }} - {%- else %} - {{- '<|im_start|>' + message.role + '\n' + content }} - {%- endif %} - {%- if message.tool_calls and message.tool_calls is iterable and message.tool_calls is not mapping %} - {%- for tool_call in message.tool_calls %} - {%- if tool_call.function is defined %} - {%- set tool_call = tool_call.function %} - {%- endif %} - {%- if loop.first %} - {%- if content|trim %} - {{- '\n\n\n\n' }} - {%- else %} - {{- '\n\n' }} - {%- endif %} - {%- else %} - {{- '\n\n\n' }} - {%- endif %} - {%- if tool_call.arguments is defined %} - {%- for args_name, args_value in tool_call.arguments|items %} - {{- '\n' }} - {%- set args_value = args_value | tojson | safe if args_value is mapping or (args_value is sequence and args_value is not string) else args_value | string %} - {{- args_value }} - {{- '\n\n' }} - {%- endfor %} - {%- endif %} - {{- '\n' }} - {%- endfor %} - {%- endif %} - {{- '<|im_end|>\n' }} - {%- elif message.role == "tool" %} - {%- if loop.previtem and loop.previtem.role != "tool" %} - {{- '<|im_start|>user' }} - {%- endif %} - {{- '\n\n' }} - {{- content }} - {{- '\n' }} - {%- if not loop.last and loop.nextitem.role != "tool" %} - {{- '<|im_end|>\n' }} - {%- elif loop.last %} - {{- '<|im_end|>\n' }} - {%- endif %} - {%- else %} - {{- raise_exception('Unexpected message role.') }} - {%- endif %} -{%- endfor %} -{%- if add_generation_prompt %} - {{- '<|im_start|>assistant\n' }} - {%- if enable_thinking is defined and enable_thinking is true %} - {{- '\n' }} - {%- else %} - {{- '\n\n\n\n' }} - {%- endif %} -{%- endif %} \ No newline at end of file diff --git a/substitutivity_original_Estonian/Qwen3.5-2B-Base_substitutivity_splits_original_features_train_substitutivity_splits_original_features_test1/checkpoint-2075/tokenizer.json b/substitutivity_original_Estonian/Qwen3.5-2B-Base_substitutivity_splits_original_features_train_substitutivity_splits_original_features_test1/checkpoint-2075/tokenizer.json deleted file mode 100644 index 67741b04f23bfdb46501f748ce27865ec82eccfb..0000000000000000000000000000000000000000 --- a/substitutivity_original_Estonian/Qwen3.5-2B-Base_substitutivity_splits_original_features_train_substitutivity_splits_original_features_test1/checkpoint-2075/tokenizer.json +++ /dev/null @@ -1,3 +0,0 @@ -version https://git-lfs.github.com/spec/v1 -oid sha256:87a7830d63fcf43bf241c3c5242e96e62dd3fdc29224ca26fed8ea333db72de4 -size 19989343 diff --git a/substitutivity_original_Estonian/Qwen3.5-2B-Base_substitutivity_splits_original_features_train_substitutivity_splits_original_features_test1/checkpoint-2075/tokenizer_config.json b/substitutivity_original_Estonian/Qwen3.5-2B-Base_substitutivity_splits_original_features_train_substitutivity_splits_original_features_test1/checkpoint-2075/tokenizer_config.json deleted file mode 100644 index b4a37b2a6fd3ab3317cd7bac72855be1a843b2bb..0000000000000000000000000000000000000000 --- a/substitutivity_original_Estonian/Qwen3.5-2B-Base_substitutivity_splits_original_features_train_substitutivity_splits_original_features_test1/checkpoint-2075/tokenizer_config.json +++ /dev/null @@ -1,31 +0,0 @@ -{ - "add_prefix_space": false, - "audio_bos_token": "<|audio_start|>", - "audio_eos_token": "<|audio_end|>", - "audio_token": "<|audio_pad|>", - "backend": "tokenizers", - "bos_token": null, - "clean_up_tokenization_spaces": false, - "eos_token": "<|endoftext|>", - "errors": "replace", - "image_token": "<|image_pad|>", - "is_local": false, - "model_max_length": 262144, - "model_specific_special_tokens": { - "audio_bos_token": "<|audio_start|>", - "audio_eos_token": "<|audio_end|>", - "audio_token": "<|audio_pad|>", - "image_token": "<|image_pad|>", - "video_token": "<|video_pad|>", - "vision_bos_token": "<|vision_start|>", - "vision_eos_token": "<|vision_end|>" - }, - "pad_token": "<|endoftext|>", - "pretokenize_regex": "(?i:'s|'t|'re|'ve|'m|'ll|'d)|[^\\r\\n\\p{L}\\p{N}]?[\\p{L}\\p{M}]+|\\p{N}| ?[^\\s\\p{L}\\p{M}\\p{N}]+[\\r\\n]*|\\s*[\\r\\n]+|\\s+(?!\\S)|\\s+", - "split_special_tokens": false, - "tokenizer_class": "TokenizersBackend", - "unk_token": null, - "video_token": "<|video_pad|>", - "vision_bos_token": "<|vision_start|>", - "vision_eos_token": "<|vision_end|>" -} diff --git a/substitutivity_original_Estonian/Qwen3.5-2B-Base_substitutivity_splits_original_features_train_substitutivity_splits_original_features_test1/checkpoint-2075/trainer_state.json b/substitutivity_original_Estonian/Qwen3.5-2B-Base_substitutivity_splits_original_features_train_substitutivity_splits_original_features_test1/checkpoint-2075/trainer_state.json deleted file mode 100644 index 3eb677e71d469992fc51577ec2a17e0bda8c3100..0000000000000000000000000000000000000000 --- a/substitutivity_original_Estonian/Qwen3.5-2B-Base_substitutivity_splits_original_features_train_substitutivity_splits_original_features_test1/checkpoint-2075/trainer_state.json +++ /dev/null @@ -1,554 +0,0 @@ -{ - "best_global_step": null, - "best_metric": null, - "best_model_checkpoint": null, - "epoch": 5.0, - "eval_steps": 500, - "global_step": 2075, - "is_hyper_param_search": false, - "is_local_process_zero": true, - "is_world_process_zero": true, - "log_history": [ - { - "entropy": 2.0473356211185454, - "epoch": 0.12062726176115803, - "grad_norm": 3.0554118156433105, - "learning_rate": 1.5682344549790923e-05, - "loss": 1.910550079345703, - "mean_token_accuracy": 0.6053373458981514, - "num_tokens": 121428.0, - "step": 50 - }, - { - "entropy": 1.0371007251739501, - "epoch": 0.24125452352231605, - "grad_norm": 2.109868049621582, - "learning_rate": 3.168473694753677e-05, - "loss": 0.9585584259033203, - "mean_token_accuracy": 0.7352996730804443, - "num_tokens": 243966.0, - "step": 100 - }, - { - "entropy": 0.9105490958690643, - "epoch": 0.3618817852834741, - "grad_norm": 1.7675701379776, - "learning_rate": 4.7687129345282605e-05, - "loss": 0.8393524932861328, - "mean_token_accuracy": 0.7588736474514007, - "num_tokens": 366649.0, - "step": 150 - }, - { - "entropy": 0.8660378754138947, - "epoch": 0.4825090470446321, - "grad_norm": 1.6502987146377563, - "learning_rate": 6.368952174302845e-05, - "loss": 0.8005844116210937, - "mean_token_accuracy": 0.7663125538825989, - "num_tokens": 488367.0, - "step": 200 - }, - { - "entropy": 0.8439123743772506, - "epoch": 0.6031363088057901, - "grad_norm": 1.538365364074707, - "learning_rate": 7.969191414077429e-05, - "loss": 0.7773619842529297, - "mean_token_accuracy": 0.7718527519702911, - "num_tokens": 603131.0, - "step": 250 - }, - { - "entropy": 0.8223806774616241, - "epoch": 0.7237635705669482, - "grad_norm": 1.1559664011001587, - "learning_rate": 9.569430653852014e-05, - "loss": 0.7587767791748047, - "mean_token_accuracy": 0.7767325991392136, - "num_tokens": 726412.0, - "step": 300 - }, - { - "entropy": 0.8256935775279999, - "epoch": 0.8443908323281062, - "grad_norm": 1.2322715520858765, - "learning_rate": 0.00011169669893626597, - "loss": 0.769886245727539, - "mean_token_accuracy": 0.773840668797493, - "num_tokens": 846516.0, - "step": 350 - }, - { - "entropy": 0.8065843790769577, - "epoch": 0.9650180940892642, - "grad_norm": 0.9908286333084106, - "learning_rate": 0.00012769909133401183, - "loss": 0.7451084899902344, - "mean_token_accuracy": 0.7794792425632476, - "num_tokens": 958978.0, - "step": 400 - }, - { - "epoch": 1.0, - "eval_entropy": 0.791056221790528, - "eval_mean_token_accuracy": 0.7507295300451557, - "eval_not_syn_loss": 0.7860428094863892, - "eval_not_syn_runtime": 53.5431, - "eval_not_syn_samples_per_second": 26.521, - "eval_not_syn_steps_per_second": 3.324, - "eval_num_tokens": 993806.0, - "step": 415 - }, - { - "epoch": 1.0, - "eval_entropy": 0.7424047408478983, - "eval_mean_token_accuracy": 0.8023804037758474, - "eval_num_tokens": 993806.0, - "eval_syn_loss": 0.7273509502410889, - "eval_syn_runtime": 54.8679, - "eval_syn_samples_per_second": 25.88, - "eval_syn_steps_per_second": 3.244, - "step": 415 - }, - { - "entropy": 0.7607260761838971, - "epoch": 1.0844390832328106, - "grad_norm": 1.0653696060180664, - "learning_rate": 0.0001327927019027069, - "loss": 0.7034613037109375, - "mean_token_accuracy": 0.7869195998317063, - "num_tokens": 1080301.0, - "step": 450 - }, - { - "entropy": 0.7370894801616669, - "epoch": 1.2050663449939687, - "grad_norm": 1.763787865638733, - "learning_rate": 0.00013265416569994628, - "loss": 0.6816973114013671, - "mean_token_accuracy": 0.7914507347345352, - "num_tokens": 1201215.0, - "step": 500 - }, - { - "entropy": 0.742829909324646, - "epoch": 1.3256936067551266, - "grad_norm": 0.97663414478302, - "learning_rate": 0.00013239847940553975, - "loss": 0.6867338562011719, - "mean_token_accuracy": 0.790328289270401, - "num_tokens": 1322895.0, - "step": 550 - }, - { - "entropy": 0.7312483406066894, - "epoch": 1.4463208685162847, - "grad_norm": 1.0874731540679932, - "learning_rate": 0.00013202609518972978, - "loss": 0.676800537109375, - "mean_token_accuracy": 0.7921512228250503, - "num_tokens": 1436749.0, - "step": 600 - }, - { - "entropy": 0.7178869074583054, - "epoch": 1.5669481302774426, - "grad_norm": 1.064300298690796, - "learning_rate": 0.0001315376715980265, - "loss": 0.6635584259033203, - "mean_token_accuracy": 0.7944776183366775, - "num_tokens": 1555400.0, - "step": 650 - }, - { - "entropy": 0.7198166918754577, - "epoch": 1.6875753920386007, - "grad_norm": 0.7779315114021301, - "learning_rate": 0.00013093407238659802, - "loss": 0.6660567474365234, - "mean_token_accuracy": 0.7968300533294678, - "num_tokens": 1676892.0, - "step": 700 - }, - { - "entropy": 0.7170061504840851, - "epoch": 1.8082026537997589, - "grad_norm": 1.089854121208191, - "learning_rate": 0.00013021636499475496, - "loss": 0.6632771301269531, - "mean_token_accuracy": 0.7963915091753005, - "num_tokens": 1802689.0, - "step": 750 - }, - { - "entropy": 0.6953610718250275, - "epoch": 1.9288299155609168, - "grad_norm": 0.7996730208396912, - "learning_rate": 0.0001293858186572295, - "loss": 0.6421419525146485, - "mean_token_accuracy": 0.8002577543258667, - "num_tokens": 1920804.0, - "step": 800 - }, - { - "epoch": 2.0, - "eval_entropy": 0.7369416523515508, - "eval_mean_token_accuracy": 0.7872022366925572, - "eval_not_syn_loss": 0.713637113571167, - "eval_not_syn_runtime": 53.2265, - "eval_not_syn_samples_per_second": 26.678, - "eval_not_syn_steps_per_second": 3.344, - "eval_num_tokens": 1987612.0, - "step": 830 - }, - { - "epoch": 2.0, - "eval_entropy": 0.6877032756135705, - "eval_mean_token_accuracy": 0.7997466372639945, - "eval_num_tokens": 1987612.0, - "eval_syn_loss": 0.6614266633987427, - "eval_syn_runtime": 54.5178, - "eval_syn_samples_per_second": 26.047, - "eval_syn_steps_per_second": 3.265, - "step": 830 - }, - { - "entropy": 0.6644553433765065, - "epoch": 2.0482509047044632, - "grad_norm": 0.7562468647956848, - "learning_rate": 0.00012844390215958852, - "loss": 0.6112185287475586, - "mean_token_accuracy": 0.806649933559726, - "num_tokens": 2039660.0, - "step": 850 - }, - { - "entropy": 0.6146580249071121, - "epoch": 2.168878166465621, - "grad_norm": 1.098607063293457, - "learning_rate": 0.00012739228124074913, - "loss": 0.5553749084472657, - "mean_token_accuracy": 0.8187588346004486, - "num_tokens": 2161772.0, - "step": 900 - }, - { - "entropy": 0.6187627577781677, - "epoch": 2.2895054282267795, - "grad_norm": 1.0338040590286255, - "learning_rate": 0.00012623281564719095, - "loss": 0.5596194076538086, - "mean_token_accuracy": 0.8182566928863525, - "num_tokens": 2280833.0, - "step": 950 - }, - { - "entropy": 0.6232365870475769, - "epoch": 2.4101326899879374, - "grad_norm": 1.0010799169540405, - "learning_rate": 0.00012496755584407422, - "loss": 0.5667143249511719, - "mean_token_accuracy": 0.8175061982870102, - "num_tokens": 2398206.0, - "step": 1000 - }, - { - "entropy": 0.6151268655061721, - "epoch": 2.5307599517490953, - "grad_norm": 1.0065034627914429, - "learning_rate": 0.0001235987393890801, - "loss": 0.5598890686035156, - "mean_token_accuracy": 0.8193449324369431, - "num_tokens": 2518888.0, - "step": 1050 - }, - { - "entropy": 0.6173626267910004, - "epoch": 2.651387213510253, - "grad_norm": 0.7958805561065674, - "learning_rate": 0.00012212878697538592, - "loss": 0.5649292755126953, - "mean_token_accuracy": 0.8172902607917786, - "num_tokens": 2639763.0, - "step": 1100 - }, - { - "entropy": 0.6264930808544159, - "epoch": 2.772014475271411, - "grad_norm": 0.7718372344970703, - "learning_rate": 0.00012056029815077323, - "loss": 0.5716455078125, - "mean_token_accuracy": 0.8170186239480972, - "num_tokens": 2755540.0, - "step": 1150 - }, - { - "entropy": 0.6203387409448624, - "epoch": 2.8926417370325694, - "grad_norm": 0.7914765477180481, - "learning_rate": 0.000118896046720439, - "loss": 0.5653961944580078, - "mean_token_accuracy": 0.8170280295610428, - "num_tokens": 2874118.0, - "step": 1200 - }, - { - "epoch": 3.0, - "eval_entropy": 0.6751241854737314, - "eval_mean_token_accuracy": 0.792226163858778, - "eval_not_syn_loss": 0.6902772784233093, - "eval_not_syn_runtime": 53.2557, - "eval_not_syn_samples_per_second": 26.664, - "eval_not_syn_steps_per_second": 3.342, - "eval_num_tokens": 2981418.0, - "step": 1245 - }, - { - "epoch": 3.0, - "eval_entropy": 0.6308753868837035, - "eval_mean_token_accuracy": 0.8131611759073278, - "eval_num_tokens": 2981418.0, - "eval_syn_loss": 0.6440441012382507, - "eval_syn_runtime": 54.7064, - "eval_syn_samples_per_second": 25.957, - "eval_syn_steps_per_second": 3.254, - "step": 1245 - }, - { - "entropy": 0.6077835421369533, - "epoch": 3.012062726176116, - "grad_norm": 1.124720573425293, - "learning_rate": 0.00011713897584164003, - "loss": 0.5557136917114258, - "mean_token_accuracy": 0.8203362726201915, - "num_tokens": 2994602.0, - "step": 1250 - }, - { - "entropy": 0.507642827630043, - "epoch": 3.132689987937274, - "grad_norm": 0.9039916396141052, - "learning_rate": 0.00011529219281884579, - "loss": 0.4467372512817383, - "mean_token_accuracy": 0.8464260965585708, - "num_tokens": 3121094.0, - "step": 1300 - }, - { - "entropy": 0.5137191081047058, - "epoch": 3.2533172496984317, - "grad_norm": 0.8580691814422607, - "learning_rate": 0.00011335896360860367, - "loss": 0.4538694763183594, - "mean_token_accuracy": 0.8448719161748887, - "num_tokens": 3248038.0, - "step": 1350 - }, - { - "entropy": 0.5293738862872124, - "epoch": 3.37394451145959, - "grad_norm": 1.0177273750305176, - "learning_rate": 0.00011134270704383492, - "loss": 0.4639348220825195, - "mean_token_accuracy": 0.8428975892066956, - "num_tokens": 3363487.0, - "step": 1400 - }, - { - "entropy": 0.5218872320652008, - "epoch": 3.494571773220748, - "grad_norm": 1.0530673265457153, - "learning_rate": 0.00010924698878777538, - "loss": 0.4584346389770508, - "mean_token_accuracy": 0.8432288640737533, - "num_tokens": 3480513.0, - "step": 1450 - }, - { - "entropy": 0.5320372316241264, - "epoch": 3.615199034981906, - "grad_norm": 1.0291194915771484, - "learning_rate": 0.0001070755150282529, - "loss": 0.4711898422241211, - "mean_token_accuracy": 0.842066935300827, - "num_tokens": 3597183.0, - "step": 1500 - }, - { - "entropy": 0.5315000694990158, - "epoch": 3.7358262967430638, - "grad_norm": 0.9788838624954224, - "learning_rate": 0.00010483212592345332, - "loss": 0.4729101943969727, - "mean_token_accuracy": 0.8404586082696914, - "num_tokens": 3714173.0, - "step": 1550 - }, - { - "entropy": 0.5331643190979958, - "epoch": 3.856453558504222, - "grad_norm": 1.1248903274536133, - "learning_rate": 0.00010252078881076532, - "loss": 0.47405517578125, - "mean_token_accuracy": 0.8397526168823242, - "num_tokens": 3832652.0, - "step": 1600 - }, - { - "entropy": 0.5343865439295769, - "epoch": 3.97708082026538, - "grad_norm": 0.9008272886276245, - "learning_rate": 0.00010014559119071452, - "loss": 0.47172779083251953, - "mean_token_accuracy": 0.8404951041936874, - "num_tokens": 3955167.0, - "step": 1650 - }, - { - "epoch": 4.0, - "eval_entropy": 0.5736920103598177, - "eval_mean_token_accuracy": 0.7981750027517255, - "eval_not_syn_loss": 0.707631528377533, - "eval_not_syn_runtime": 52.2858, - "eval_not_syn_samples_per_second": 27.158, - "eval_not_syn_steps_per_second": 3.404, - "eval_num_tokens": 3975224.0, - "step": 1660 - }, - { - "epoch": 4.0, - "eval_entropy": 0.5363499959198277, - "eval_mean_token_accuracy": 0.8063885539435269, - "eval_num_tokens": 3975224.0, - "eval_syn_loss": 0.6610181331634521, - "eval_syn_runtime": 54.2167, - "eval_syn_samples_per_second": 26.191, - "eval_syn_steps_per_second": 3.283, - "step": 1660 - }, - { - "entropy": 0.4512602482179199, - "epoch": 4.0965018094089265, - "grad_norm": 0.8079906105995178, - "learning_rate": 9.771073349839416e-05, - "loss": 0.3778452682495117, - "mean_token_accuracy": 0.8684436221315404, - "num_tokens": 4072361.0, - "step": 1700 - }, - { - "entropy": 0.4228873434662819, - "epoch": 4.217129071170085, - "grad_norm": 1.1206746101379395, - "learning_rate": 9.52205216751758e-05, - "loss": 0.34671127319335937, - "mean_token_accuracy": 0.8763950079679489, - "num_tokens": 4194978.0, - "step": 1750 - }, - { - "entropy": 0.42808119297027586, - "epoch": 4.337756332931242, - "grad_norm": 1.1268125772476196, - "learning_rate": 9.26793595538368e-05, - "loss": 0.3484600830078125, - "mean_token_accuracy": 0.8753323918581009, - "num_tokens": 4315548.0, - "step": 1800 - }, - { - "entropy": 0.42506767451763156, - "epoch": 4.458383594692401, - "grad_norm": 1.1051807403564453, - "learning_rate": 9.009174107057088e-05, - "loss": 0.35434780120849607, - "mean_token_accuracy": 0.8734411865472793, - "num_tokens": 4431469.0, - "step": 1850 - }, - { - "entropy": 0.43314908266067503, - "epoch": 4.579010856453559, - "grad_norm": 0.9782640337944031, - "learning_rate": 8.746224231765484e-05, - "loss": 0.36225860595703124, - "mean_token_accuracy": 0.8715140378475189, - "num_tokens": 4551803.0, - "step": 1900 - }, - { - "entropy": 0.42984022706747055, - "epoch": 4.699638118214716, - "grad_norm": 0.9772566556930542, - "learning_rate": 8.479551345082568e-05, - "loss": 0.356609001159668, - "mean_token_accuracy": 0.8717925781011582, - "num_tokens": 4674468.0, - "step": 1950 - }, - { - "entropy": 0.4474307084083557, - "epoch": 4.820265379975875, - "grad_norm": 0.9547265768051147, - "learning_rate": 8.209627046567966e-05, - "loss": 0.3654902267456055, - "mean_token_accuracy": 0.8724055212736129, - "num_tokens": 4788027.0, - "step": 2000 - }, - { - "entropy": 0.4377285006642342, - "epoch": 4.940892641737032, - "grad_norm": 1.0390652418136597, - "learning_rate": 7.936928685763635e-05, - "loss": 0.3645761871337891, - "mean_token_accuracy": 0.8707609891891479, - "num_tokens": 4907561.0, - "step": 2050 - }, - { - "epoch": 5.0, - "eval_entropy": 0.5173388388049737, - "eval_mean_token_accuracy": 0.7918405432379647, - "eval_not_syn_loss": 0.7532826662063599, - "eval_not_syn_runtime": 53.0766, - "eval_not_syn_samples_per_second": 26.754, - "eval_not_syn_steps_per_second": 3.354, - "eval_num_tokens": 4969030.0, - "step": 2075 - }, - { - "epoch": 5.0, - "eval_entropy": 0.4831916272975086, - "eval_mean_token_accuracy": 0.8204237202580056, - "eval_num_tokens": 4969030.0, - "eval_syn_loss": 0.6971890330314636, - "eval_syn_runtime": 54.8755, - "eval_syn_samples_per_second": 25.877, - "eval_syn_steps_per_second": 3.244, - "step": 2075 - } - ], - "logging_steps": 50, - "max_steps": 4150, - "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": 1.0572192305263104e+17, - "train_batch_size": 8, - "trial_name": null, - "trial_params": null -} diff --git a/substitutivity_original_Estonian/Qwen3.5-2B-Base_substitutivity_splits_original_features_train_substitutivity_splits_original_features_test1/checkpoint-2075/training_args.bin b/substitutivity_original_Estonian/Qwen3.5-2B-Base_substitutivity_splits_original_features_train_substitutivity_splits_original_features_test1/checkpoint-2075/training_args.bin deleted file mode 100644 index 48901efb30506830185bfaafecbe236e7095a006..0000000000000000000000000000000000000000 --- a/substitutivity_original_Estonian/Qwen3.5-2B-Base_substitutivity_splits_original_features_train_substitutivity_splits_original_features_test1/checkpoint-2075/training_args.bin +++ /dev/null @@ -1,3 +0,0 @@ -version https://git-lfs.github.com/spec/v1 -oid sha256:770fc15ebc27a54d780af597d32c57ca14b3b528f14788d142bdfce09b555bef -size 6033 diff --git a/substitutivity_original_Estonian/Qwen3.5-2B-Base_substitutivity_splits_original_features_train_substitutivity_splits_original_features_test1/checkpoint-2490/README.md b/substitutivity_original_Estonian/Qwen3.5-2B-Base_substitutivity_splits_original_features_train_substitutivity_splits_original_features_test1/checkpoint-2490/README.md deleted file mode 100644 index ece204d005f7bf161d0e448efdd56fd297b7c8a2..0000000000000000000000000000000000000000 --- a/substitutivity_original_Estonian/Qwen3.5-2B-Base_substitutivity_splits_original_features_train_substitutivity_splits_original_features_test1/checkpoint-2490/README.md +++ /dev/null @@ -1,209 +0,0 @@ ---- -base_model: Qwen/Qwen3.5-2B-Base -library_name: peft -pipeline_tag: text-generation -tags: -- base_model:adapter:Qwen/Qwen3.5-2B-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/substitutivity_original_Estonian/Qwen3.5-2B-Base_substitutivity_splits_original_features_train_substitutivity_splits_original_features_test1/checkpoint-2490/adapter_config.json b/substitutivity_original_Estonian/Qwen3.5-2B-Base_substitutivity_splits_original_features_train_substitutivity_splits_original_features_test1/checkpoint-2490/adapter_config.json deleted file mode 100644 index a64145ce23adcbdbe246679087a13f4cb63ff664..0000000000000000000000000000000000000000 --- a/substitutivity_original_Estonian/Qwen3.5-2B-Base_substitutivity_splits_original_features_train_substitutivity_splits_original_features_test1/checkpoint-2490/adapter_config.json +++ /dev/null @@ -1,46 +0,0 @@ -{ - "alora_invocation_tokens": null, - "alpha_pattern": {}, - "arrow_config": null, - "auto_mapping": null, - "base_model_name_or_path": "Qwen/Qwen3.5-2B-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": 256, - "lora_bias": false, - "lora_dropout": 0.027050353915701653, - "megatron_config": null, - "megatron_core": "megatron.core", - "modules_to_save": null, - "peft_type": "LORA", - "peft_version": "0.18.1", - "qalora_group_size": 16, - "r": 128, - "rank_pattern": {}, - "revision": null, - "target_modules": [ - "o_proj", - "q_proj", - "k_proj", - "down_proj", - "up_proj", - "v_proj", - "gate_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/substitutivity_original_Estonian/Qwen3.5-2B-Base_substitutivity_splits_original_features_train_substitutivity_splits_original_features_test1/checkpoint-2490/adapter_model.safetensors b/substitutivity_original_Estonian/Qwen3.5-2B-Base_substitutivity_splits_original_features_train_substitutivity_splits_original_features_test1/checkpoint-2490/adapter_model.safetensors deleted file mode 100644 index 2012874dedad6ce282c3be7f4b8c0e27021ba186..0000000000000000000000000000000000000000 --- a/substitutivity_original_Estonian/Qwen3.5-2B-Base_substitutivity_splits_original_features_train_substitutivity_splits_original_features_test1/checkpoint-2490/adapter_model.safetensors +++ /dev/null @@ -1,3 +0,0 @@ -version https://git-lfs.github.com/spec/v1 -oid sha256:275a7bde3ab6e98ce13f2b9e139f10d35926ca663e5d15c5533c4b1e5cf007af -size 349201608 diff --git a/substitutivity_original_Estonian/Qwen3.5-2B-Base_substitutivity_splits_original_features_train_substitutivity_splits_original_features_test1/checkpoint-2490/chat_template.jinja b/substitutivity_original_Estonian/Qwen3.5-2B-Base_substitutivity_splits_original_features_train_substitutivity_splits_original_features_test1/checkpoint-2490/chat_template.jinja deleted file mode 100644 index 0ef09f214eaa6d9bca297988afc1454b5827b2c7..0000000000000000000000000000000000000000 --- a/substitutivity_original_Estonian/Qwen3.5-2B-Base_substitutivity_splits_original_features_train_substitutivity_splits_original_features_test1/checkpoint-2490/chat_template.jinja +++ /dev/null @@ -1,154 +0,0 @@ -{%- set image_count = namespace(value=0) %} -{%- set video_count = namespace(value=0) %} -{%- macro render_content(content, do_vision_count, is_system_content=false) %} - {%- if content is string %} - {{- content }} - {%- elif content is iterable and content is not mapping %} - {%- for item in content %} - {%- if 'image' in item or 'image_url' in item or item.type == 'image' %} - {%- if is_system_content %} - {{- raise_exception('System message cannot contain images.') }} - {%- endif %} - {%- if do_vision_count %} - {%- set image_count.value = image_count.value + 1 %} - {%- endif %} - {%- if add_vision_id %} - {{- 'Picture ' ~ image_count.value ~ ': ' }} - {%- endif %} - {{- '<|vision_start|><|image_pad|><|vision_end|>' }} - {%- elif 'video' in item or item.type == 'video' %} - {%- if is_system_content %} - {{- raise_exception('System message cannot contain videos.') }} - {%- endif %} - {%- if do_vision_count %} - {%- set video_count.value = video_count.value + 1 %} - {%- endif %} - {%- if add_vision_id %} - {{- 'Video ' ~ video_count.value ~ ': ' }} - {%- endif %} - {{- '<|vision_start|><|video_pad|><|vision_end|>' }} - {%- elif 'text' in item %} - {{- item.text }} - {%- else %} - {{- raise_exception('Unexpected item type in content.') }} - {%- endif %} - {%- endfor %} - {%- elif content is none or content is undefined %} - {{- '' }} - {%- else %} - {{- raise_exception('Unexpected content type.') }} - {%- endif %} -{%- endmacro %} -{%- if not messages %} - {{- raise_exception('No messages provided.') }} -{%- endif %} -{%- if tools and tools is iterable and tools is not mapping %} - {{- '<|im_start|>system\n' }} - {{- "# Tools\n\nYou have access to the following functions:\n\n" }} - {%- for tool in tools %} - {{- "\n" }} - {{- tool | tojson }} - {%- endfor %} - {{- "\n" }} - {{- '\n\nIf you choose to call a function ONLY reply in the following format with NO suffix:\n\n\n\n\nvalue_1\n\n\nThis is the value for the second parameter\nthat can span\nmultiple lines\n\n\n\n\n\nReminder:\n- Function calls MUST follow the specified format: an inner block must be nested within XML tags\n- Required parameters MUST be specified\n- You may provide optional reasoning for your function call in natural language BEFORE the function call, but NOT after\n- If there is no function call available, answer the question like normal with your current knowledge and do not tell the user about function calls\n' }} - {%- if messages[0].role == 'system' %} - {%- set content = render_content(messages[0].content, false, true)|trim %} - {%- if content %} - {{- '\n\n' + content }} - {%- endif %} - {%- endif %} - {{- '<|im_end|>\n' }} -{%- else %} - {%- if messages[0].role == 'system' %} - {%- set content = render_content(messages[0].content, false, true)|trim %} - {{- '<|im_start|>system\n' + 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" %} - {%- set content = render_content(message.content, false)|trim %} - {%- if not(content.startswith('') and content.endswith('')) %} - {%- set ns.multi_step_tool = false %} - {%- set ns.last_query_index = index %} - {%- endif %} - {%- endif %} -{%- endfor %} -{%- if ns.multi_step_tool %} - {{- raise_exception('No user query found in messages.') }} -{%- endif %} -{%- for message in messages %} - {%- set content = render_content(message.content, true)|trim %} - {%- if message.role == "system" %} - {%- if not loop.first %} - {{- raise_exception('System message must be at the beginning.') }} - {%- endif %} - {%- elif message.role == "user" %} - {{- '<|im_start|>' + message.role + '\n' + content + '<|im_end|>' + '\n' }} - {%- elif message.role == "assistant" %} - {%- set reasoning_content = '' %} - {%- if message.reasoning_content is string %} - {%- set reasoning_content = message.reasoning_content %} - {%- else %} - {%- if '' in content %} - {%- set reasoning_content = content.split('')[0].rstrip('\n').split('')[-1].lstrip('\n') %} - {%- set content = content.split('')[-1].lstrip('\n') %} - {%- endif %} - {%- endif %} - {%- set reasoning_content = reasoning_content|trim %} - {%- if loop.index0 > ns.last_query_index %} - {{- '<|im_start|>' + message.role + '\n\n' + reasoning_content + '\n\n\n' + content }} - {%- else %} - {{- '<|im_start|>' + message.role + '\n' + content }} - {%- endif %} - {%- if message.tool_calls and message.tool_calls is iterable and message.tool_calls is not mapping %} - {%- for tool_call in message.tool_calls %} - {%- if tool_call.function is defined %} - {%- set tool_call = tool_call.function %} - {%- endif %} - {%- if loop.first %} - {%- if content|trim %} - {{- '\n\n\n\n' }} - {%- else %} - {{- '\n\n' }} - {%- endif %} - {%- else %} - {{- '\n\n\n' }} - {%- endif %} - {%- if tool_call.arguments is defined %} - {%- for args_name, args_value in tool_call.arguments|items %} - {{- '\n' }} - {%- set args_value = args_value | tojson | safe if args_value is mapping or (args_value is sequence and args_value is not string) else args_value | string %} - {{- args_value }} - {{- '\n\n' }} - {%- endfor %} - {%- endif %} - {{- '\n' }} - {%- endfor %} - {%- endif %} - {{- '<|im_end|>\n' }} - {%- elif message.role == "tool" %} - {%- if loop.previtem and loop.previtem.role != "tool" %} - {{- '<|im_start|>user' }} - {%- endif %} - {{- '\n\n' }} - {{- content }} - {{- '\n' }} - {%- if not loop.last and loop.nextitem.role != "tool" %} - {{- '<|im_end|>\n' }} - {%- elif loop.last %} - {{- '<|im_end|>\n' }} - {%- endif %} - {%- else %} - {{- raise_exception('Unexpected message role.') }} - {%- endif %} -{%- endfor %} -{%- if add_generation_prompt %} - {{- '<|im_start|>assistant\n' }} - {%- if enable_thinking is defined and enable_thinking is true %} - {{- '\n' }} - {%- else %} - {{- '\n\n\n\n' }} - {%- endif %} -{%- endif %} \ No newline at end of file diff --git a/substitutivity_original_Estonian/Qwen3.5-2B-Base_substitutivity_splits_original_features_train_substitutivity_splits_original_features_test1/checkpoint-2490/tokenizer.json b/substitutivity_original_Estonian/Qwen3.5-2B-Base_substitutivity_splits_original_features_train_substitutivity_splits_original_features_test1/checkpoint-2490/tokenizer.json deleted file mode 100644 index 67741b04f23bfdb46501f748ce27865ec82eccfb..0000000000000000000000000000000000000000 --- a/substitutivity_original_Estonian/Qwen3.5-2B-Base_substitutivity_splits_original_features_train_substitutivity_splits_original_features_test1/checkpoint-2490/tokenizer.json +++ /dev/null @@ -1,3 +0,0 @@ -version https://git-lfs.github.com/spec/v1 -oid sha256:87a7830d63fcf43bf241c3c5242e96e62dd3fdc29224ca26fed8ea333db72de4 -size 19989343 diff --git a/substitutivity_original_Estonian/Qwen3.5-2B-Base_substitutivity_splits_original_features_train_substitutivity_splits_original_features_test1/checkpoint-2490/tokenizer_config.json b/substitutivity_original_Estonian/Qwen3.5-2B-Base_substitutivity_splits_original_features_train_substitutivity_splits_original_features_test1/checkpoint-2490/tokenizer_config.json deleted file mode 100644 index b4a37b2a6fd3ab3317cd7bac72855be1a843b2bb..0000000000000000000000000000000000000000 --- a/substitutivity_original_Estonian/Qwen3.5-2B-Base_substitutivity_splits_original_features_train_substitutivity_splits_original_features_test1/checkpoint-2490/tokenizer_config.json +++ /dev/null @@ -1,31 +0,0 @@ -{ - "add_prefix_space": false, - "audio_bos_token": "<|audio_start|>", - "audio_eos_token": "<|audio_end|>", - "audio_token": "<|audio_pad|>", - "backend": "tokenizers", - "bos_token": null, - "clean_up_tokenization_spaces": false, - "eos_token": "<|endoftext|>", - "errors": "replace", - "image_token": "<|image_pad|>", - "is_local": false, - "model_max_length": 262144, - "model_specific_special_tokens": { - "audio_bos_token": "<|audio_start|>", - "audio_eos_token": "<|audio_end|>", - "audio_token": "<|audio_pad|>", - "image_token": "<|image_pad|>", - "video_token": "<|video_pad|>", - "vision_bos_token": "<|vision_start|>", - "vision_eos_token": "<|vision_end|>" - }, - "pad_token": "<|endoftext|>", - "pretokenize_regex": "(?i:'s|'t|'re|'ve|'m|'ll|'d)|[^\\r\\n\\p{L}\\p{N}]?[\\p{L}\\p{M}]+|\\p{N}| ?[^\\s\\p{L}\\p{M}\\p{N}]+[\\r\\n]*|\\s*[\\r\\n]+|\\s+(?!\\S)|\\s+", - "split_special_tokens": false, - "tokenizer_class": "TokenizersBackend", - "unk_token": null, - "video_token": "<|video_pad|>", - "vision_bos_token": "<|vision_start|>", - "vision_eos_token": "<|vision_end|>" -} diff --git a/substitutivity_original_Estonian/Qwen3.5-2B-Base_substitutivity_splits_original_features_train_substitutivity_splits_original_features_test1/checkpoint-2490/trainer_state.json b/substitutivity_original_Estonian/Qwen3.5-2B-Base_substitutivity_splits_original_features_train_substitutivity_splits_original_features_test1/checkpoint-2490/trainer_state.json deleted file mode 100644 index 665a062e057efd0b33ab1e052b36c6e7d33d2520..0000000000000000000000000000000000000000 --- a/substitutivity_original_Estonian/Qwen3.5-2B-Base_substitutivity_splits_original_features_train_substitutivity_splits_original_features_test1/checkpoint-2490/trainer_state.json +++ /dev/null @@ -1,656 +0,0 @@ -{ - "best_global_step": null, - "best_metric": null, - "best_model_checkpoint": null, - "epoch": 6.0, - "eval_steps": 500, - "global_step": 2490, - "is_hyper_param_search": false, - "is_local_process_zero": true, - "is_world_process_zero": true, - "log_history": [ - { - "entropy": 2.0473356211185454, - "epoch": 0.12062726176115803, - "grad_norm": 3.0554118156433105, - "learning_rate": 1.5682344549790923e-05, - "loss": 1.910550079345703, - "mean_token_accuracy": 0.6053373458981514, - "num_tokens": 121428.0, - "step": 50 - }, - { - "entropy": 1.0371007251739501, - "epoch": 0.24125452352231605, - "grad_norm": 2.109868049621582, - "learning_rate": 3.168473694753677e-05, - "loss": 0.9585584259033203, - "mean_token_accuracy": 0.7352996730804443, - "num_tokens": 243966.0, - "step": 100 - }, - { - "entropy": 0.9105490958690643, - "epoch": 0.3618817852834741, - "grad_norm": 1.7675701379776, - "learning_rate": 4.7687129345282605e-05, - "loss": 0.8393524932861328, - "mean_token_accuracy": 0.7588736474514007, - "num_tokens": 366649.0, - "step": 150 - }, - { - "entropy": 0.8660378754138947, - "epoch": 0.4825090470446321, - "grad_norm": 1.6502987146377563, - "learning_rate": 6.368952174302845e-05, - "loss": 0.8005844116210937, - "mean_token_accuracy": 0.7663125538825989, - "num_tokens": 488367.0, - "step": 200 - }, - { - "entropy": 0.8439123743772506, - "epoch": 0.6031363088057901, - "grad_norm": 1.538365364074707, - "learning_rate": 7.969191414077429e-05, - "loss": 0.7773619842529297, - "mean_token_accuracy": 0.7718527519702911, - "num_tokens": 603131.0, - "step": 250 - }, - { - "entropy": 0.8223806774616241, - "epoch": 0.7237635705669482, - "grad_norm": 1.1559664011001587, - "learning_rate": 9.569430653852014e-05, - "loss": 0.7587767791748047, - "mean_token_accuracy": 0.7767325991392136, - "num_tokens": 726412.0, - "step": 300 - }, - { - "entropy": 0.8256935775279999, - "epoch": 0.8443908323281062, - "grad_norm": 1.2322715520858765, - "learning_rate": 0.00011169669893626597, - "loss": 0.769886245727539, - "mean_token_accuracy": 0.773840668797493, - "num_tokens": 846516.0, - "step": 350 - }, - { - "entropy": 0.8065843790769577, - "epoch": 0.9650180940892642, - "grad_norm": 0.9908286333084106, - "learning_rate": 0.00012769909133401183, - "loss": 0.7451084899902344, - "mean_token_accuracy": 0.7794792425632476, - "num_tokens": 958978.0, - "step": 400 - }, - { - "epoch": 1.0, - "eval_entropy": 0.791056221790528, - "eval_mean_token_accuracy": 0.7507295300451557, - "eval_not_syn_loss": 0.7860428094863892, - "eval_not_syn_runtime": 53.5431, - "eval_not_syn_samples_per_second": 26.521, - "eval_not_syn_steps_per_second": 3.324, - "eval_num_tokens": 993806.0, - "step": 415 - }, - { - "epoch": 1.0, - "eval_entropy": 0.7424047408478983, - "eval_mean_token_accuracy": 0.8023804037758474, - "eval_num_tokens": 993806.0, - "eval_syn_loss": 0.7273509502410889, - "eval_syn_runtime": 54.8679, - "eval_syn_samples_per_second": 25.88, - "eval_syn_steps_per_second": 3.244, - "step": 415 - }, - { - "entropy": 0.7607260761838971, - "epoch": 1.0844390832328106, - "grad_norm": 1.0653696060180664, - "learning_rate": 0.0001327927019027069, - "loss": 0.7034613037109375, - "mean_token_accuracy": 0.7869195998317063, - "num_tokens": 1080301.0, - "step": 450 - }, - { - "entropy": 0.7370894801616669, - "epoch": 1.2050663449939687, - "grad_norm": 1.763787865638733, - "learning_rate": 0.00013265416569994628, - "loss": 0.6816973114013671, - "mean_token_accuracy": 0.7914507347345352, - "num_tokens": 1201215.0, - "step": 500 - }, - { - "entropy": 0.742829909324646, - "epoch": 1.3256936067551266, - "grad_norm": 0.97663414478302, - "learning_rate": 0.00013239847940553975, - "loss": 0.6867338562011719, - "mean_token_accuracy": 0.790328289270401, - "num_tokens": 1322895.0, - "step": 550 - }, - { - "entropy": 0.7312483406066894, - "epoch": 1.4463208685162847, - "grad_norm": 1.0874731540679932, - "learning_rate": 0.00013202609518972978, - "loss": 0.676800537109375, - "mean_token_accuracy": 0.7921512228250503, - "num_tokens": 1436749.0, - "step": 600 - }, - { - "entropy": 0.7178869074583054, - "epoch": 1.5669481302774426, - "grad_norm": 1.064300298690796, - "learning_rate": 0.0001315376715980265, - "loss": 0.6635584259033203, - "mean_token_accuracy": 0.7944776183366775, - "num_tokens": 1555400.0, - "step": 650 - }, - { - "entropy": 0.7198166918754577, - "epoch": 1.6875753920386007, - "grad_norm": 0.7779315114021301, - "learning_rate": 0.00013093407238659802, - "loss": 0.6660567474365234, - "mean_token_accuracy": 0.7968300533294678, - "num_tokens": 1676892.0, - "step": 700 - }, - { - "entropy": 0.7170061504840851, - "epoch": 1.8082026537997589, - "grad_norm": 1.089854121208191, - "learning_rate": 0.00013021636499475496, - "loss": 0.6632771301269531, - "mean_token_accuracy": 0.7963915091753005, - "num_tokens": 1802689.0, - "step": 750 - }, - { - "entropy": 0.6953610718250275, - "epoch": 1.9288299155609168, - "grad_norm": 0.7996730208396912, - "learning_rate": 0.0001293858186572295, - "loss": 0.6421419525146485, - "mean_token_accuracy": 0.8002577543258667, - "num_tokens": 1920804.0, - "step": 800 - }, - { - "epoch": 2.0, - "eval_entropy": 0.7369416523515508, - "eval_mean_token_accuracy": 0.7872022366925572, - "eval_not_syn_loss": 0.713637113571167, - "eval_not_syn_runtime": 53.2265, - "eval_not_syn_samples_per_second": 26.678, - "eval_not_syn_steps_per_second": 3.344, - "eval_num_tokens": 1987612.0, - "step": 830 - }, - { - "epoch": 2.0, - "eval_entropy": 0.6877032756135705, - "eval_mean_token_accuracy": 0.7997466372639945, - "eval_num_tokens": 1987612.0, - "eval_syn_loss": 0.6614266633987427, - "eval_syn_runtime": 54.5178, - "eval_syn_samples_per_second": 26.047, - "eval_syn_steps_per_second": 3.265, - "step": 830 - }, - { - "entropy": 0.6644553433765065, - "epoch": 2.0482509047044632, - "grad_norm": 0.7562468647956848, - "learning_rate": 0.00012844390215958852, - "loss": 0.6112185287475586, - "mean_token_accuracy": 0.806649933559726, - "num_tokens": 2039660.0, - "step": 850 - }, - { - "entropy": 0.6146580249071121, - "epoch": 2.168878166465621, - "grad_norm": 1.098607063293457, - "learning_rate": 0.00012739228124074913, - "loss": 0.5553749084472657, - "mean_token_accuracy": 0.8187588346004486, - "num_tokens": 2161772.0, - "step": 900 - }, - { - "entropy": 0.6187627577781677, - "epoch": 2.2895054282267795, - "grad_norm": 1.0338040590286255, - "learning_rate": 0.00012623281564719095, - "loss": 0.5596194076538086, - "mean_token_accuracy": 0.8182566928863525, - "num_tokens": 2280833.0, - "step": 950 - }, - { - "entropy": 0.6232365870475769, - "epoch": 2.4101326899879374, - "grad_norm": 1.0010799169540405, - "learning_rate": 0.00012496755584407422, - "loss": 0.5667143249511719, - "mean_token_accuracy": 0.8175061982870102, - "num_tokens": 2398206.0, - "step": 1000 - }, - { - "entropy": 0.6151268655061721, - "epoch": 2.5307599517490953, - "grad_norm": 1.0065034627914429, - "learning_rate": 0.0001235987393890801, - "loss": 0.5598890686035156, - "mean_token_accuracy": 0.8193449324369431, - "num_tokens": 2518888.0, - "step": 1050 - }, - { - "entropy": 0.6173626267910004, - "epoch": 2.651387213510253, - "grad_norm": 0.7958805561065674, - "learning_rate": 0.00012212878697538592, - "loss": 0.5649292755126953, - "mean_token_accuracy": 0.8172902607917786, - "num_tokens": 2639763.0, - "step": 1100 - }, - { - "entropy": 0.6264930808544159, - "epoch": 2.772014475271411, - "grad_norm": 0.7718372344970703, - "learning_rate": 0.00012056029815077323, - "loss": 0.5716455078125, - "mean_token_accuracy": 0.8170186239480972, - "num_tokens": 2755540.0, - "step": 1150 - }, - { - "entropy": 0.6203387409448624, - "epoch": 2.8926417370325694, - "grad_norm": 0.7914765477180481, - "learning_rate": 0.000118896046720439, - "loss": 0.5653961944580078, - "mean_token_accuracy": 0.8170280295610428, - "num_tokens": 2874118.0, - "step": 1200 - }, - { - "epoch": 3.0, - "eval_entropy": 0.6751241854737314, - "eval_mean_token_accuracy": 0.792226163858778, - "eval_not_syn_loss": 0.6902772784233093, - "eval_not_syn_runtime": 53.2557, - "eval_not_syn_samples_per_second": 26.664, - "eval_not_syn_steps_per_second": 3.342, - "eval_num_tokens": 2981418.0, - "step": 1245 - }, - { - "epoch": 3.0, - "eval_entropy": 0.6308753868837035, - "eval_mean_token_accuracy": 0.8131611759073278, - "eval_num_tokens": 2981418.0, - "eval_syn_loss": 0.6440441012382507, - "eval_syn_runtime": 54.7064, - "eval_syn_samples_per_second": 25.957, - "eval_syn_steps_per_second": 3.254, - "step": 1245 - }, - { - "entropy": 0.6077835421369533, - "epoch": 3.012062726176116, - "grad_norm": 1.124720573425293, - "learning_rate": 0.00011713897584164003, - "loss": 0.5557136917114258, - "mean_token_accuracy": 0.8203362726201915, - "num_tokens": 2994602.0, - "step": 1250 - }, - { - "entropy": 0.507642827630043, - "epoch": 3.132689987937274, - "grad_norm": 0.9039916396141052, - "learning_rate": 0.00011529219281884579, - "loss": 0.4467372512817383, - "mean_token_accuracy": 0.8464260965585708, - "num_tokens": 3121094.0, - "step": 1300 - }, - { - "entropy": 0.5137191081047058, - "epoch": 3.2533172496984317, - "grad_norm": 0.8580691814422607, - "learning_rate": 0.00011335896360860367, - "loss": 0.4538694763183594, - "mean_token_accuracy": 0.8448719161748887, - "num_tokens": 3248038.0, - "step": 1350 - }, - { - "entropy": 0.5293738862872124, - "epoch": 3.37394451145959, - "grad_norm": 1.0177273750305176, - "learning_rate": 0.00011134270704383492, - "loss": 0.4639348220825195, - "mean_token_accuracy": 0.8428975892066956, - "num_tokens": 3363487.0, - "step": 1400 - }, - { - "entropy": 0.5218872320652008, - "epoch": 3.494571773220748, - "grad_norm": 1.0530673265457153, - "learning_rate": 0.00010924698878777538, - "loss": 0.4584346389770508, - "mean_token_accuracy": 0.8432288640737533, - "num_tokens": 3480513.0, - "step": 1450 - }, - { - "entropy": 0.5320372316241264, - "epoch": 3.615199034981906, - "grad_norm": 1.0291194915771484, - "learning_rate": 0.0001070755150282529, - "loss": 0.4711898422241211, - "mean_token_accuracy": 0.842066935300827, - "num_tokens": 3597183.0, - "step": 1500 - }, - { - "entropy": 0.5315000694990158, - "epoch": 3.7358262967430638, - "grad_norm": 0.9788838624954224, - "learning_rate": 0.00010483212592345332, - "loss": 0.4729101943969727, - "mean_token_accuracy": 0.8404586082696914, - "num_tokens": 3714173.0, - "step": 1550 - }, - { - "entropy": 0.5331643190979958, - "epoch": 3.856453558504222, - "grad_norm": 1.1248903274536133, - "learning_rate": 0.00010252078881076532, - "loss": 0.47405517578125, - "mean_token_accuracy": 0.8397526168823242, - "num_tokens": 3832652.0, - "step": 1600 - }, - { - "entropy": 0.5343865439295769, - "epoch": 3.97708082026538, - "grad_norm": 0.9008272886276245, - "learning_rate": 0.00010014559119071452, - "loss": 0.47172779083251953, - "mean_token_accuracy": 0.8404951041936874, - "num_tokens": 3955167.0, - "step": 1650 - }, - { - "epoch": 4.0, - "eval_entropy": 0.5736920103598177, - "eval_mean_token_accuracy": 0.7981750027517255, - "eval_not_syn_loss": 0.707631528377533, - "eval_not_syn_runtime": 52.2858, - "eval_not_syn_samples_per_second": 27.158, - "eval_not_syn_steps_per_second": 3.404, - "eval_num_tokens": 3975224.0, - "step": 1660 - }, - { - "epoch": 4.0, - "eval_entropy": 0.5363499959198277, - "eval_mean_token_accuracy": 0.8063885539435269, - "eval_num_tokens": 3975224.0, - "eval_syn_loss": 0.6610181331634521, - "eval_syn_runtime": 54.2167, - "eval_syn_samples_per_second": 26.191, - "eval_syn_steps_per_second": 3.283, - "step": 1660 - }, - { - "entropy": 0.4512602482179199, - "epoch": 4.0965018094089265, - "grad_norm": 0.8079906105995178, - "learning_rate": 9.771073349839416e-05, - "loss": 0.3778452682495117, - "mean_token_accuracy": 0.8684436221315404, - "num_tokens": 4072361.0, - "step": 1700 - }, - { - "entropy": 0.4228873434662819, - "epoch": 4.217129071170085, - "grad_norm": 1.1206746101379395, - "learning_rate": 9.52205216751758e-05, - "loss": 0.34671127319335937, - "mean_token_accuracy": 0.8763950079679489, - "num_tokens": 4194978.0, - "step": 1750 - }, - { - "entropy": 0.42808119297027586, - "epoch": 4.337756332931242, - "grad_norm": 1.1268125772476196, - "learning_rate": 9.26793595538368e-05, - "loss": 0.3484600830078125, - "mean_token_accuracy": 0.8753323918581009, - "num_tokens": 4315548.0, - "step": 1800 - }, - { - "entropy": 0.42506767451763156, - "epoch": 4.458383594692401, - "grad_norm": 1.1051807403564453, - "learning_rate": 9.009174107057088e-05, - "loss": 0.35434780120849607, - "mean_token_accuracy": 0.8734411865472793, - "num_tokens": 4431469.0, - "step": 1850 - }, - { - "entropy": 0.43314908266067503, - "epoch": 4.579010856453559, - "grad_norm": 0.9782640337944031, - "learning_rate": 8.746224231765484e-05, - "loss": 0.36225860595703124, - "mean_token_accuracy": 0.8715140378475189, - "num_tokens": 4551803.0, - "step": 1900 - }, - { - "entropy": 0.42984022706747055, - "epoch": 4.699638118214716, - "grad_norm": 0.9772566556930542, - "learning_rate": 8.479551345082568e-05, - "loss": 0.356609001159668, - "mean_token_accuracy": 0.8717925781011582, - "num_tokens": 4674468.0, - "step": 1950 - }, - { - "entropy": 0.4474307084083557, - "epoch": 4.820265379975875, - "grad_norm": 0.9547265768051147, - "learning_rate": 8.209627046567966e-05, - "loss": 0.3654902267456055, - "mean_token_accuracy": 0.8724055212736129, - "num_tokens": 4788027.0, - "step": 2000 - }, - { - "entropy": 0.4377285006642342, - "epoch": 4.940892641737032, - "grad_norm": 1.0390652418136597, - "learning_rate": 7.936928685763635e-05, - "loss": 0.3645761871337891, - "mean_token_accuracy": 0.8707609891891479, - "num_tokens": 4907561.0, - "step": 2050 - }, - { - "epoch": 5.0, - "eval_entropy": 0.5173388388049737, - "eval_mean_token_accuracy": 0.7918405432379647, - "eval_not_syn_loss": 0.7532826662063599, - "eval_not_syn_runtime": 53.0766, - "eval_not_syn_samples_per_second": 26.754, - "eval_not_syn_steps_per_second": 3.354, - "eval_num_tokens": 4969030.0, - "step": 2075 - }, - { - "epoch": 5.0, - "eval_entropy": 0.4831916272975086, - "eval_mean_token_accuracy": 0.8204237202580056, - "eval_num_tokens": 4969030.0, - "eval_syn_loss": 0.6971890330314636, - "eval_syn_runtime": 54.8755, - "eval_syn_samples_per_second": 25.877, - "eval_syn_steps_per_second": 3.244, - "step": 2075 - }, - { - "entropy": 0.3804885711934831, - "epoch": 5.060313630880579, - "grad_norm": 0.8717736005783081, - "learning_rate": 7.661938518021687e-05, - "loss": 0.3034685707092285, - "mean_token_accuracy": 0.8910005471923135, - "num_tokens": 5026625.0, - "step": 2100 - }, - { - "entropy": 0.3174623078107834, - "epoch": 5.180940892641737, - "grad_norm": 0.9808429479598999, - "learning_rate": 7.385142851656478e-05, - "loss": 0.23829608917236328, - "mean_token_accuracy": 0.9131860065460206, - "num_tokens": 5141010.0, - "step": 2150 - }, - { - "entropy": 0.3187542732059956, - "epoch": 5.301568154402895, - "grad_norm": 1.3739993572235107, - "learning_rate": 7.107031187929228e-05, - "loss": 0.24024734497070313, - "mean_token_accuracy": 0.910272576212883, - "num_tokens": 5261240.0, - "step": 2200 - }, - { - "entropy": 0.31624249830842016, - "epoch": 5.422195416164053, - "grad_norm": 1.1618341207504272, - "learning_rate": 6.82809535538603e-05, - "loss": 0.24144052505493163, - "mean_token_accuracy": 0.9122701978683472, - "num_tokens": 5382160.0, - "step": 2250 - }, - { - "entropy": 0.31793304443359377, - "epoch": 5.542822677925211, - "grad_norm": 1.4960980415344238, - "learning_rate": 6.548828640080172e-05, - "loss": 0.24051647186279296, - "mean_token_accuracy": 0.9116496992111206, - "num_tokens": 5502002.0, - "step": 2300 - }, - { - "entropy": 0.3180307722091675, - "epoch": 5.66344993968637, - "grad_norm": 1.103051781654358, - "learning_rate": 6.269724913216913e-05, - "loss": 0.24294727325439452, - "mean_token_accuracy": 0.9102917027473449, - "num_tokens": 5629969.0, - "step": 2350 - }, - { - "entropy": 0.32020670652389527, - "epoch": 5.784077201447527, - "grad_norm": 1.0690243244171143, - "learning_rate": 5.991277757763442e-05, - "loss": 0.24328670501708985, - "mean_token_accuracy": 0.9101310813426972, - "num_tokens": 5750596.0, - "step": 2400 - }, - { - "entropy": 0.31846788004040716, - "epoch": 5.904704463208685, - "grad_norm": 1.203026294708252, - "learning_rate": 5.7139795955685784e-05, - "loss": 0.244984130859375, - "mean_token_accuracy": 0.9100374561548233, - "num_tokens": 5868110.0, - "step": 2450 - }, - { - "epoch": 6.0, - "eval_entropy": 0.45794740654109567, - "eval_mean_token_accuracy": 0.7891001142142864, - "eval_not_syn_loss": 0.8338615894317627, - "eval_not_syn_runtime": 53.3841, - "eval_not_syn_samples_per_second": 26.6, - "eval_not_syn_steps_per_second": 3.334, - "eval_num_tokens": 5962836.0, - "step": 2490 - }, - { - "epoch": 6.0, - "eval_entropy": 0.4288232589370749, - "eval_mean_token_accuracy": 0.8114104933953017, - "eval_num_tokens": 5962836.0, - "eval_syn_loss": 0.7791845798492432, - "eval_syn_runtime": 54.7877, - "eval_syn_samples_per_second": 25.918, - "eval_syn_steps_per_second": 3.249, - "step": 2490 - } - ], - "logging_steps": 50, - "max_steps": 4150, - "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": 1.2673564949952768e+17, - "train_batch_size": 8, - "trial_name": null, - "trial_params": null -} diff --git a/substitutivity_original_Estonian/Qwen3.5-2B-Base_substitutivity_splits_original_features_train_substitutivity_splits_original_features_test1/checkpoint-2490/training_args.bin b/substitutivity_original_Estonian/Qwen3.5-2B-Base_substitutivity_splits_original_features_train_substitutivity_splits_original_features_test1/checkpoint-2490/training_args.bin deleted file mode 100644 index 48901efb30506830185bfaafecbe236e7095a006..0000000000000000000000000000000000000000 --- a/substitutivity_original_Estonian/Qwen3.5-2B-Base_substitutivity_splits_original_features_train_substitutivity_splits_original_features_test1/checkpoint-2490/training_args.bin +++ /dev/null @@ -1,3 +0,0 @@ -version https://git-lfs.github.com/spec/v1 -oid sha256:770fc15ebc27a54d780af597d32c57ca14b3b528f14788d142bdfce09b555bef -size 6033 diff --git a/substitutivity_original_Estonian/Qwen3.5-2B-Base_substitutivity_splits_original_features_train_substitutivity_splits_original_features_test1/checkpoint-2905/README.md b/substitutivity_original_Estonian/Qwen3.5-2B-Base_substitutivity_splits_original_features_train_substitutivity_splits_original_features_test1/checkpoint-2905/README.md deleted file mode 100644 index ece204d005f7bf161d0e448efdd56fd297b7c8a2..0000000000000000000000000000000000000000 --- a/substitutivity_original_Estonian/Qwen3.5-2B-Base_substitutivity_splits_original_features_train_substitutivity_splits_original_features_test1/checkpoint-2905/README.md +++ /dev/null @@ -1,209 +0,0 @@ ---- -base_model: Qwen/Qwen3.5-2B-Base -library_name: peft -pipeline_tag: text-generation -tags: -- base_model:adapter:Qwen/Qwen3.5-2B-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/substitutivity_original_Estonian/Qwen3.5-2B-Base_substitutivity_splits_original_features_train_substitutivity_splits_original_features_test1/checkpoint-2905/adapter_config.json b/substitutivity_original_Estonian/Qwen3.5-2B-Base_substitutivity_splits_original_features_train_substitutivity_splits_original_features_test1/checkpoint-2905/adapter_config.json deleted file mode 100644 index a64145ce23adcbdbe246679087a13f4cb63ff664..0000000000000000000000000000000000000000 --- a/substitutivity_original_Estonian/Qwen3.5-2B-Base_substitutivity_splits_original_features_train_substitutivity_splits_original_features_test1/checkpoint-2905/adapter_config.json +++ /dev/null @@ -1,46 +0,0 @@ -{ - "alora_invocation_tokens": null, - "alpha_pattern": {}, - "arrow_config": null, - "auto_mapping": null, - "base_model_name_or_path": "Qwen/Qwen3.5-2B-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": 256, - "lora_bias": false, - "lora_dropout": 0.027050353915701653, - "megatron_config": null, - "megatron_core": "megatron.core", - "modules_to_save": null, - "peft_type": "LORA", - "peft_version": "0.18.1", - "qalora_group_size": 16, - "r": 128, - "rank_pattern": {}, - "revision": null, - "target_modules": [ - "o_proj", - "q_proj", - "k_proj", - "down_proj", - "up_proj", - "v_proj", - "gate_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/substitutivity_original_Estonian/Qwen3.5-2B-Base_substitutivity_splits_original_features_train_substitutivity_splits_original_features_test1/checkpoint-2905/adapter_model.safetensors b/substitutivity_original_Estonian/Qwen3.5-2B-Base_substitutivity_splits_original_features_train_substitutivity_splits_original_features_test1/checkpoint-2905/adapter_model.safetensors deleted file mode 100644 index ee1a380911c0359942a14f76ba145fa1b4f02b32..0000000000000000000000000000000000000000 --- a/substitutivity_original_Estonian/Qwen3.5-2B-Base_substitutivity_splits_original_features_train_substitutivity_splits_original_features_test1/checkpoint-2905/adapter_model.safetensors +++ /dev/null @@ -1,3 +0,0 @@ -version https://git-lfs.github.com/spec/v1 -oid sha256:86d444de5a2715bc79592e7c0468576e52e59a3d77c90640aa14e87a29046869 -size 349201608 diff --git a/substitutivity_original_Estonian/Qwen3.5-2B-Base_substitutivity_splits_original_features_train_substitutivity_splits_original_features_test1/checkpoint-2905/chat_template.jinja b/substitutivity_original_Estonian/Qwen3.5-2B-Base_substitutivity_splits_original_features_train_substitutivity_splits_original_features_test1/checkpoint-2905/chat_template.jinja deleted file mode 100644 index 0ef09f214eaa6d9bca297988afc1454b5827b2c7..0000000000000000000000000000000000000000 --- a/substitutivity_original_Estonian/Qwen3.5-2B-Base_substitutivity_splits_original_features_train_substitutivity_splits_original_features_test1/checkpoint-2905/chat_template.jinja +++ /dev/null @@ -1,154 +0,0 @@ -{%- set image_count = namespace(value=0) %} -{%- set video_count = namespace(value=0) %} -{%- macro render_content(content, do_vision_count, is_system_content=false) %} - {%- if content is string %} - {{- content }} - {%- elif content is iterable and content is not mapping %} - {%- for item in content %} - {%- if 'image' in item or 'image_url' in item or item.type == 'image' %} - {%- if is_system_content %} - {{- raise_exception('System message cannot contain images.') }} - {%- endif %} - {%- if do_vision_count %} - {%- set image_count.value = image_count.value + 1 %} - {%- endif %} - {%- if add_vision_id %} - {{- 'Picture ' ~ image_count.value ~ ': ' }} - {%- endif %} - {{- '<|vision_start|><|image_pad|><|vision_end|>' }} - {%- elif 'video' in item or item.type == 'video' %} - {%- if is_system_content %} - {{- raise_exception('System message cannot contain videos.') }} - {%- endif %} - {%- if do_vision_count %} - {%- set video_count.value = video_count.value + 1 %} - {%- endif %} - {%- if add_vision_id %} - {{- 'Video ' ~ video_count.value ~ ': ' }} - {%- endif %} - {{- '<|vision_start|><|video_pad|><|vision_end|>' }} - {%- elif 'text' in item %} - {{- item.text }} - {%- else %} - {{- raise_exception('Unexpected item type in content.') }} - {%- endif %} - {%- endfor %} - {%- elif content is none or content is undefined %} - {{- '' }} - {%- else %} - {{- raise_exception('Unexpected content type.') }} - {%- endif %} -{%- endmacro %} -{%- if not messages %} - {{- raise_exception('No messages provided.') }} -{%- endif %} -{%- if tools and tools is iterable and tools is not mapping %} - {{- '<|im_start|>system\n' }} - {{- "# Tools\n\nYou have access to the following functions:\n\n" }} - {%- for tool in tools %} - {{- "\n" }} - {{- tool | tojson }} - {%- endfor %} - {{- "\n" }} - {{- '\n\nIf you choose to call a function ONLY reply in the following format with NO suffix:\n\n\n\n\nvalue_1\n\n\nThis is the value for the second parameter\nthat can span\nmultiple lines\n\n\n\n\n\nReminder:\n- Function calls MUST follow the specified format: an inner block must be nested within XML tags\n- Required parameters MUST be specified\n- You may provide optional reasoning for your function call in natural language BEFORE the function call, but NOT after\n- If there is no function call available, answer the question like normal with your current knowledge and do not tell the user about function calls\n' }} - {%- if messages[0].role == 'system' %} - {%- set content = render_content(messages[0].content, false, true)|trim %} - {%- if content %} - {{- '\n\n' + content }} - {%- endif %} - {%- endif %} - {{- '<|im_end|>\n' }} -{%- else %} - {%- if messages[0].role == 'system' %} - {%- set content = render_content(messages[0].content, false, true)|trim %} - {{- '<|im_start|>system\n' + 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" %} - {%- set content = render_content(message.content, false)|trim %} - {%- if not(content.startswith('') and content.endswith('')) %} - {%- set ns.multi_step_tool = false %} - {%- set ns.last_query_index = index %} - {%- endif %} - {%- endif %} -{%- endfor %} -{%- if ns.multi_step_tool %} - {{- raise_exception('No user query found in messages.') }} -{%- endif %} -{%- for message in messages %} - {%- set content = render_content(message.content, true)|trim %} - {%- if message.role == "system" %} - {%- if not loop.first %} - {{- raise_exception('System message must be at the beginning.') }} - {%- endif %} - {%- elif message.role == "user" %} - {{- '<|im_start|>' + message.role + '\n' + content + '<|im_end|>' + '\n' }} - {%- elif message.role == "assistant" %} - {%- set reasoning_content = '' %} - {%- if message.reasoning_content is string %} - {%- set reasoning_content = message.reasoning_content %} - {%- else %} - {%- if '' in content %} - {%- set reasoning_content = content.split('')[0].rstrip('\n').split('')[-1].lstrip('\n') %} - {%- set content = content.split('')[-1].lstrip('\n') %} - {%- endif %} - {%- endif %} - {%- set reasoning_content = reasoning_content|trim %} - {%- if loop.index0 > ns.last_query_index %} - {{- '<|im_start|>' + message.role + '\n\n' + reasoning_content + '\n\n\n' + content }} - {%- else %} - {{- '<|im_start|>' + message.role + '\n' + content }} - {%- endif %} - {%- if message.tool_calls and message.tool_calls is iterable and message.tool_calls is not mapping %} - {%- for tool_call in message.tool_calls %} - {%- if tool_call.function is defined %} - {%- set tool_call = tool_call.function %} - {%- endif %} - {%- if loop.first %} - {%- if content|trim %} - {{- '\n\n\n\n' }} - {%- else %} - {{- '\n\n' }} - {%- endif %} - {%- else %} - {{- '\n\n\n' }} - {%- endif %} - {%- if tool_call.arguments is defined %} - {%- for args_name, args_value in tool_call.arguments|items %} - {{- '\n' }} - {%- set args_value = args_value | tojson | safe if args_value is mapping or (args_value is sequence and args_value is not string) else args_value | string %} - {{- args_value }} - {{- '\n\n' }} - {%- endfor %} - {%- endif %} - {{- '\n' }} - {%- endfor %} - {%- endif %} - {{- '<|im_end|>\n' }} - {%- elif message.role == "tool" %} - {%- if loop.previtem and loop.previtem.role != "tool" %} - {{- '<|im_start|>user' }} - {%- endif %} - {{- '\n\n' }} - {{- content }} - {{- '\n' }} - {%- if not loop.last and loop.nextitem.role != "tool" %} - {{- '<|im_end|>\n' }} - {%- elif loop.last %} - {{- '<|im_end|>\n' }} - {%- endif %} - {%- else %} - {{- raise_exception('Unexpected message role.') }} - {%- endif %} -{%- endfor %} -{%- if add_generation_prompt %} - {{- '<|im_start|>assistant\n' }} - {%- if enable_thinking is defined and enable_thinking is true %} - {{- '\n' }} - {%- else %} - {{- '\n\n\n\n' }} - {%- endif %} -{%- endif %} \ No newline at end of file diff --git a/substitutivity_original_Estonian/Qwen3.5-2B-Base_substitutivity_splits_original_features_train_substitutivity_splits_original_features_test1/checkpoint-2905/tokenizer.json b/substitutivity_original_Estonian/Qwen3.5-2B-Base_substitutivity_splits_original_features_train_substitutivity_splits_original_features_test1/checkpoint-2905/tokenizer.json deleted file mode 100644 index 67741b04f23bfdb46501f748ce27865ec82eccfb..0000000000000000000000000000000000000000 --- a/substitutivity_original_Estonian/Qwen3.5-2B-Base_substitutivity_splits_original_features_train_substitutivity_splits_original_features_test1/checkpoint-2905/tokenizer.json +++ /dev/null @@ -1,3 +0,0 @@ -version https://git-lfs.github.com/spec/v1 -oid sha256:87a7830d63fcf43bf241c3c5242e96e62dd3fdc29224ca26fed8ea333db72de4 -size 19989343 diff --git a/substitutivity_original_Estonian/Qwen3.5-2B-Base_substitutivity_splits_original_features_train_substitutivity_splits_original_features_test1/checkpoint-2905/tokenizer_config.json b/substitutivity_original_Estonian/Qwen3.5-2B-Base_substitutivity_splits_original_features_train_substitutivity_splits_original_features_test1/checkpoint-2905/tokenizer_config.json deleted file mode 100644 index b4a37b2a6fd3ab3317cd7bac72855be1a843b2bb..0000000000000000000000000000000000000000 --- a/substitutivity_original_Estonian/Qwen3.5-2B-Base_substitutivity_splits_original_features_train_substitutivity_splits_original_features_test1/checkpoint-2905/tokenizer_config.json +++ /dev/null @@ -1,31 +0,0 @@ -{ - "add_prefix_space": false, - "audio_bos_token": "<|audio_start|>", - "audio_eos_token": "<|audio_end|>", - "audio_token": "<|audio_pad|>", - "backend": "tokenizers", - "bos_token": null, - "clean_up_tokenization_spaces": false, - "eos_token": "<|endoftext|>", - "errors": "replace", - "image_token": "<|image_pad|>", - "is_local": false, - "model_max_length": 262144, - "model_specific_special_tokens": { - "audio_bos_token": "<|audio_start|>", - "audio_eos_token": "<|audio_end|>", - "audio_token": "<|audio_pad|>", - "image_token": "<|image_pad|>", - "video_token": "<|video_pad|>", - "vision_bos_token": "<|vision_start|>", - "vision_eos_token": "<|vision_end|>" - }, - "pad_token": "<|endoftext|>", - "pretokenize_regex": "(?i:'s|'t|'re|'ve|'m|'ll|'d)|[^\\r\\n\\p{L}\\p{N}]?[\\p{L}\\p{M}]+|\\p{N}| ?[^\\s\\p{L}\\p{M}\\p{N}]+[\\r\\n]*|\\s*[\\r\\n]+|\\s+(?!\\S)|\\s+", - "split_special_tokens": false, - "tokenizer_class": "TokenizersBackend", - "unk_token": null, - "video_token": "<|video_pad|>", - "vision_bos_token": "<|vision_start|>", - "vision_eos_token": "<|vision_end|>" -} diff --git a/substitutivity_original_Estonian/Qwen3.5-2B-Base_substitutivity_splits_original_features_train_substitutivity_splits_original_features_test1/checkpoint-2905/trainer_state.json b/substitutivity_original_Estonian/Qwen3.5-2B-Base_substitutivity_splits_original_features_train_substitutivity_splits_original_features_test1/checkpoint-2905/trainer_state.json deleted file mode 100644 index fdbdc3deba01cf6577b271c86a61f6fb29b764df..0000000000000000000000000000000000000000 --- a/substitutivity_original_Estonian/Qwen3.5-2B-Base_substitutivity_splits_original_features_train_substitutivity_splits_original_features_test1/checkpoint-2905/trainer_state.json +++ /dev/null @@ -1,768 +0,0 @@ -{ - "best_global_step": null, - "best_metric": null, - "best_model_checkpoint": null, - "epoch": 7.0, - "eval_steps": 500, - "global_step": 2905, - "is_hyper_param_search": false, - "is_local_process_zero": true, - "is_world_process_zero": true, - "log_history": [ - { - "entropy": 2.0473356211185454, - "epoch": 0.12062726176115803, - "grad_norm": 3.0554118156433105, - "learning_rate": 1.5682344549790923e-05, - "loss": 1.910550079345703, - "mean_token_accuracy": 0.6053373458981514, - "num_tokens": 121428.0, - "step": 50 - }, - { - "entropy": 1.0371007251739501, - "epoch": 0.24125452352231605, - "grad_norm": 2.109868049621582, - "learning_rate": 3.168473694753677e-05, - "loss": 0.9585584259033203, - "mean_token_accuracy": 0.7352996730804443, - "num_tokens": 243966.0, - "step": 100 - }, - { - "entropy": 0.9105490958690643, - "epoch": 0.3618817852834741, - "grad_norm": 1.7675701379776, - "learning_rate": 4.7687129345282605e-05, - "loss": 0.8393524932861328, - "mean_token_accuracy": 0.7588736474514007, - "num_tokens": 366649.0, - "step": 150 - }, - { - "entropy": 0.8660378754138947, - "epoch": 0.4825090470446321, - "grad_norm": 1.6502987146377563, - "learning_rate": 6.368952174302845e-05, - "loss": 0.8005844116210937, - "mean_token_accuracy": 0.7663125538825989, - "num_tokens": 488367.0, - "step": 200 - }, - { - "entropy": 0.8439123743772506, - "epoch": 0.6031363088057901, - "grad_norm": 1.538365364074707, - "learning_rate": 7.969191414077429e-05, - "loss": 0.7773619842529297, - "mean_token_accuracy": 0.7718527519702911, - "num_tokens": 603131.0, - "step": 250 - }, - { - "entropy": 0.8223806774616241, - "epoch": 0.7237635705669482, - "grad_norm": 1.1559664011001587, - "learning_rate": 9.569430653852014e-05, - "loss": 0.7587767791748047, - "mean_token_accuracy": 0.7767325991392136, - "num_tokens": 726412.0, - "step": 300 - }, - { - "entropy": 0.8256935775279999, - "epoch": 0.8443908323281062, - "grad_norm": 1.2322715520858765, - "learning_rate": 0.00011169669893626597, - "loss": 0.769886245727539, - "mean_token_accuracy": 0.773840668797493, - "num_tokens": 846516.0, - "step": 350 - }, - { - "entropy": 0.8065843790769577, - "epoch": 0.9650180940892642, - "grad_norm": 0.9908286333084106, - "learning_rate": 0.00012769909133401183, - "loss": 0.7451084899902344, - "mean_token_accuracy": 0.7794792425632476, - "num_tokens": 958978.0, - "step": 400 - }, - { - "epoch": 1.0, - "eval_entropy": 0.791056221790528, - "eval_mean_token_accuracy": 0.7507295300451557, - "eval_not_syn_loss": 0.7860428094863892, - "eval_not_syn_runtime": 53.5431, - "eval_not_syn_samples_per_second": 26.521, - "eval_not_syn_steps_per_second": 3.324, - "eval_num_tokens": 993806.0, - "step": 415 - }, - { - "epoch": 1.0, - "eval_entropy": 0.7424047408478983, - "eval_mean_token_accuracy": 0.8023804037758474, - "eval_num_tokens": 993806.0, - "eval_syn_loss": 0.7273509502410889, - "eval_syn_runtime": 54.8679, - "eval_syn_samples_per_second": 25.88, - "eval_syn_steps_per_second": 3.244, - "step": 415 - }, - { - "entropy": 0.7607260761838971, - "epoch": 1.0844390832328106, - "grad_norm": 1.0653696060180664, - "learning_rate": 0.0001327927019027069, - "loss": 0.7034613037109375, - "mean_token_accuracy": 0.7869195998317063, - "num_tokens": 1080301.0, - "step": 450 - }, - { - "entropy": 0.7370894801616669, - "epoch": 1.2050663449939687, - "grad_norm": 1.763787865638733, - "learning_rate": 0.00013265416569994628, - "loss": 0.6816973114013671, - "mean_token_accuracy": 0.7914507347345352, - "num_tokens": 1201215.0, - "step": 500 - }, - { - "entropy": 0.742829909324646, - "epoch": 1.3256936067551266, - "grad_norm": 0.97663414478302, - "learning_rate": 0.00013239847940553975, - "loss": 0.6867338562011719, - "mean_token_accuracy": 0.790328289270401, - "num_tokens": 1322895.0, - "step": 550 - }, - { - "entropy": 0.7312483406066894, - "epoch": 1.4463208685162847, - "grad_norm": 1.0874731540679932, - "learning_rate": 0.00013202609518972978, - "loss": 0.676800537109375, - "mean_token_accuracy": 0.7921512228250503, - "num_tokens": 1436749.0, - "step": 600 - }, - { - "entropy": 0.7178869074583054, - "epoch": 1.5669481302774426, - "grad_norm": 1.064300298690796, - "learning_rate": 0.0001315376715980265, - "loss": 0.6635584259033203, - "mean_token_accuracy": 0.7944776183366775, - "num_tokens": 1555400.0, - "step": 650 - }, - { - "entropy": 0.7198166918754577, - "epoch": 1.6875753920386007, - "grad_norm": 0.7779315114021301, - "learning_rate": 0.00013093407238659802, - "loss": 0.6660567474365234, - "mean_token_accuracy": 0.7968300533294678, - "num_tokens": 1676892.0, - "step": 700 - }, - { - "entropy": 0.7170061504840851, - "epoch": 1.8082026537997589, - "grad_norm": 1.089854121208191, - "learning_rate": 0.00013021636499475496, - "loss": 0.6632771301269531, - "mean_token_accuracy": 0.7963915091753005, - "num_tokens": 1802689.0, - "step": 750 - }, - { - "entropy": 0.6953610718250275, - "epoch": 1.9288299155609168, - "grad_norm": 0.7996730208396912, - "learning_rate": 0.0001293858186572295, - "loss": 0.6421419525146485, - "mean_token_accuracy": 0.8002577543258667, - "num_tokens": 1920804.0, - "step": 800 - }, - { - "epoch": 2.0, - "eval_entropy": 0.7369416523515508, - "eval_mean_token_accuracy": 0.7872022366925572, - "eval_not_syn_loss": 0.713637113571167, - "eval_not_syn_runtime": 53.2265, - "eval_not_syn_samples_per_second": 26.678, - "eval_not_syn_steps_per_second": 3.344, - "eval_num_tokens": 1987612.0, - "step": 830 - }, - { - "epoch": 2.0, - "eval_entropy": 0.6877032756135705, - "eval_mean_token_accuracy": 0.7997466372639945, - "eval_num_tokens": 1987612.0, - "eval_syn_loss": 0.6614266633987427, - "eval_syn_runtime": 54.5178, - "eval_syn_samples_per_second": 26.047, - "eval_syn_steps_per_second": 3.265, - "step": 830 - }, - { - "entropy": 0.6644553433765065, - "epoch": 2.0482509047044632, - "grad_norm": 0.7562468647956848, - "learning_rate": 0.00012844390215958852, - "loss": 0.6112185287475586, - "mean_token_accuracy": 0.806649933559726, - "num_tokens": 2039660.0, - "step": 850 - }, - { - "entropy": 0.6146580249071121, - "epoch": 2.168878166465621, - "grad_norm": 1.098607063293457, - "learning_rate": 0.00012739228124074913, - "loss": 0.5553749084472657, - "mean_token_accuracy": 0.8187588346004486, - "num_tokens": 2161772.0, - "step": 900 - }, - { - "entropy": 0.6187627577781677, - "epoch": 2.2895054282267795, - "grad_norm": 1.0338040590286255, - "learning_rate": 0.00012623281564719095, - "loss": 0.5596194076538086, - "mean_token_accuracy": 0.8182566928863525, - "num_tokens": 2280833.0, - "step": 950 - }, - { - "entropy": 0.6232365870475769, - "epoch": 2.4101326899879374, - "grad_norm": 1.0010799169540405, - "learning_rate": 0.00012496755584407422, - "loss": 0.5667143249511719, - "mean_token_accuracy": 0.8175061982870102, - "num_tokens": 2398206.0, - "step": 1000 - }, - { - "entropy": 0.6151268655061721, - "epoch": 2.5307599517490953, - "grad_norm": 1.0065034627914429, - "learning_rate": 0.0001235987393890801, - "loss": 0.5598890686035156, - "mean_token_accuracy": 0.8193449324369431, - "num_tokens": 2518888.0, - "step": 1050 - }, - { - "entropy": 0.6173626267910004, - "epoch": 2.651387213510253, - "grad_norm": 0.7958805561065674, - "learning_rate": 0.00012212878697538592, - "loss": 0.5649292755126953, - "mean_token_accuracy": 0.8172902607917786, - "num_tokens": 2639763.0, - "step": 1100 - }, - { - "entropy": 0.6264930808544159, - "epoch": 2.772014475271411, - "grad_norm": 0.7718372344970703, - "learning_rate": 0.00012056029815077323, - "loss": 0.5716455078125, - "mean_token_accuracy": 0.8170186239480972, - "num_tokens": 2755540.0, - "step": 1150 - }, - { - "entropy": 0.6203387409448624, - "epoch": 2.8926417370325694, - "grad_norm": 0.7914765477180481, - "learning_rate": 0.000118896046720439, - "loss": 0.5653961944580078, - "mean_token_accuracy": 0.8170280295610428, - "num_tokens": 2874118.0, - "step": 1200 - }, - { - "epoch": 3.0, - "eval_entropy": 0.6751241854737314, - "eval_mean_token_accuracy": 0.792226163858778, - "eval_not_syn_loss": 0.6902772784233093, - "eval_not_syn_runtime": 53.2557, - "eval_not_syn_samples_per_second": 26.664, - "eval_not_syn_steps_per_second": 3.342, - "eval_num_tokens": 2981418.0, - "step": 1245 - }, - { - "epoch": 3.0, - "eval_entropy": 0.6308753868837035, - "eval_mean_token_accuracy": 0.8131611759073278, - "eval_num_tokens": 2981418.0, - "eval_syn_loss": 0.6440441012382507, - "eval_syn_runtime": 54.7064, - "eval_syn_samples_per_second": 25.957, - "eval_syn_steps_per_second": 3.254, - "step": 1245 - }, - { - "entropy": 0.6077835421369533, - "epoch": 3.012062726176116, - "grad_norm": 1.124720573425293, - "learning_rate": 0.00011713897584164003, - "loss": 0.5557136917114258, - "mean_token_accuracy": 0.8203362726201915, - "num_tokens": 2994602.0, - "step": 1250 - }, - { - "entropy": 0.507642827630043, - "epoch": 3.132689987937274, - "grad_norm": 0.9039916396141052, - "learning_rate": 0.00011529219281884579, - "loss": 0.4467372512817383, - "mean_token_accuracy": 0.8464260965585708, - "num_tokens": 3121094.0, - "step": 1300 - }, - { - "entropy": 0.5137191081047058, - "epoch": 3.2533172496984317, - "grad_norm": 0.8580691814422607, - "learning_rate": 0.00011335896360860367, - "loss": 0.4538694763183594, - "mean_token_accuracy": 0.8448719161748887, - "num_tokens": 3248038.0, - "step": 1350 - }, - { - "entropy": 0.5293738862872124, - "epoch": 3.37394451145959, - "grad_norm": 1.0177273750305176, - "learning_rate": 0.00011134270704383492, - "loss": 0.4639348220825195, - "mean_token_accuracy": 0.8428975892066956, - "num_tokens": 3363487.0, - "step": 1400 - }, - { - "entropy": 0.5218872320652008, - "epoch": 3.494571773220748, - "grad_norm": 1.0530673265457153, - "learning_rate": 0.00010924698878777538, - "loss": 0.4584346389770508, - "mean_token_accuracy": 0.8432288640737533, - "num_tokens": 3480513.0, - "step": 1450 - }, - { - "entropy": 0.5320372316241264, - "epoch": 3.615199034981906, - "grad_norm": 1.0291194915771484, - "learning_rate": 0.0001070755150282529, - "loss": 0.4711898422241211, - "mean_token_accuracy": 0.842066935300827, - "num_tokens": 3597183.0, - "step": 1500 - }, - { - "entropy": 0.5315000694990158, - "epoch": 3.7358262967430638, - "grad_norm": 0.9788838624954224, - "learning_rate": 0.00010483212592345332, - "loss": 0.4729101943969727, - "mean_token_accuracy": 0.8404586082696914, - "num_tokens": 3714173.0, - "step": 1550 - }, - { - "entropy": 0.5331643190979958, - "epoch": 3.856453558504222, - "grad_norm": 1.1248903274536133, - "learning_rate": 0.00010252078881076532, - "loss": 0.47405517578125, - "mean_token_accuracy": 0.8397526168823242, - "num_tokens": 3832652.0, - "step": 1600 - }, - { - "entropy": 0.5343865439295769, - "epoch": 3.97708082026538, - "grad_norm": 0.9008272886276245, - "learning_rate": 0.00010014559119071452, - "loss": 0.47172779083251953, - "mean_token_accuracy": 0.8404951041936874, - "num_tokens": 3955167.0, - "step": 1650 - }, - { - "epoch": 4.0, - "eval_entropy": 0.5736920103598177, - "eval_mean_token_accuracy": 0.7981750027517255, - "eval_not_syn_loss": 0.707631528377533, - "eval_not_syn_runtime": 52.2858, - "eval_not_syn_samples_per_second": 27.158, - "eval_not_syn_steps_per_second": 3.404, - "eval_num_tokens": 3975224.0, - "step": 1660 - }, - { - "epoch": 4.0, - "eval_entropy": 0.5363499959198277, - "eval_mean_token_accuracy": 0.8063885539435269, - "eval_num_tokens": 3975224.0, - "eval_syn_loss": 0.6610181331634521, - "eval_syn_runtime": 54.2167, - "eval_syn_samples_per_second": 26.191, - "eval_syn_steps_per_second": 3.283, - "step": 1660 - }, - { - "entropy": 0.4512602482179199, - "epoch": 4.0965018094089265, - "grad_norm": 0.8079906105995178, - "learning_rate": 9.771073349839416e-05, - "loss": 0.3778452682495117, - "mean_token_accuracy": 0.8684436221315404, - "num_tokens": 4072361.0, - "step": 1700 - }, - { - "entropy": 0.4228873434662819, - "epoch": 4.217129071170085, - "grad_norm": 1.1206746101379395, - "learning_rate": 9.52205216751758e-05, - "loss": 0.34671127319335937, - "mean_token_accuracy": 0.8763950079679489, - "num_tokens": 4194978.0, - "step": 1750 - }, - { - "entropy": 0.42808119297027586, - "epoch": 4.337756332931242, - "grad_norm": 1.1268125772476196, - "learning_rate": 9.26793595538368e-05, - "loss": 0.3484600830078125, - "mean_token_accuracy": 0.8753323918581009, - "num_tokens": 4315548.0, - "step": 1800 - }, - { - "entropy": 0.42506767451763156, - "epoch": 4.458383594692401, - "grad_norm": 1.1051807403564453, - "learning_rate": 9.009174107057088e-05, - "loss": 0.35434780120849607, - "mean_token_accuracy": 0.8734411865472793, - "num_tokens": 4431469.0, - "step": 1850 - }, - { - "entropy": 0.43314908266067503, - "epoch": 4.579010856453559, - "grad_norm": 0.9782640337944031, - "learning_rate": 8.746224231765484e-05, - "loss": 0.36225860595703124, - "mean_token_accuracy": 0.8715140378475189, - "num_tokens": 4551803.0, - "step": 1900 - }, - { - "entropy": 0.42984022706747055, - "epoch": 4.699638118214716, - "grad_norm": 0.9772566556930542, - "learning_rate": 8.479551345082568e-05, - "loss": 0.356609001159668, - "mean_token_accuracy": 0.8717925781011582, - "num_tokens": 4674468.0, - "step": 1950 - }, - { - "entropy": 0.4474307084083557, - "epoch": 4.820265379975875, - "grad_norm": 0.9547265768051147, - "learning_rate": 8.209627046567966e-05, - "loss": 0.3654902267456055, - "mean_token_accuracy": 0.8724055212736129, - "num_tokens": 4788027.0, - "step": 2000 - }, - { - "entropy": 0.4377285006642342, - "epoch": 4.940892641737032, - "grad_norm": 1.0390652418136597, - "learning_rate": 7.936928685763635e-05, - "loss": 0.3645761871337891, - "mean_token_accuracy": 0.8707609891891479, - "num_tokens": 4907561.0, - "step": 2050 - }, - { - "epoch": 5.0, - "eval_entropy": 0.5173388388049737, - "eval_mean_token_accuracy": 0.7918405432379647, - "eval_not_syn_loss": 0.7532826662063599, - "eval_not_syn_runtime": 53.0766, - "eval_not_syn_samples_per_second": 26.754, - "eval_not_syn_steps_per_second": 3.354, - "eval_num_tokens": 4969030.0, - "step": 2075 - }, - { - "epoch": 5.0, - "eval_entropy": 0.4831916272975086, - "eval_mean_token_accuracy": 0.8204237202580056, - "eval_num_tokens": 4969030.0, - "eval_syn_loss": 0.6971890330314636, - "eval_syn_runtime": 54.8755, - "eval_syn_samples_per_second": 25.877, - "eval_syn_steps_per_second": 3.244, - "step": 2075 - }, - { - "entropy": 0.3804885711934831, - "epoch": 5.060313630880579, - "grad_norm": 0.8717736005783081, - "learning_rate": 7.661938518021687e-05, - "loss": 0.3034685707092285, - "mean_token_accuracy": 0.8910005471923135, - "num_tokens": 5026625.0, - "step": 2100 - }, - { - "entropy": 0.3174623078107834, - "epoch": 5.180940892641737, - "grad_norm": 0.9808429479598999, - "learning_rate": 7.385142851656478e-05, - "loss": 0.23829608917236328, - "mean_token_accuracy": 0.9131860065460206, - "num_tokens": 5141010.0, - "step": 2150 - }, - { - "entropy": 0.3187542732059956, - "epoch": 5.301568154402895, - "grad_norm": 1.3739993572235107, - "learning_rate": 7.107031187929228e-05, - "loss": 0.24024734497070313, - "mean_token_accuracy": 0.910272576212883, - "num_tokens": 5261240.0, - "step": 2200 - }, - { - "entropy": 0.31624249830842016, - "epoch": 5.422195416164053, - "grad_norm": 1.1618341207504272, - "learning_rate": 6.82809535538603e-05, - "loss": 0.24144052505493163, - "mean_token_accuracy": 0.9122701978683472, - "num_tokens": 5382160.0, - "step": 2250 - }, - { - "entropy": 0.31793304443359377, - "epoch": 5.542822677925211, - "grad_norm": 1.4960980415344238, - "learning_rate": 6.548828640080172e-05, - "loss": 0.24051647186279296, - "mean_token_accuracy": 0.9116496992111206, - "num_tokens": 5502002.0, - "step": 2300 - }, - { - "entropy": 0.3180307722091675, - "epoch": 5.66344993968637, - "grad_norm": 1.103051781654358, - "learning_rate": 6.269724913216913e-05, - "loss": 0.24294727325439452, - "mean_token_accuracy": 0.9102917027473449, - "num_tokens": 5629969.0, - "step": 2350 - }, - { - "entropy": 0.32020670652389527, - "epoch": 5.784077201447527, - "grad_norm": 1.0690243244171143, - "learning_rate": 5.991277757763442e-05, - "loss": 0.24328670501708985, - "mean_token_accuracy": 0.9101310813426972, - "num_tokens": 5750596.0, - "step": 2400 - }, - { - "entropy": 0.31846788004040716, - "epoch": 5.904704463208685, - "grad_norm": 1.203026294708252, - "learning_rate": 5.7139795955685784e-05, - "loss": 0.244984130859375, - "mean_token_accuracy": 0.9100374561548233, - "num_tokens": 5868110.0, - "step": 2450 - }, - { - "epoch": 6.0, - "eval_entropy": 0.45794740654109567, - "eval_mean_token_accuracy": 0.7891001142142864, - "eval_not_syn_loss": 0.8338615894317627, - "eval_not_syn_runtime": 53.3841, - "eval_not_syn_samples_per_second": 26.6, - "eval_not_syn_steps_per_second": 3.334, - "eval_num_tokens": 5962836.0, - "step": 2490 - }, - { - "epoch": 6.0, - "eval_entropy": 0.4288232589370749, - "eval_mean_token_accuracy": 0.8114104933953017, - "eval_num_tokens": 5962836.0, - "eval_syn_loss": 0.7791845798492432, - "eval_syn_runtime": 54.7877, - "eval_syn_samples_per_second": 25.918, - "eval_syn_steps_per_second": 3.249, - "step": 2490 - }, - { - "entropy": 0.2972855128423132, - "epoch": 6.024125452352232, - "grad_norm": 1.1748684644699097, - "learning_rate": 5.438320816535859e-05, - "loss": 0.22051204681396486, - "mean_token_accuracy": 0.9190925630656156, - "num_tokens": 5990107.0, - "step": 2500 - }, - { - "entropy": 0.2216818816959858, - "epoch": 6.144752714113389, - "grad_norm": 1.0899405479431152, - "learning_rate": 5.164788911390022e-05, - "loss": 0.1487915325164795, - "mean_token_accuracy": 0.9462692165374755, - "num_tokens": 6104230.0, - "step": 2550 - }, - { - "entropy": 0.22189782992005347, - "epoch": 6.265379975874548, - "grad_norm": 1.0967661142349243, - "learning_rate": 4.8938676095705914e-05, - "loss": 0.149844388961792, - "mean_token_accuracy": 0.9452104204893113, - "num_tokens": 6223548.0, - "step": 2600 - }, - { - "entropy": 0.2155836282670498, - "epoch": 6.386007237635706, - "grad_norm": 1.1443830728530884, - "learning_rate": 4.626036023777093e-05, - "loss": 0.1472412109375, - "mean_token_accuracy": 0.946697844862938, - "num_tokens": 6349102.0, - "step": 2650 - }, - { - "entropy": 0.22189507275819778, - "epoch": 6.5066344993968634, - "grad_norm": 1.2138874530792236, - "learning_rate": 4.3617678026788244e-05, - "loss": 0.15155397415161131, - "mean_token_accuracy": 0.9441326081752777, - "num_tokens": 6468966.0, - "step": 2700 - }, - { - "entropy": 0.22405965447425844, - "epoch": 6.627261761158022, - "grad_norm": 1.2444716691970825, - "learning_rate": 4.10153029328748e-05, - "loss": 0.15516298294067382, - "mean_token_accuracy": 0.9442053025960923, - "num_tokens": 6583296.0, - "step": 2750 - }, - { - "entropy": 0.2216674068570137, - "epoch": 6.74788902291918, - "grad_norm": 1.0056185722351074, - "learning_rate": 3.845783714474033e-05, - "loss": 0.15423223495483399, - "mean_token_accuracy": 0.9445771616697312, - "num_tokens": 6698696.0, - "step": 2800 - }, - { - "entropy": 0.21455578804016112, - "epoch": 6.868516284680338, - "grad_norm": 1.0748720169067383, - "learning_rate": 3.59498034309138e-05, - "loss": 0.14870482444763183, - "mean_token_accuracy": 0.9457976347208024, - "num_tokens": 6819711.0, - "step": 2850 - }, - { - "entropy": 0.21227611482143402, - "epoch": 6.989143546441496, - "grad_norm": 0.978204607963562, - "learning_rate": 3.349563714142155e-05, - "loss": 0.14802613258361816, - "mean_token_accuracy": 0.9466447854042053, - "num_tokens": 6944941.0, - "step": 2900 - }, - { - "epoch": 7.0, - "eval_entropy": 0.38817436741978933, - "eval_mean_token_accuracy": 0.7801904045463948, - "eval_not_syn_loss": 0.9901206493377686, - "eval_not_syn_runtime": 53.1137, - "eval_not_syn_samples_per_second": 26.735, - "eval_not_syn_steps_per_second": 3.351, - "eval_num_tokens": 6956642.0, - "step": 2905 - }, - { - "epoch": 7.0, - "eval_entropy": 0.3640736076938972, - "eval_mean_token_accuracy": 0.8158511425671953, - "eval_num_tokens": 6956642.0, - "eval_syn_loss": 0.8903024792671204, - "eval_syn_runtime": 54.3768, - "eval_syn_samples_per_second": 26.114, - "eval_syn_steps_per_second": 3.273, - "step": 2905 - } - ], - "logging_steps": 50, - "max_steps": 4150, - "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": 1.4783731624475712e+17, - "train_batch_size": 8, - "trial_name": null, - "trial_params": null -} diff --git a/substitutivity_original_Estonian/Qwen3.5-2B-Base_substitutivity_splits_original_features_train_substitutivity_splits_original_features_test1/checkpoint-2905/training_args.bin b/substitutivity_original_Estonian/Qwen3.5-2B-Base_substitutivity_splits_original_features_train_substitutivity_splits_original_features_test1/checkpoint-2905/training_args.bin deleted file mode 100644 index 48901efb30506830185bfaafecbe236e7095a006..0000000000000000000000000000000000000000 --- a/substitutivity_original_Estonian/Qwen3.5-2B-Base_substitutivity_splits_original_features_train_substitutivity_splits_original_features_test1/checkpoint-2905/training_args.bin +++ /dev/null @@ -1,3 +0,0 @@ -version https://git-lfs.github.com/spec/v1 -oid sha256:770fc15ebc27a54d780af597d32c57ca14b3b528f14788d142bdfce09b555bef -size 6033 diff --git a/substitutivity_original_Estonian/Qwen3.5-2B-Base_substitutivity_splits_original_features_train_substitutivity_splits_original_features_test1/checkpoint-3320/README.md b/substitutivity_original_Estonian/Qwen3.5-2B-Base_substitutivity_splits_original_features_train_substitutivity_splits_original_features_test1/checkpoint-3320/README.md deleted file mode 100644 index ece204d005f7bf161d0e448efdd56fd297b7c8a2..0000000000000000000000000000000000000000 --- a/substitutivity_original_Estonian/Qwen3.5-2B-Base_substitutivity_splits_original_features_train_substitutivity_splits_original_features_test1/checkpoint-3320/README.md +++ /dev/null @@ -1,209 +0,0 @@ ---- -base_model: Qwen/Qwen3.5-2B-Base -library_name: peft -pipeline_tag: text-generation -tags: -- base_model:adapter:Qwen/Qwen3.5-2B-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/substitutivity_original_Estonian/Qwen3.5-2B-Base_substitutivity_splits_original_features_train_substitutivity_splits_original_features_test1/checkpoint-3320/adapter_config.json b/substitutivity_original_Estonian/Qwen3.5-2B-Base_substitutivity_splits_original_features_train_substitutivity_splits_original_features_test1/checkpoint-3320/adapter_config.json deleted file mode 100644 index a64145ce23adcbdbe246679087a13f4cb63ff664..0000000000000000000000000000000000000000 --- a/substitutivity_original_Estonian/Qwen3.5-2B-Base_substitutivity_splits_original_features_train_substitutivity_splits_original_features_test1/checkpoint-3320/adapter_config.json +++ /dev/null @@ -1,46 +0,0 @@ -{ - "alora_invocation_tokens": null, - "alpha_pattern": {}, - "arrow_config": null, - "auto_mapping": null, - "base_model_name_or_path": "Qwen/Qwen3.5-2B-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": 256, - "lora_bias": false, - "lora_dropout": 0.027050353915701653, - "megatron_config": null, - "megatron_core": "megatron.core", - "modules_to_save": null, - "peft_type": "LORA", - "peft_version": "0.18.1", - "qalora_group_size": 16, - "r": 128, - "rank_pattern": {}, - "revision": null, - "target_modules": [ - "o_proj", - "q_proj", - "k_proj", - "down_proj", - "up_proj", - "v_proj", - "gate_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/substitutivity_original_Estonian/Qwen3.5-2B-Base_substitutivity_splits_original_features_train_substitutivity_splits_original_features_test1/checkpoint-3320/adapter_model.safetensors b/substitutivity_original_Estonian/Qwen3.5-2B-Base_substitutivity_splits_original_features_train_substitutivity_splits_original_features_test1/checkpoint-3320/adapter_model.safetensors deleted file mode 100644 index 364e4ec699f1e3bcb1077f30974001234705fd6e..0000000000000000000000000000000000000000 --- a/substitutivity_original_Estonian/Qwen3.5-2B-Base_substitutivity_splits_original_features_train_substitutivity_splits_original_features_test1/checkpoint-3320/adapter_model.safetensors +++ /dev/null @@ -1,3 +0,0 @@ -version https://git-lfs.github.com/spec/v1 -oid sha256:eb103a11b545a6a1bfadb00e9f36fe586c11850287cbaf9f1a507a083a379f5a -size 349201608 diff --git a/substitutivity_original_Estonian/Qwen3.5-2B-Base_substitutivity_splits_original_features_train_substitutivity_splits_original_features_test1/checkpoint-3320/chat_template.jinja b/substitutivity_original_Estonian/Qwen3.5-2B-Base_substitutivity_splits_original_features_train_substitutivity_splits_original_features_test1/checkpoint-3320/chat_template.jinja deleted file mode 100644 index 0ef09f214eaa6d9bca297988afc1454b5827b2c7..0000000000000000000000000000000000000000 --- a/substitutivity_original_Estonian/Qwen3.5-2B-Base_substitutivity_splits_original_features_train_substitutivity_splits_original_features_test1/checkpoint-3320/chat_template.jinja +++ /dev/null @@ -1,154 +0,0 @@ -{%- set image_count = namespace(value=0) %} -{%- set video_count = namespace(value=0) %} -{%- macro render_content(content, do_vision_count, is_system_content=false) %} - {%- if content is string %} - {{- content }} - {%- elif content is iterable and content is not mapping %} - {%- for item in content %} - {%- if 'image' in item or 'image_url' in item or item.type == 'image' %} - {%- if is_system_content %} - {{- raise_exception('System message cannot contain images.') }} - {%- endif %} - {%- if do_vision_count %} - {%- set image_count.value = image_count.value + 1 %} - {%- endif %} - {%- if add_vision_id %} - {{- 'Picture ' ~ image_count.value ~ ': ' }} - {%- endif %} - {{- '<|vision_start|><|image_pad|><|vision_end|>' }} - {%- elif 'video' in item or item.type == 'video' %} - {%- if is_system_content %} - {{- raise_exception('System message cannot contain videos.') }} - {%- endif %} - {%- if do_vision_count %} - {%- set video_count.value = video_count.value + 1 %} - {%- endif %} - {%- if add_vision_id %} - {{- 'Video ' ~ video_count.value ~ ': ' }} - {%- endif %} - {{- '<|vision_start|><|video_pad|><|vision_end|>' }} - {%- elif 'text' in item %} - {{- item.text }} - {%- else %} - {{- raise_exception('Unexpected item type in content.') }} - {%- endif %} - {%- endfor %} - {%- elif content is none or content is undefined %} - {{- '' }} - {%- else %} - {{- raise_exception('Unexpected content type.') }} - {%- endif %} -{%- endmacro %} -{%- if not messages %} - {{- raise_exception('No messages provided.') }} -{%- endif %} -{%- if tools and tools is iterable and tools is not mapping %} - {{- '<|im_start|>system\n' }} - {{- "# Tools\n\nYou have access to the following functions:\n\n" }} - {%- for tool in tools %} - {{- "\n" }} - {{- tool | tojson }} - {%- endfor %} - {{- "\n" }} - {{- '\n\nIf you choose to call a function ONLY reply in the following format with NO suffix:\n\n\n\n\nvalue_1\n\n\nThis is the value for the second parameter\nthat can span\nmultiple lines\n\n\n\n\n\nReminder:\n- Function calls MUST follow the specified format: an inner block must be nested within XML tags\n- Required parameters MUST be specified\n- You may provide optional reasoning for your function call in natural language BEFORE the function call, but NOT after\n- If there is no function call available, answer the question like normal with your current knowledge and do not tell the user about function calls\n' }} - {%- if messages[0].role == 'system' %} - {%- set content = render_content(messages[0].content, false, true)|trim %} - {%- if content %} - {{- '\n\n' + content }} - {%- endif %} - {%- endif %} - {{- '<|im_end|>\n' }} -{%- else %} - {%- if messages[0].role == 'system' %} - {%- set content = render_content(messages[0].content, false, true)|trim %} - {{- '<|im_start|>system\n' + 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" %} - {%- set content = render_content(message.content, false)|trim %} - {%- if not(content.startswith('') and content.endswith('')) %} - {%- set ns.multi_step_tool = false %} - {%- set ns.last_query_index = index %} - {%- endif %} - {%- endif %} -{%- endfor %} -{%- if ns.multi_step_tool %} - {{- raise_exception('No user query found in messages.') }} -{%- endif %} -{%- for message in messages %} - {%- set content = render_content(message.content, true)|trim %} - {%- if message.role == "system" %} - {%- if not loop.first %} - {{- raise_exception('System message must be at the beginning.') }} - {%- endif %} - {%- elif message.role == "user" %} - {{- '<|im_start|>' + message.role + '\n' + content + '<|im_end|>' + '\n' }} - {%- elif message.role == "assistant" %} - {%- set reasoning_content = '' %} - {%- if message.reasoning_content is string %} - {%- set reasoning_content = message.reasoning_content %} - {%- else %} - {%- if '' in content %} - {%- set reasoning_content = content.split('')[0].rstrip('\n').split('')[-1].lstrip('\n') %} - {%- set content = content.split('')[-1].lstrip('\n') %} - {%- endif %} - {%- endif %} - {%- set reasoning_content = reasoning_content|trim %} - {%- if loop.index0 > ns.last_query_index %} - {{- '<|im_start|>' + message.role + '\n\n' + reasoning_content + '\n\n\n' + content }} - {%- else %} - {{- '<|im_start|>' + message.role + '\n' + content }} - {%- endif %} - {%- if message.tool_calls and message.tool_calls is iterable and message.tool_calls is not mapping %} - {%- for tool_call in message.tool_calls %} - {%- if tool_call.function is defined %} - {%- set tool_call = tool_call.function %} - {%- endif %} - {%- if loop.first %} - {%- if content|trim %} - {{- '\n\n\n\n' }} - {%- else %} - {{- '\n\n' }} - {%- endif %} - {%- else %} - {{- '\n\n\n' }} - {%- endif %} - {%- if tool_call.arguments is defined %} - {%- for args_name, args_value in tool_call.arguments|items %} - {{- '\n' }} - {%- set args_value = args_value | tojson | safe if args_value is mapping or (args_value is sequence and args_value is not string) else args_value | string %} - {{- args_value }} - {{- '\n\n' }} - {%- endfor %} - {%- endif %} - {{- '\n' }} - {%- endfor %} - {%- endif %} - {{- '<|im_end|>\n' }} - {%- elif message.role == "tool" %} - {%- if loop.previtem and loop.previtem.role != "tool" %} - {{- '<|im_start|>user' }} - {%- endif %} - {{- '\n\n' }} - {{- content }} - {{- '\n' }} - {%- if not loop.last and loop.nextitem.role != "tool" %} - {{- '<|im_end|>\n' }} - {%- elif loop.last %} - {{- '<|im_end|>\n' }} - {%- endif %} - {%- else %} - {{- raise_exception('Unexpected message role.') }} - {%- endif %} -{%- endfor %} -{%- if add_generation_prompt %} - {{- '<|im_start|>assistant\n' }} - {%- if enable_thinking is defined and enable_thinking is true %} - {{- '\n' }} - {%- else %} - {{- '\n\n\n\n' }} - {%- endif %} -{%- endif %} \ No newline at end of file diff --git a/substitutivity_original_Estonian/Qwen3.5-2B-Base_substitutivity_splits_original_features_train_substitutivity_splits_original_features_test1/checkpoint-3320/tokenizer.json b/substitutivity_original_Estonian/Qwen3.5-2B-Base_substitutivity_splits_original_features_train_substitutivity_splits_original_features_test1/checkpoint-3320/tokenizer.json deleted file mode 100644 index 67741b04f23bfdb46501f748ce27865ec82eccfb..0000000000000000000000000000000000000000 --- a/substitutivity_original_Estonian/Qwen3.5-2B-Base_substitutivity_splits_original_features_train_substitutivity_splits_original_features_test1/checkpoint-3320/tokenizer.json +++ /dev/null @@ -1,3 +0,0 @@ -version https://git-lfs.github.com/spec/v1 -oid sha256:87a7830d63fcf43bf241c3c5242e96e62dd3fdc29224ca26fed8ea333db72de4 -size 19989343 diff --git a/substitutivity_original_Estonian/Qwen3.5-2B-Base_substitutivity_splits_original_features_train_substitutivity_splits_original_features_test1/checkpoint-3320/tokenizer_config.json b/substitutivity_original_Estonian/Qwen3.5-2B-Base_substitutivity_splits_original_features_train_substitutivity_splits_original_features_test1/checkpoint-3320/tokenizer_config.json deleted file mode 100644 index b4a37b2a6fd3ab3317cd7bac72855be1a843b2bb..0000000000000000000000000000000000000000 --- a/substitutivity_original_Estonian/Qwen3.5-2B-Base_substitutivity_splits_original_features_train_substitutivity_splits_original_features_test1/checkpoint-3320/tokenizer_config.json +++ /dev/null @@ -1,31 +0,0 @@ -{ - "add_prefix_space": false, - "audio_bos_token": "<|audio_start|>", - "audio_eos_token": "<|audio_end|>", - "audio_token": "<|audio_pad|>", - "backend": "tokenizers", - "bos_token": null, - "clean_up_tokenization_spaces": false, - "eos_token": "<|endoftext|>", - "errors": "replace", - "image_token": "<|image_pad|>", - "is_local": false, - "model_max_length": 262144, - "model_specific_special_tokens": { - "audio_bos_token": "<|audio_start|>", - "audio_eos_token": "<|audio_end|>", - "audio_token": "<|audio_pad|>", - "image_token": "<|image_pad|>", - "video_token": "<|video_pad|>", - "vision_bos_token": "<|vision_start|>", - "vision_eos_token": "<|vision_end|>" - }, - "pad_token": "<|endoftext|>", - "pretokenize_regex": "(?i:'s|'t|'re|'ve|'m|'ll|'d)|[^\\r\\n\\p{L}\\p{N}]?[\\p{L}\\p{M}]+|\\p{N}| ?[^\\s\\p{L}\\p{M}\\p{N}]+[\\r\\n]*|\\s*[\\r\\n]+|\\s+(?!\\S)|\\s+", - "split_special_tokens": false, - "tokenizer_class": "TokenizersBackend", - "unk_token": null, - "video_token": "<|video_pad|>", - "vision_bos_token": "<|vision_start|>", - "vision_eos_token": "<|vision_end|>" -} diff --git a/substitutivity_original_Estonian/Qwen3.5-2B-Base_substitutivity_splits_original_features_train_substitutivity_splits_original_features_test1/checkpoint-3320/trainer_state.json b/substitutivity_original_Estonian/Qwen3.5-2B-Base_substitutivity_splits_original_features_train_substitutivity_splits_original_features_test1/checkpoint-3320/trainer_state.json deleted file mode 100644 index 512082fd48302333ff52c73452f60c2fa54f259e..0000000000000000000000000000000000000000 --- a/substitutivity_original_Estonian/Qwen3.5-2B-Base_substitutivity_splits_original_features_train_substitutivity_splits_original_features_test1/checkpoint-3320/trainer_state.json +++ /dev/null @@ -1,870 +0,0 @@ -{ - "best_global_step": null, - "best_metric": null, - "best_model_checkpoint": null, - "epoch": 8.0, - "eval_steps": 500, - "global_step": 3320, - "is_hyper_param_search": false, - "is_local_process_zero": true, - "is_world_process_zero": true, - "log_history": [ - { - "entropy": 2.0473356211185454, - "epoch": 0.12062726176115803, - "grad_norm": 3.0554118156433105, - "learning_rate": 1.5682344549790923e-05, - "loss": 1.910550079345703, - "mean_token_accuracy": 0.6053373458981514, - "num_tokens": 121428.0, - "step": 50 - }, - { - "entropy": 1.0371007251739501, - "epoch": 0.24125452352231605, - "grad_norm": 2.109868049621582, - "learning_rate": 3.168473694753677e-05, - "loss": 0.9585584259033203, - "mean_token_accuracy": 0.7352996730804443, - "num_tokens": 243966.0, - "step": 100 - }, - { - "entropy": 0.9105490958690643, - "epoch": 0.3618817852834741, - "grad_norm": 1.7675701379776, - "learning_rate": 4.7687129345282605e-05, - "loss": 0.8393524932861328, - "mean_token_accuracy": 0.7588736474514007, - "num_tokens": 366649.0, - "step": 150 - }, - { - "entropy": 0.8660378754138947, - "epoch": 0.4825090470446321, - "grad_norm": 1.6502987146377563, - "learning_rate": 6.368952174302845e-05, - "loss": 0.8005844116210937, - "mean_token_accuracy": 0.7663125538825989, - "num_tokens": 488367.0, - "step": 200 - }, - { - "entropy": 0.8439123743772506, - "epoch": 0.6031363088057901, - "grad_norm": 1.538365364074707, - "learning_rate": 7.969191414077429e-05, - "loss": 0.7773619842529297, - "mean_token_accuracy": 0.7718527519702911, - "num_tokens": 603131.0, - "step": 250 - }, - { - "entropy": 0.8223806774616241, - "epoch": 0.7237635705669482, - "grad_norm": 1.1559664011001587, - "learning_rate": 9.569430653852014e-05, - "loss": 0.7587767791748047, - "mean_token_accuracy": 0.7767325991392136, - "num_tokens": 726412.0, - "step": 300 - }, - { - "entropy": 0.8256935775279999, - "epoch": 0.8443908323281062, - "grad_norm": 1.2322715520858765, - "learning_rate": 0.00011169669893626597, - "loss": 0.769886245727539, - "mean_token_accuracy": 0.773840668797493, - "num_tokens": 846516.0, - "step": 350 - }, - { - "entropy": 0.8065843790769577, - "epoch": 0.9650180940892642, - "grad_norm": 0.9908286333084106, - "learning_rate": 0.00012769909133401183, - "loss": 0.7451084899902344, - "mean_token_accuracy": 0.7794792425632476, - "num_tokens": 958978.0, - "step": 400 - }, - { - "epoch": 1.0, - "eval_entropy": 0.791056221790528, - "eval_mean_token_accuracy": 0.7507295300451557, - "eval_not_syn_loss": 0.7860428094863892, - "eval_not_syn_runtime": 53.5431, - "eval_not_syn_samples_per_second": 26.521, - "eval_not_syn_steps_per_second": 3.324, - "eval_num_tokens": 993806.0, - "step": 415 - }, - { - "epoch": 1.0, - "eval_entropy": 0.7424047408478983, - "eval_mean_token_accuracy": 0.8023804037758474, - "eval_num_tokens": 993806.0, - "eval_syn_loss": 0.7273509502410889, - "eval_syn_runtime": 54.8679, - "eval_syn_samples_per_second": 25.88, - "eval_syn_steps_per_second": 3.244, - "step": 415 - }, - { - "entropy": 0.7607260761838971, - "epoch": 1.0844390832328106, - "grad_norm": 1.0653696060180664, - "learning_rate": 0.0001327927019027069, - "loss": 0.7034613037109375, - "mean_token_accuracy": 0.7869195998317063, - "num_tokens": 1080301.0, - "step": 450 - }, - { - "entropy": 0.7370894801616669, - "epoch": 1.2050663449939687, - "grad_norm": 1.763787865638733, - "learning_rate": 0.00013265416569994628, - "loss": 0.6816973114013671, - "mean_token_accuracy": 0.7914507347345352, - "num_tokens": 1201215.0, - "step": 500 - }, - { - "entropy": 0.742829909324646, - "epoch": 1.3256936067551266, - "grad_norm": 0.97663414478302, - "learning_rate": 0.00013239847940553975, - "loss": 0.6867338562011719, - "mean_token_accuracy": 0.790328289270401, - "num_tokens": 1322895.0, - "step": 550 - }, - { - "entropy": 0.7312483406066894, - "epoch": 1.4463208685162847, - "grad_norm": 1.0874731540679932, - "learning_rate": 0.00013202609518972978, - "loss": 0.676800537109375, - "mean_token_accuracy": 0.7921512228250503, - "num_tokens": 1436749.0, - "step": 600 - }, - { - "entropy": 0.7178869074583054, - "epoch": 1.5669481302774426, - "grad_norm": 1.064300298690796, - "learning_rate": 0.0001315376715980265, - "loss": 0.6635584259033203, - "mean_token_accuracy": 0.7944776183366775, - "num_tokens": 1555400.0, - "step": 650 - }, - { - "entropy": 0.7198166918754577, - "epoch": 1.6875753920386007, - "grad_norm": 0.7779315114021301, - "learning_rate": 0.00013093407238659802, - "loss": 0.6660567474365234, - "mean_token_accuracy": 0.7968300533294678, - "num_tokens": 1676892.0, - "step": 700 - }, - { - "entropy": 0.7170061504840851, - "epoch": 1.8082026537997589, - "grad_norm": 1.089854121208191, - "learning_rate": 0.00013021636499475496, - "loss": 0.6632771301269531, - "mean_token_accuracy": 0.7963915091753005, - "num_tokens": 1802689.0, - "step": 750 - }, - { - "entropy": 0.6953610718250275, - "epoch": 1.9288299155609168, - "grad_norm": 0.7996730208396912, - "learning_rate": 0.0001293858186572295, - "loss": 0.6421419525146485, - "mean_token_accuracy": 0.8002577543258667, - "num_tokens": 1920804.0, - "step": 800 - }, - { - "epoch": 2.0, - "eval_entropy": 0.7369416523515508, - "eval_mean_token_accuracy": 0.7872022366925572, - "eval_not_syn_loss": 0.713637113571167, - "eval_not_syn_runtime": 53.2265, - "eval_not_syn_samples_per_second": 26.678, - "eval_not_syn_steps_per_second": 3.344, - "eval_num_tokens": 1987612.0, - "step": 830 - }, - { - "epoch": 2.0, - "eval_entropy": 0.6877032756135705, - "eval_mean_token_accuracy": 0.7997466372639945, - "eval_num_tokens": 1987612.0, - "eval_syn_loss": 0.6614266633987427, - "eval_syn_runtime": 54.5178, - "eval_syn_samples_per_second": 26.047, - "eval_syn_steps_per_second": 3.265, - "step": 830 - }, - { - "entropy": 0.6644553433765065, - "epoch": 2.0482509047044632, - "grad_norm": 0.7562468647956848, - "learning_rate": 0.00012844390215958852, - "loss": 0.6112185287475586, - "mean_token_accuracy": 0.806649933559726, - "num_tokens": 2039660.0, - "step": 850 - }, - { - "entropy": 0.6146580249071121, - "epoch": 2.168878166465621, - "grad_norm": 1.098607063293457, - "learning_rate": 0.00012739228124074913, - "loss": 0.5553749084472657, - "mean_token_accuracy": 0.8187588346004486, - "num_tokens": 2161772.0, - "step": 900 - }, - { - "entropy": 0.6187627577781677, - "epoch": 2.2895054282267795, - "grad_norm": 1.0338040590286255, - "learning_rate": 0.00012623281564719095, - "loss": 0.5596194076538086, - "mean_token_accuracy": 0.8182566928863525, - "num_tokens": 2280833.0, - "step": 950 - }, - { - "entropy": 0.6232365870475769, - "epoch": 2.4101326899879374, - "grad_norm": 1.0010799169540405, - "learning_rate": 0.00012496755584407422, - "loss": 0.5667143249511719, - "mean_token_accuracy": 0.8175061982870102, - "num_tokens": 2398206.0, - "step": 1000 - }, - { - "entropy": 0.6151268655061721, - "epoch": 2.5307599517490953, - "grad_norm": 1.0065034627914429, - "learning_rate": 0.0001235987393890801, - "loss": 0.5598890686035156, - "mean_token_accuracy": 0.8193449324369431, - "num_tokens": 2518888.0, - "step": 1050 - }, - { - "entropy": 0.6173626267910004, - "epoch": 2.651387213510253, - "grad_norm": 0.7958805561065674, - "learning_rate": 0.00012212878697538592, - "loss": 0.5649292755126953, - "mean_token_accuracy": 0.8172902607917786, - "num_tokens": 2639763.0, - "step": 1100 - }, - { - "entropy": 0.6264930808544159, - "epoch": 2.772014475271411, - "grad_norm": 0.7718372344970703, - "learning_rate": 0.00012056029815077323, - "loss": 0.5716455078125, - "mean_token_accuracy": 0.8170186239480972, - "num_tokens": 2755540.0, - "step": 1150 - }, - { - "entropy": 0.6203387409448624, - "epoch": 2.8926417370325694, - "grad_norm": 0.7914765477180481, - "learning_rate": 0.000118896046720439, - "loss": 0.5653961944580078, - "mean_token_accuracy": 0.8170280295610428, - "num_tokens": 2874118.0, - "step": 1200 - }, - { - "epoch": 3.0, - "eval_entropy": 0.6751241854737314, - "eval_mean_token_accuracy": 0.792226163858778, - "eval_not_syn_loss": 0.6902772784233093, - "eval_not_syn_runtime": 53.2557, - "eval_not_syn_samples_per_second": 26.664, - "eval_not_syn_steps_per_second": 3.342, - "eval_num_tokens": 2981418.0, - "step": 1245 - }, - { - "epoch": 3.0, - "eval_entropy": 0.6308753868837035, - "eval_mean_token_accuracy": 0.8131611759073278, - "eval_num_tokens": 2981418.0, - "eval_syn_loss": 0.6440441012382507, - "eval_syn_runtime": 54.7064, - "eval_syn_samples_per_second": 25.957, - "eval_syn_steps_per_second": 3.254, - "step": 1245 - }, - { - "entropy": 0.6077835421369533, - "epoch": 3.012062726176116, - "grad_norm": 1.124720573425293, - "learning_rate": 0.00011713897584164003, - "loss": 0.5557136917114258, - "mean_token_accuracy": 0.8203362726201915, - "num_tokens": 2994602.0, - "step": 1250 - }, - { - "entropy": 0.507642827630043, - "epoch": 3.132689987937274, - "grad_norm": 0.9039916396141052, - "learning_rate": 0.00011529219281884579, - "loss": 0.4467372512817383, - "mean_token_accuracy": 0.8464260965585708, - "num_tokens": 3121094.0, - "step": 1300 - }, - { - "entropy": 0.5137191081047058, - "epoch": 3.2533172496984317, - "grad_norm": 0.8580691814422607, - "learning_rate": 0.00011335896360860367, - "loss": 0.4538694763183594, - "mean_token_accuracy": 0.8448719161748887, - "num_tokens": 3248038.0, - "step": 1350 - }, - { - "entropy": 0.5293738862872124, - "epoch": 3.37394451145959, - "grad_norm": 1.0177273750305176, - "learning_rate": 0.00011134270704383492, - "loss": 0.4639348220825195, - "mean_token_accuracy": 0.8428975892066956, - "num_tokens": 3363487.0, - "step": 1400 - }, - { - "entropy": 0.5218872320652008, - "epoch": 3.494571773220748, - "grad_norm": 1.0530673265457153, - "learning_rate": 0.00010924698878777538, - "loss": 0.4584346389770508, - "mean_token_accuracy": 0.8432288640737533, - "num_tokens": 3480513.0, - "step": 1450 - }, - { - "entropy": 0.5320372316241264, - "epoch": 3.615199034981906, - "grad_norm": 1.0291194915771484, - "learning_rate": 0.0001070755150282529, - "loss": 0.4711898422241211, - "mean_token_accuracy": 0.842066935300827, - "num_tokens": 3597183.0, - "step": 1500 - }, - { - "entropy": 0.5315000694990158, - "epoch": 3.7358262967430638, - "grad_norm": 0.9788838624954224, - "learning_rate": 0.00010483212592345332, - "loss": 0.4729101943969727, - "mean_token_accuracy": 0.8404586082696914, - "num_tokens": 3714173.0, - "step": 1550 - }, - { - "entropy": 0.5331643190979958, - "epoch": 3.856453558504222, - "grad_norm": 1.1248903274536133, - "learning_rate": 0.00010252078881076532, - "loss": 0.47405517578125, - "mean_token_accuracy": 0.8397526168823242, - "num_tokens": 3832652.0, - "step": 1600 - }, - { - "entropy": 0.5343865439295769, - "epoch": 3.97708082026538, - "grad_norm": 0.9008272886276245, - "learning_rate": 0.00010014559119071452, - "loss": 0.47172779083251953, - "mean_token_accuracy": 0.8404951041936874, - "num_tokens": 3955167.0, - "step": 1650 - }, - { - "epoch": 4.0, - "eval_entropy": 0.5736920103598177, - "eval_mean_token_accuracy": 0.7981750027517255, - "eval_not_syn_loss": 0.707631528377533, - "eval_not_syn_runtime": 52.2858, - "eval_not_syn_samples_per_second": 27.158, - "eval_not_syn_steps_per_second": 3.404, - "eval_num_tokens": 3975224.0, - "step": 1660 - }, - { - "epoch": 4.0, - "eval_entropy": 0.5363499959198277, - "eval_mean_token_accuracy": 0.8063885539435269, - "eval_num_tokens": 3975224.0, - "eval_syn_loss": 0.6610181331634521, - "eval_syn_runtime": 54.2167, - "eval_syn_samples_per_second": 26.191, - "eval_syn_steps_per_second": 3.283, - "step": 1660 - }, - { - "entropy": 0.4512602482179199, - "epoch": 4.0965018094089265, - "grad_norm": 0.8079906105995178, - "learning_rate": 9.771073349839416e-05, - "loss": 0.3778452682495117, - "mean_token_accuracy": 0.8684436221315404, - "num_tokens": 4072361.0, - "step": 1700 - }, - { - "entropy": 0.4228873434662819, - "epoch": 4.217129071170085, - "grad_norm": 1.1206746101379395, - "learning_rate": 9.52205216751758e-05, - "loss": 0.34671127319335937, - "mean_token_accuracy": 0.8763950079679489, - "num_tokens": 4194978.0, - "step": 1750 - }, - { - "entropy": 0.42808119297027586, - "epoch": 4.337756332931242, - "grad_norm": 1.1268125772476196, - "learning_rate": 9.26793595538368e-05, - "loss": 0.3484600830078125, - "mean_token_accuracy": 0.8753323918581009, - "num_tokens": 4315548.0, - "step": 1800 - }, - { - "entropy": 0.42506767451763156, - "epoch": 4.458383594692401, - "grad_norm": 1.1051807403564453, - "learning_rate": 9.009174107057088e-05, - "loss": 0.35434780120849607, - "mean_token_accuracy": 0.8734411865472793, - "num_tokens": 4431469.0, - "step": 1850 - }, - { - "entropy": 0.43314908266067503, - "epoch": 4.579010856453559, - "grad_norm": 0.9782640337944031, - "learning_rate": 8.746224231765484e-05, - "loss": 0.36225860595703124, - "mean_token_accuracy": 0.8715140378475189, - "num_tokens": 4551803.0, - "step": 1900 - }, - { - "entropy": 0.42984022706747055, - "epoch": 4.699638118214716, - "grad_norm": 0.9772566556930542, - "learning_rate": 8.479551345082568e-05, - "loss": 0.356609001159668, - "mean_token_accuracy": 0.8717925781011582, - "num_tokens": 4674468.0, - "step": 1950 - }, - { - "entropy": 0.4474307084083557, - "epoch": 4.820265379975875, - "grad_norm": 0.9547265768051147, - "learning_rate": 8.209627046567966e-05, - "loss": 0.3654902267456055, - "mean_token_accuracy": 0.8724055212736129, - "num_tokens": 4788027.0, - "step": 2000 - }, - { - "entropy": 0.4377285006642342, - "epoch": 4.940892641737032, - "grad_norm": 1.0390652418136597, - "learning_rate": 7.936928685763635e-05, - "loss": 0.3645761871337891, - "mean_token_accuracy": 0.8707609891891479, - "num_tokens": 4907561.0, - "step": 2050 - }, - { - "epoch": 5.0, - "eval_entropy": 0.5173388388049737, - "eval_mean_token_accuracy": 0.7918405432379647, - "eval_not_syn_loss": 0.7532826662063599, - "eval_not_syn_runtime": 53.0766, - "eval_not_syn_samples_per_second": 26.754, - "eval_not_syn_steps_per_second": 3.354, - "eval_num_tokens": 4969030.0, - "step": 2075 - }, - { - "epoch": 5.0, - "eval_entropy": 0.4831916272975086, - "eval_mean_token_accuracy": 0.8204237202580056, - "eval_num_tokens": 4969030.0, - "eval_syn_loss": 0.6971890330314636, - "eval_syn_runtime": 54.8755, - "eval_syn_samples_per_second": 25.877, - "eval_syn_steps_per_second": 3.244, - "step": 2075 - }, - { - "entropy": 0.3804885711934831, - "epoch": 5.060313630880579, - "grad_norm": 0.8717736005783081, - "learning_rate": 7.661938518021687e-05, - "loss": 0.3034685707092285, - "mean_token_accuracy": 0.8910005471923135, - "num_tokens": 5026625.0, - "step": 2100 - }, - { - "entropy": 0.3174623078107834, - "epoch": 5.180940892641737, - "grad_norm": 0.9808429479598999, - "learning_rate": 7.385142851656478e-05, - "loss": 0.23829608917236328, - "mean_token_accuracy": 0.9131860065460206, - "num_tokens": 5141010.0, - "step": 2150 - }, - { - "entropy": 0.3187542732059956, - "epoch": 5.301568154402895, - "grad_norm": 1.3739993572235107, - "learning_rate": 7.107031187929228e-05, - "loss": 0.24024734497070313, - "mean_token_accuracy": 0.910272576212883, - "num_tokens": 5261240.0, - "step": 2200 - }, - { - "entropy": 0.31624249830842016, - "epoch": 5.422195416164053, - "grad_norm": 1.1618341207504272, - "learning_rate": 6.82809535538603e-05, - "loss": 0.24144052505493163, - "mean_token_accuracy": 0.9122701978683472, - "num_tokens": 5382160.0, - "step": 2250 - }, - { - "entropy": 0.31793304443359377, - "epoch": 5.542822677925211, - "grad_norm": 1.4960980415344238, - "learning_rate": 6.548828640080172e-05, - "loss": 0.24051647186279296, - "mean_token_accuracy": 0.9116496992111206, - "num_tokens": 5502002.0, - "step": 2300 - }, - { - "entropy": 0.3180307722091675, - "epoch": 5.66344993968637, - "grad_norm": 1.103051781654358, - "learning_rate": 6.269724913216913e-05, - "loss": 0.24294727325439452, - "mean_token_accuracy": 0.9102917027473449, - "num_tokens": 5629969.0, - "step": 2350 - }, - { - "entropy": 0.32020670652389527, - "epoch": 5.784077201447527, - "grad_norm": 1.0690243244171143, - "learning_rate": 5.991277757763442e-05, - "loss": 0.24328670501708985, - "mean_token_accuracy": 0.9101310813426972, - "num_tokens": 5750596.0, - "step": 2400 - }, - { - "entropy": 0.31846788004040716, - "epoch": 5.904704463208685, - "grad_norm": 1.203026294708252, - "learning_rate": 5.7139795955685784e-05, - "loss": 0.244984130859375, - "mean_token_accuracy": 0.9100374561548233, - "num_tokens": 5868110.0, - "step": 2450 - }, - { - "epoch": 6.0, - "eval_entropy": 0.45794740654109567, - "eval_mean_token_accuracy": 0.7891001142142864, - "eval_not_syn_loss": 0.8338615894317627, - "eval_not_syn_runtime": 53.3841, - "eval_not_syn_samples_per_second": 26.6, - "eval_not_syn_steps_per_second": 3.334, - "eval_num_tokens": 5962836.0, - "step": 2490 - }, - { - "epoch": 6.0, - "eval_entropy": 0.4288232589370749, - "eval_mean_token_accuracy": 0.8114104933953017, - "eval_num_tokens": 5962836.0, - "eval_syn_loss": 0.7791845798492432, - "eval_syn_runtime": 54.7877, - "eval_syn_samples_per_second": 25.918, - "eval_syn_steps_per_second": 3.249, - "step": 2490 - }, - { - "entropy": 0.2972855128423132, - "epoch": 6.024125452352232, - "grad_norm": 1.1748684644699097, - "learning_rate": 5.438320816535859e-05, - "loss": 0.22051204681396486, - "mean_token_accuracy": 0.9190925630656156, - "num_tokens": 5990107.0, - "step": 2500 - }, - { - "entropy": 0.2216818816959858, - "epoch": 6.144752714113389, - "grad_norm": 1.0899405479431152, - "learning_rate": 5.164788911390022e-05, - "loss": 0.1487915325164795, - "mean_token_accuracy": 0.9462692165374755, - "num_tokens": 6104230.0, - "step": 2550 - }, - { - "entropy": 0.22189782992005347, - "epoch": 6.265379975874548, - "grad_norm": 1.0967661142349243, - "learning_rate": 4.8938676095705914e-05, - "loss": 0.149844388961792, - "mean_token_accuracy": 0.9452104204893113, - "num_tokens": 6223548.0, - "step": 2600 - }, - { - "entropy": 0.2155836282670498, - "epoch": 6.386007237635706, - "grad_norm": 1.1443830728530884, - "learning_rate": 4.626036023777093e-05, - "loss": 0.1472412109375, - "mean_token_accuracy": 0.946697844862938, - "num_tokens": 6349102.0, - "step": 2650 - }, - { - "entropy": 0.22189507275819778, - "epoch": 6.5066344993968634, - "grad_norm": 1.2138874530792236, - "learning_rate": 4.3617678026788244e-05, - "loss": 0.15155397415161131, - "mean_token_accuracy": 0.9441326081752777, - "num_tokens": 6468966.0, - "step": 2700 - }, - { - "entropy": 0.22405965447425844, - "epoch": 6.627261761158022, - "grad_norm": 1.2444716691970825, - "learning_rate": 4.10153029328748e-05, - "loss": 0.15516298294067382, - "mean_token_accuracy": 0.9442053025960923, - "num_tokens": 6583296.0, - "step": 2750 - }, - { - "entropy": 0.2216674068570137, - "epoch": 6.74788902291918, - "grad_norm": 1.0056185722351074, - "learning_rate": 3.845783714474033e-05, - "loss": 0.15423223495483399, - "mean_token_accuracy": 0.9445771616697312, - "num_tokens": 6698696.0, - "step": 2800 - }, - { - "entropy": 0.21455578804016112, - "epoch": 6.868516284680338, - "grad_norm": 1.0748720169067383, - "learning_rate": 3.59498034309138e-05, - "loss": 0.14870482444763183, - "mean_token_accuracy": 0.9457976347208024, - "num_tokens": 6819711.0, - "step": 2850 - }, - { - "entropy": 0.21227611482143402, - "epoch": 6.989143546441496, - "grad_norm": 0.978204607963562, - "learning_rate": 3.349563714142155e-05, - "loss": 0.14802613258361816, - "mean_token_accuracy": 0.9466447854042053, - "num_tokens": 6944941.0, - "step": 2900 - }, - { - "epoch": 7.0, - "eval_entropy": 0.38817436741978933, - "eval_mean_token_accuracy": 0.7801904045463948, - "eval_not_syn_loss": 0.9901206493377686, - "eval_not_syn_runtime": 53.1137, - "eval_not_syn_samples_per_second": 26.735, - "eval_not_syn_steps_per_second": 3.351, - "eval_num_tokens": 6956642.0, - "step": 2905 - }, - { - "epoch": 7.0, - "eval_entropy": 0.3640736076938972, - "eval_mean_token_accuracy": 0.8158511425671953, - "eval_num_tokens": 6956642.0, - "eval_syn_loss": 0.8903024792671204, - "eval_syn_runtime": 54.3768, - "eval_syn_samples_per_second": 26.114, - "eval_syn_steps_per_second": 3.273, - "step": 2905 - }, - { - "entropy": 0.16986672166321012, - "epoch": 7.108564535585042, - "grad_norm": 0.7068538069725037, - "learning_rate": 3.1099678364060876e-05, - "loss": 0.10039780616760254, - "mean_token_accuracy": 0.964904587678235, - "num_tokens": 7064818.0, - "step": 2950 - }, - { - "entropy": 0.16062987893819808, - "epoch": 7.2291917973462, - "grad_norm": 0.963703453540802, - "learning_rate": 2.8766164249141098e-05, - "loss": 0.09831766128540038, - "mean_token_accuracy": 0.9666522133350373, - "num_tokens": 7183624.0, - "step": 3000 - }, - { - "entropy": 0.15993704989552499, - "epoch": 7.349819059107358, - "grad_norm": 0.8077407479286194, - "learning_rate": 2.6499221516264785e-05, - "loss": 0.09768331527709961, - "mean_token_accuracy": 0.9664828568696976, - "num_tokens": 7304122.0, - "step": 3050 - }, - { - "entropy": 0.16501803942024706, - "epoch": 7.470446320868517, - "grad_norm": 0.8302854299545288, - "learning_rate": 2.4302859156401512e-05, - "loss": 0.10022387504577637, - "mean_token_accuracy": 0.9645011806488037, - "num_tokens": 7422217.0, - "step": 3100 - }, - { - "entropy": 0.1541054170578718, - "epoch": 7.591073582629674, - "grad_norm": 0.8342278599739075, - "learning_rate": 2.218096134215895e-05, - "loss": 0.09620431900024413, - "mean_token_accuracy": 0.9666509574651718, - "num_tokens": 7547071.0, - "step": 3150 - }, - { - "entropy": 0.16526211962103843, - "epoch": 7.711700844390832, - "grad_norm": 0.9860979914665222, - "learning_rate": 2.0137280558790573e-05, - "loss": 0.10015854835510254, - "mean_token_accuracy": 0.9650658935308456, - "num_tokens": 7663184.0, - "step": 3200 - }, - { - "entropy": 0.1566712998598814, - "epoch": 7.832328106151991, - "grad_norm": 0.6179924011230469, - "learning_rate": 1.8175430968086255e-05, - "loss": 0.09698445320129395, - "mean_token_accuracy": 0.9661821335554123, - "num_tokens": 7785293.0, - "step": 3250 - }, - { - "entropy": 0.16459605664014818, - "epoch": 7.952955367913148, - "grad_norm": 0.7428495287895203, - "learning_rate": 1.6298882016882273e-05, - "loss": 0.09972883224487304, - "mean_token_accuracy": 0.9650641173124314, - "num_tokens": 7903616.0, - "step": 3300 - }, - { - "epoch": 8.0, - "eval_entropy": 0.339001152455137, - "eval_mean_token_accuracy": 0.7834485541568713, - "eval_not_syn_loss": 1.1380223035812378, - "eval_not_syn_runtime": 52.6766, - "eval_not_syn_samples_per_second": 26.957, - "eval_not_syn_steps_per_second": 3.379, - "eval_num_tokens": 7950448.0, - "step": 3320 - }, - { - "epoch": 8.0, - "eval_entropy": 0.3175283433513695, - "eval_mean_token_accuracy": 0.8104708164595487, - "eval_num_tokens": 7950448.0, - "eval_syn_loss": 1.0303504467010498, - "eval_syn_runtime": 54.8193, - "eval_syn_samples_per_second": 25.903, - "eval_syn_steps_per_second": 3.247, - "step": 3320 - } - ], - "logging_steps": 50, - "max_steps": 4150, - "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": 1.6904481806551872e+17, - "train_batch_size": 8, - "trial_name": null, - "trial_params": null -} diff --git a/substitutivity_original_Estonian/Qwen3.5-2B-Base_substitutivity_splits_original_features_train_substitutivity_splits_original_features_test1/checkpoint-3320/training_args.bin b/substitutivity_original_Estonian/Qwen3.5-2B-Base_substitutivity_splits_original_features_train_substitutivity_splits_original_features_test1/checkpoint-3320/training_args.bin deleted file mode 100644 index 48901efb30506830185bfaafecbe236e7095a006..0000000000000000000000000000000000000000 --- a/substitutivity_original_Estonian/Qwen3.5-2B-Base_substitutivity_splits_original_features_train_substitutivity_splits_original_features_test1/checkpoint-3320/training_args.bin +++ /dev/null @@ -1,3 +0,0 @@ -version https://git-lfs.github.com/spec/v1 -oid sha256:770fc15ebc27a54d780af597d32c57ca14b3b528f14788d142bdfce09b555bef -size 6033 diff --git a/substitutivity_original_Estonian/Qwen3.5-2B-Base_substitutivity_splits_original_features_train_substitutivity_splits_original_features_test1/checkpoint-3735/README.md b/substitutivity_original_Estonian/Qwen3.5-2B-Base_substitutivity_splits_original_features_train_substitutivity_splits_original_features_test1/checkpoint-3735/README.md deleted file mode 100644 index ece204d005f7bf161d0e448efdd56fd297b7c8a2..0000000000000000000000000000000000000000 --- a/substitutivity_original_Estonian/Qwen3.5-2B-Base_substitutivity_splits_original_features_train_substitutivity_splits_original_features_test1/checkpoint-3735/README.md +++ /dev/null @@ -1,209 +0,0 @@ ---- -base_model: Qwen/Qwen3.5-2B-Base -library_name: peft -pipeline_tag: text-generation -tags: -- base_model:adapter:Qwen/Qwen3.5-2B-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/substitutivity_original_Estonian/Qwen3.5-2B-Base_substitutivity_splits_original_features_train_substitutivity_splits_original_features_test1/checkpoint-3735/adapter_config.json b/substitutivity_original_Estonian/Qwen3.5-2B-Base_substitutivity_splits_original_features_train_substitutivity_splits_original_features_test1/checkpoint-3735/adapter_config.json deleted file mode 100644 index a64145ce23adcbdbe246679087a13f4cb63ff664..0000000000000000000000000000000000000000 --- a/substitutivity_original_Estonian/Qwen3.5-2B-Base_substitutivity_splits_original_features_train_substitutivity_splits_original_features_test1/checkpoint-3735/adapter_config.json +++ /dev/null @@ -1,46 +0,0 @@ -{ - "alora_invocation_tokens": null, - "alpha_pattern": {}, - "arrow_config": null, - "auto_mapping": null, - "base_model_name_or_path": "Qwen/Qwen3.5-2B-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": 256, - "lora_bias": false, - "lora_dropout": 0.027050353915701653, - "megatron_config": null, - "megatron_core": "megatron.core", - "modules_to_save": null, - "peft_type": "LORA", - "peft_version": "0.18.1", - "qalora_group_size": 16, - "r": 128, - "rank_pattern": {}, - "revision": null, - "target_modules": [ - "o_proj", - "q_proj", - "k_proj", - "down_proj", - "up_proj", - "v_proj", - "gate_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/substitutivity_original_Estonian/Qwen3.5-2B-Base_substitutivity_splits_original_features_train_substitutivity_splits_original_features_test1/checkpoint-3735/adapter_model.safetensors b/substitutivity_original_Estonian/Qwen3.5-2B-Base_substitutivity_splits_original_features_train_substitutivity_splits_original_features_test1/checkpoint-3735/adapter_model.safetensors deleted file mode 100644 index bd33da7fc4d9d00c5465d9375bd00bb5343ff5bc..0000000000000000000000000000000000000000 --- a/substitutivity_original_Estonian/Qwen3.5-2B-Base_substitutivity_splits_original_features_train_substitutivity_splits_original_features_test1/checkpoint-3735/adapter_model.safetensors +++ /dev/null @@ -1,3 +0,0 @@ -version https://git-lfs.github.com/spec/v1 -oid sha256:83532f622e1a65123912fc580f987262f305c4cf7dc1632ca8fcdc04c05a40f3 -size 349201608 diff --git a/substitutivity_original_Estonian/Qwen3.5-2B-Base_substitutivity_splits_original_features_train_substitutivity_splits_original_features_test1/checkpoint-3735/chat_template.jinja b/substitutivity_original_Estonian/Qwen3.5-2B-Base_substitutivity_splits_original_features_train_substitutivity_splits_original_features_test1/checkpoint-3735/chat_template.jinja deleted file mode 100644 index 0ef09f214eaa6d9bca297988afc1454b5827b2c7..0000000000000000000000000000000000000000 --- a/substitutivity_original_Estonian/Qwen3.5-2B-Base_substitutivity_splits_original_features_train_substitutivity_splits_original_features_test1/checkpoint-3735/chat_template.jinja +++ /dev/null @@ -1,154 +0,0 @@ -{%- set image_count = namespace(value=0) %} -{%- set video_count = namespace(value=0) %} -{%- macro render_content(content, do_vision_count, is_system_content=false) %} - {%- if content is string %} - {{- content }} - {%- elif content is iterable and content is not mapping %} - {%- for item in content %} - {%- if 'image' in item or 'image_url' in item or item.type == 'image' %} - {%- if is_system_content %} - {{- raise_exception('System message cannot contain images.') }} - {%- endif %} - {%- if do_vision_count %} - {%- set image_count.value = image_count.value + 1 %} - {%- endif %} - {%- if add_vision_id %} - {{- 'Picture ' ~ image_count.value ~ ': ' }} - {%- endif %} - {{- '<|vision_start|><|image_pad|><|vision_end|>' }} - {%- elif 'video' in item or item.type == 'video' %} - {%- if is_system_content %} - {{- raise_exception('System message cannot contain videos.') }} - {%- endif %} - {%- if do_vision_count %} - {%- set video_count.value = video_count.value + 1 %} - {%- endif %} - {%- if add_vision_id %} - {{- 'Video ' ~ video_count.value ~ ': ' }} - {%- endif %} - {{- '<|vision_start|><|video_pad|><|vision_end|>' }} - {%- elif 'text' in item %} - {{- item.text }} - {%- else %} - {{- raise_exception('Unexpected item type in content.') }} - {%- endif %} - {%- endfor %} - {%- elif content is none or content is undefined %} - {{- '' }} - {%- else %} - {{- raise_exception('Unexpected content type.') }} - {%- endif %} -{%- endmacro %} -{%- if not messages %} - {{- raise_exception('No messages provided.') }} -{%- endif %} -{%- if tools and tools is iterable and tools is not mapping %} - {{- '<|im_start|>system\n' }} - {{- "# Tools\n\nYou have access to the following functions:\n\n" }} - {%- for tool in tools %} - {{- "\n" }} - {{- tool | tojson }} - {%- endfor %} - {{- "\n" }} - {{- '\n\nIf you choose to call a function ONLY reply in the following format with NO suffix:\n\n\n\n\nvalue_1\n\n\nThis is the value for the second parameter\nthat can span\nmultiple lines\n\n\n\n\n\nReminder:\n- Function calls MUST follow the specified format: an inner block must be nested within XML tags\n- Required parameters MUST be specified\n- You may provide optional reasoning for your function call in natural language BEFORE the function call, but NOT after\n- If there is no function call available, answer the question like normal with your current knowledge and do not tell the user about function calls\n' }} - {%- if messages[0].role == 'system' %} - {%- set content = render_content(messages[0].content, false, true)|trim %} - {%- if content %} - {{- '\n\n' + content }} - {%- endif %} - {%- endif %} - {{- '<|im_end|>\n' }} -{%- else %} - {%- if messages[0].role == 'system' %} - {%- set content = render_content(messages[0].content, false, true)|trim %} - {{- '<|im_start|>system\n' + 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" %} - {%- set content = render_content(message.content, false)|trim %} - {%- if not(content.startswith('') and content.endswith('')) %} - {%- set ns.multi_step_tool = false %} - {%- set ns.last_query_index = index %} - {%- endif %} - {%- endif %} -{%- endfor %} -{%- if ns.multi_step_tool %} - {{- raise_exception('No user query found in messages.') }} -{%- endif %} -{%- for message in messages %} - {%- set content = render_content(message.content, true)|trim %} - {%- if message.role == "system" %} - {%- if not loop.first %} - {{- raise_exception('System message must be at the beginning.') }} - {%- endif %} - {%- elif message.role == "user" %} - {{- '<|im_start|>' + message.role + '\n' + content + '<|im_end|>' + '\n' }} - {%- elif message.role == "assistant" %} - {%- set reasoning_content = '' %} - {%- if message.reasoning_content is string %} - {%- set reasoning_content = message.reasoning_content %} - {%- else %} - {%- if '' in content %} - {%- set reasoning_content = content.split('')[0].rstrip('\n').split('')[-1].lstrip('\n') %} - {%- set content = content.split('')[-1].lstrip('\n') %} - {%- endif %} - {%- endif %} - {%- set reasoning_content = reasoning_content|trim %} - {%- if loop.index0 > ns.last_query_index %} - {{- '<|im_start|>' + message.role + '\n\n' + reasoning_content + '\n\n\n' + content }} - {%- else %} - {{- '<|im_start|>' + message.role + '\n' + content }} - {%- endif %} - {%- if message.tool_calls and message.tool_calls is iterable and message.tool_calls is not mapping %} - {%- for tool_call in message.tool_calls %} - {%- if tool_call.function is defined %} - {%- set tool_call = tool_call.function %} - {%- endif %} - {%- if loop.first %} - {%- if content|trim %} - {{- '\n\n\n\n' }} - {%- else %} - {{- '\n\n' }} - {%- endif %} - {%- else %} - {{- '\n\n\n' }} - {%- endif %} - {%- if tool_call.arguments is defined %} - {%- for args_name, args_value in tool_call.arguments|items %} - {{- '\n' }} - {%- set args_value = args_value | tojson | safe if args_value is mapping or (args_value is sequence and args_value is not string) else args_value | string %} - {{- args_value }} - {{- '\n\n' }} - {%- endfor %} - {%- endif %} - {{- '\n' }} - {%- endfor %} - {%- endif %} - {{- '<|im_end|>\n' }} - {%- elif message.role == "tool" %} - {%- if loop.previtem and loop.previtem.role != "tool" %} - {{- '<|im_start|>user' }} - {%- endif %} - {{- '\n\n' }} - {{- content }} - {{- '\n' }} - {%- if not loop.last and loop.nextitem.role != "tool" %} - {{- '<|im_end|>\n' }} - {%- elif loop.last %} - {{- '<|im_end|>\n' }} - {%- endif %} - {%- else %} - {{- raise_exception('Unexpected message role.') }} - {%- endif %} -{%- endfor %} -{%- if add_generation_prompt %} - {{- '<|im_start|>assistant\n' }} - {%- if enable_thinking is defined and enable_thinking is true %} - {{- '\n' }} - {%- else %} - {{- '\n\n\n\n' }} - {%- endif %} -{%- endif %} \ No newline at end of file diff --git a/substitutivity_original_Estonian/Qwen3.5-2B-Base_substitutivity_splits_original_features_train_substitutivity_splits_original_features_test1/checkpoint-3735/tokenizer.json b/substitutivity_original_Estonian/Qwen3.5-2B-Base_substitutivity_splits_original_features_train_substitutivity_splits_original_features_test1/checkpoint-3735/tokenizer.json deleted file mode 100644 index 67741b04f23bfdb46501f748ce27865ec82eccfb..0000000000000000000000000000000000000000 --- a/substitutivity_original_Estonian/Qwen3.5-2B-Base_substitutivity_splits_original_features_train_substitutivity_splits_original_features_test1/checkpoint-3735/tokenizer.json +++ /dev/null @@ -1,3 +0,0 @@ -version https://git-lfs.github.com/spec/v1 -oid sha256:87a7830d63fcf43bf241c3c5242e96e62dd3fdc29224ca26fed8ea333db72de4 -size 19989343 diff --git a/substitutivity_original_Estonian/Qwen3.5-2B-Base_substitutivity_splits_original_features_train_substitutivity_splits_original_features_test1/checkpoint-3735/tokenizer_config.json b/substitutivity_original_Estonian/Qwen3.5-2B-Base_substitutivity_splits_original_features_train_substitutivity_splits_original_features_test1/checkpoint-3735/tokenizer_config.json deleted file mode 100644 index b4a37b2a6fd3ab3317cd7bac72855be1a843b2bb..0000000000000000000000000000000000000000 --- a/substitutivity_original_Estonian/Qwen3.5-2B-Base_substitutivity_splits_original_features_train_substitutivity_splits_original_features_test1/checkpoint-3735/tokenizer_config.json +++ /dev/null @@ -1,31 +0,0 @@ -{ - "add_prefix_space": false, - "audio_bos_token": "<|audio_start|>", - "audio_eos_token": "<|audio_end|>", - "audio_token": "<|audio_pad|>", - "backend": "tokenizers", - "bos_token": null, - "clean_up_tokenization_spaces": false, - "eos_token": "<|endoftext|>", - "errors": "replace", - "image_token": "<|image_pad|>", - "is_local": false, - "model_max_length": 262144, - "model_specific_special_tokens": { - "audio_bos_token": "<|audio_start|>", - "audio_eos_token": "<|audio_end|>", - "audio_token": "<|audio_pad|>", - "image_token": "<|image_pad|>", - "video_token": "<|video_pad|>", - "vision_bos_token": "<|vision_start|>", - "vision_eos_token": "<|vision_end|>" - }, - "pad_token": "<|endoftext|>", - "pretokenize_regex": "(?i:'s|'t|'re|'ve|'m|'ll|'d)|[^\\r\\n\\p{L}\\p{N}]?[\\p{L}\\p{M}]+|\\p{N}| ?[^\\s\\p{L}\\p{M}\\p{N}]+[\\r\\n]*|\\s*[\\r\\n]+|\\s+(?!\\S)|\\s+", - "split_special_tokens": false, - "tokenizer_class": "TokenizersBackend", - "unk_token": null, - "video_token": "<|video_pad|>", - "vision_bos_token": "<|vision_start|>", - "vision_eos_token": "<|vision_end|>" -} diff --git a/substitutivity_original_Estonian/Qwen3.5-2B-Base_substitutivity_splits_original_features_train_substitutivity_splits_original_features_test1/checkpoint-3735/trainer_state.json b/substitutivity_original_Estonian/Qwen3.5-2B-Base_substitutivity_splits_original_features_train_substitutivity_splits_original_features_test1/checkpoint-3735/trainer_state.json deleted file mode 100644 index 6d7532845f86e354c0b9f8ae074b850b311c2be9..0000000000000000000000000000000000000000 --- a/substitutivity_original_Estonian/Qwen3.5-2B-Base_substitutivity_splits_original_features_train_substitutivity_splits_original_features_test1/checkpoint-3735/trainer_state.json +++ /dev/null @@ -1,972 +0,0 @@ -{ - "best_global_step": null, - "best_metric": null, - "best_model_checkpoint": null, - "epoch": 9.0, - "eval_steps": 500, - "global_step": 3735, - "is_hyper_param_search": false, - "is_local_process_zero": true, - "is_world_process_zero": true, - "log_history": [ - { - "entropy": 2.0473356211185454, - "epoch": 0.12062726176115803, - "grad_norm": 3.0554118156433105, - "learning_rate": 1.5682344549790923e-05, - "loss": 1.910550079345703, - "mean_token_accuracy": 0.6053373458981514, - "num_tokens": 121428.0, - "step": 50 - }, - { - "entropy": 1.0371007251739501, - "epoch": 0.24125452352231605, - "grad_norm": 2.109868049621582, - "learning_rate": 3.168473694753677e-05, - "loss": 0.9585584259033203, - "mean_token_accuracy": 0.7352996730804443, - "num_tokens": 243966.0, - "step": 100 - }, - { - "entropy": 0.9105490958690643, - "epoch": 0.3618817852834741, - "grad_norm": 1.7675701379776, - "learning_rate": 4.7687129345282605e-05, - "loss": 0.8393524932861328, - "mean_token_accuracy": 0.7588736474514007, - "num_tokens": 366649.0, - "step": 150 - }, - { - "entropy": 0.8660378754138947, - "epoch": 0.4825090470446321, - "grad_norm": 1.6502987146377563, - "learning_rate": 6.368952174302845e-05, - "loss": 0.8005844116210937, - "mean_token_accuracy": 0.7663125538825989, - "num_tokens": 488367.0, - "step": 200 - }, - { - "entropy": 0.8439123743772506, - "epoch": 0.6031363088057901, - "grad_norm": 1.538365364074707, - "learning_rate": 7.969191414077429e-05, - "loss": 0.7773619842529297, - "mean_token_accuracy": 0.7718527519702911, - "num_tokens": 603131.0, - "step": 250 - }, - { - "entropy": 0.8223806774616241, - "epoch": 0.7237635705669482, - "grad_norm": 1.1559664011001587, - "learning_rate": 9.569430653852014e-05, - "loss": 0.7587767791748047, - "mean_token_accuracy": 0.7767325991392136, - "num_tokens": 726412.0, - "step": 300 - }, - { - "entropy": 0.8256935775279999, - "epoch": 0.8443908323281062, - "grad_norm": 1.2322715520858765, - "learning_rate": 0.00011169669893626597, - "loss": 0.769886245727539, - "mean_token_accuracy": 0.773840668797493, - "num_tokens": 846516.0, - "step": 350 - }, - { - "entropy": 0.8065843790769577, - "epoch": 0.9650180940892642, - "grad_norm": 0.9908286333084106, - "learning_rate": 0.00012769909133401183, - "loss": 0.7451084899902344, - "mean_token_accuracy": 0.7794792425632476, - "num_tokens": 958978.0, - "step": 400 - }, - { - "epoch": 1.0, - "eval_entropy": 0.791056221790528, - "eval_mean_token_accuracy": 0.7507295300451557, - "eval_not_syn_loss": 0.7860428094863892, - "eval_not_syn_runtime": 53.5431, - "eval_not_syn_samples_per_second": 26.521, - "eval_not_syn_steps_per_second": 3.324, - "eval_num_tokens": 993806.0, - "step": 415 - }, - { - "epoch": 1.0, - "eval_entropy": 0.7424047408478983, - "eval_mean_token_accuracy": 0.8023804037758474, - "eval_num_tokens": 993806.0, - "eval_syn_loss": 0.7273509502410889, - "eval_syn_runtime": 54.8679, - "eval_syn_samples_per_second": 25.88, - "eval_syn_steps_per_second": 3.244, - "step": 415 - }, - { - "entropy": 0.7607260761838971, - "epoch": 1.0844390832328106, - "grad_norm": 1.0653696060180664, - "learning_rate": 0.0001327927019027069, - "loss": 0.7034613037109375, - "mean_token_accuracy": 0.7869195998317063, - "num_tokens": 1080301.0, - "step": 450 - }, - { - "entropy": 0.7370894801616669, - "epoch": 1.2050663449939687, - "grad_norm": 1.763787865638733, - "learning_rate": 0.00013265416569994628, - "loss": 0.6816973114013671, - "mean_token_accuracy": 0.7914507347345352, - "num_tokens": 1201215.0, - "step": 500 - }, - { - "entropy": 0.742829909324646, - "epoch": 1.3256936067551266, - "grad_norm": 0.97663414478302, - "learning_rate": 0.00013239847940553975, - "loss": 0.6867338562011719, - "mean_token_accuracy": 0.790328289270401, - "num_tokens": 1322895.0, - "step": 550 - }, - { - "entropy": 0.7312483406066894, - "epoch": 1.4463208685162847, - "grad_norm": 1.0874731540679932, - "learning_rate": 0.00013202609518972978, - "loss": 0.676800537109375, - "mean_token_accuracy": 0.7921512228250503, - "num_tokens": 1436749.0, - "step": 600 - }, - { - "entropy": 0.7178869074583054, - "epoch": 1.5669481302774426, - "grad_norm": 1.064300298690796, - "learning_rate": 0.0001315376715980265, - "loss": 0.6635584259033203, - "mean_token_accuracy": 0.7944776183366775, - "num_tokens": 1555400.0, - "step": 650 - }, - { - "entropy": 0.7198166918754577, - "epoch": 1.6875753920386007, - "grad_norm": 0.7779315114021301, - "learning_rate": 0.00013093407238659802, - "loss": 0.6660567474365234, - "mean_token_accuracy": 0.7968300533294678, - "num_tokens": 1676892.0, - "step": 700 - }, - { - "entropy": 0.7170061504840851, - "epoch": 1.8082026537997589, - "grad_norm": 1.089854121208191, - "learning_rate": 0.00013021636499475496, - "loss": 0.6632771301269531, - "mean_token_accuracy": 0.7963915091753005, - "num_tokens": 1802689.0, - "step": 750 - }, - { - "entropy": 0.6953610718250275, - "epoch": 1.9288299155609168, - "grad_norm": 0.7996730208396912, - "learning_rate": 0.0001293858186572295, - "loss": 0.6421419525146485, - "mean_token_accuracy": 0.8002577543258667, - "num_tokens": 1920804.0, - "step": 800 - }, - { - "epoch": 2.0, - "eval_entropy": 0.7369416523515508, - "eval_mean_token_accuracy": 0.7872022366925572, - "eval_not_syn_loss": 0.713637113571167, - "eval_not_syn_runtime": 53.2265, - "eval_not_syn_samples_per_second": 26.678, - "eval_not_syn_steps_per_second": 3.344, - "eval_num_tokens": 1987612.0, - "step": 830 - }, - { - "epoch": 2.0, - "eval_entropy": 0.6877032756135705, - "eval_mean_token_accuracy": 0.7997466372639945, - "eval_num_tokens": 1987612.0, - "eval_syn_loss": 0.6614266633987427, - "eval_syn_runtime": 54.5178, - "eval_syn_samples_per_second": 26.047, - "eval_syn_steps_per_second": 3.265, - "step": 830 - }, - { - "entropy": 0.6644553433765065, - "epoch": 2.0482509047044632, - "grad_norm": 0.7562468647956848, - "learning_rate": 0.00012844390215958852, - "loss": 0.6112185287475586, - "mean_token_accuracy": 0.806649933559726, - "num_tokens": 2039660.0, - "step": 850 - }, - { - "entropy": 0.6146580249071121, - "epoch": 2.168878166465621, - "grad_norm": 1.098607063293457, - "learning_rate": 0.00012739228124074913, - "loss": 0.5553749084472657, - "mean_token_accuracy": 0.8187588346004486, - "num_tokens": 2161772.0, - "step": 900 - }, - { - "entropy": 0.6187627577781677, - "epoch": 2.2895054282267795, - "grad_norm": 1.0338040590286255, - "learning_rate": 0.00012623281564719095, - "loss": 0.5596194076538086, - "mean_token_accuracy": 0.8182566928863525, - "num_tokens": 2280833.0, - "step": 950 - }, - { - "entropy": 0.6232365870475769, - "epoch": 2.4101326899879374, - "grad_norm": 1.0010799169540405, - "learning_rate": 0.00012496755584407422, - "loss": 0.5667143249511719, - "mean_token_accuracy": 0.8175061982870102, - "num_tokens": 2398206.0, - "step": 1000 - }, - { - "entropy": 0.6151268655061721, - "epoch": 2.5307599517490953, - "grad_norm": 1.0065034627914429, - "learning_rate": 0.0001235987393890801, - "loss": 0.5598890686035156, - "mean_token_accuracy": 0.8193449324369431, - "num_tokens": 2518888.0, - "step": 1050 - }, - { - "entropy": 0.6173626267910004, - "epoch": 2.651387213510253, - "grad_norm": 0.7958805561065674, - "learning_rate": 0.00012212878697538592, - "loss": 0.5649292755126953, - "mean_token_accuracy": 0.8172902607917786, - "num_tokens": 2639763.0, - "step": 1100 - }, - { - "entropy": 0.6264930808544159, - "epoch": 2.772014475271411, - "grad_norm": 0.7718372344970703, - "learning_rate": 0.00012056029815077323, - "loss": 0.5716455078125, - "mean_token_accuracy": 0.8170186239480972, - "num_tokens": 2755540.0, - "step": 1150 - }, - { - "entropy": 0.6203387409448624, - "epoch": 2.8926417370325694, - "grad_norm": 0.7914765477180481, - "learning_rate": 0.000118896046720439, - "loss": 0.5653961944580078, - "mean_token_accuracy": 0.8170280295610428, - "num_tokens": 2874118.0, - "step": 1200 - }, - { - "epoch": 3.0, - "eval_entropy": 0.6751241854737314, - "eval_mean_token_accuracy": 0.792226163858778, - "eval_not_syn_loss": 0.6902772784233093, - "eval_not_syn_runtime": 53.2557, - "eval_not_syn_samples_per_second": 26.664, - "eval_not_syn_steps_per_second": 3.342, - "eval_num_tokens": 2981418.0, - "step": 1245 - }, - { - "epoch": 3.0, - "eval_entropy": 0.6308753868837035, - "eval_mean_token_accuracy": 0.8131611759073278, - "eval_num_tokens": 2981418.0, - "eval_syn_loss": 0.6440441012382507, - "eval_syn_runtime": 54.7064, - "eval_syn_samples_per_second": 25.957, - "eval_syn_steps_per_second": 3.254, - "step": 1245 - }, - { - "entropy": 0.6077835421369533, - "epoch": 3.012062726176116, - "grad_norm": 1.124720573425293, - "learning_rate": 0.00011713897584164003, - "loss": 0.5557136917114258, - "mean_token_accuracy": 0.8203362726201915, - "num_tokens": 2994602.0, - "step": 1250 - }, - { - "entropy": 0.507642827630043, - "epoch": 3.132689987937274, - "grad_norm": 0.9039916396141052, - "learning_rate": 0.00011529219281884579, - "loss": 0.4467372512817383, - "mean_token_accuracy": 0.8464260965585708, - "num_tokens": 3121094.0, - "step": 1300 - }, - { - "entropy": 0.5137191081047058, - "epoch": 3.2533172496984317, - "grad_norm": 0.8580691814422607, - "learning_rate": 0.00011335896360860367, - "loss": 0.4538694763183594, - "mean_token_accuracy": 0.8448719161748887, - "num_tokens": 3248038.0, - "step": 1350 - }, - { - "entropy": 0.5293738862872124, - "epoch": 3.37394451145959, - "grad_norm": 1.0177273750305176, - "learning_rate": 0.00011134270704383492, - "loss": 0.4639348220825195, - "mean_token_accuracy": 0.8428975892066956, - "num_tokens": 3363487.0, - "step": 1400 - }, - { - "entropy": 0.5218872320652008, - "epoch": 3.494571773220748, - "grad_norm": 1.0530673265457153, - "learning_rate": 0.00010924698878777538, - "loss": 0.4584346389770508, - "mean_token_accuracy": 0.8432288640737533, - "num_tokens": 3480513.0, - "step": 1450 - }, - { - "entropy": 0.5320372316241264, - "epoch": 3.615199034981906, - "grad_norm": 1.0291194915771484, - "learning_rate": 0.0001070755150282529, - "loss": 0.4711898422241211, - "mean_token_accuracy": 0.842066935300827, - "num_tokens": 3597183.0, - "step": 1500 - }, - { - "entropy": 0.5315000694990158, - "epoch": 3.7358262967430638, - "grad_norm": 0.9788838624954224, - "learning_rate": 0.00010483212592345332, - "loss": 0.4729101943969727, - "mean_token_accuracy": 0.8404586082696914, - "num_tokens": 3714173.0, - "step": 1550 - }, - { - "entropy": 0.5331643190979958, - "epoch": 3.856453558504222, - "grad_norm": 1.1248903274536133, - "learning_rate": 0.00010252078881076532, - "loss": 0.47405517578125, - "mean_token_accuracy": 0.8397526168823242, - "num_tokens": 3832652.0, - "step": 1600 - }, - { - "entropy": 0.5343865439295769, - "epoch": 3.97708082026538, - "grad_norm": 0.9008272886276245, - "learning_rate": 0.00010014559119071452, - "loss": 0.47172779083251953, - "mean_token_accuracy": 0.8404951041936874, - "num_tokens": 3955167.0, - "step": 1650 - }, - { - "epoch": 4.0, - "eval_entropy": 0.5736920103598177, - "eval_mean_token_accuracy": 0.7981750027517255, - "eval_not_syn_loss": 0.707631528377533, - "eval_not_syn_runtime": 52.2858, - "eval_not_syn_samples_per_second": 27.158, - "eval_not_syn_steps_per_second": 3.404, - "eval_num_tokens": 3975224.0, - "step": 1660 - }, - { - "epoch": 4.0, - "eval_entropy": 0.5363499959198277, - "eval_mean_token_accuracy": 0.8063885539435269, - "eval_num_tokens": 3975224.0, - "eval_syn_loss": 0.6610181331634521, - "eval_syn_runtime": 54.2167, - "eval_syn_samples_per_second": 26.191, - "eval_syn_steps_per_second": 3.283, - "step": 1660 - }, - { - "entropy": 0.4512602482179199, - "epoch": 4.0965018094089265, - "grad_norm": 0.8079906105995178, - "learning_rate": 9.771073349839416e-05, - "loss": 0.3778452682495117, - "mean_token_accuracy": 0.8684436221315404, - "num_tokens": 4072361.0, - "step": 1700 - }, - { - "entropy": 0.4228873434662819, - "epoch": 4.217129071170085, - "grad_norm": 1.1206746101379395, - "learning_rate": 9.52205216751758e-05, - "loss": 0.34671127319335937, - "mean_token_accuracy": 0.8763950079679489, - "num_tokens": 4194978.0, - "step": 1750 - }, - { - "entropy": 0.42808119297027586, - "epoch": 4.337756332931242, - "grad_norm": 1.1268125772476196, - "learning_rate": 9.26793595538368e-05, - "loss": 0.3484600830078125, - "mean_token_accuracy": 0.8753323918581009, - "num_tokens": 4315548.0, - "step": 1800 - }, - { - "entropy": 0.42506767451763156, - "epoch": 4.458383594692401, - "grad_norm": 1.1051807403564453, - "learning_rate": 9.009174107057088e-05, - "loss": 0.35434780120849607, - "mean_token_accuracy": 0.8734411865472793, - "num_tokens": 4431469.0, - "step": 1850 - }, - { - "entropy": 0.43314908266067503, - "epoch": 4.579010856453559, - "grad_norm": 0.9782640337944031, - "learning_rate": 8.746224231765484e-05, - "loss": 0.36225860595703124, - "mean_token_accuracy": 0.8715140378475189, - "num_tokens": 4551803.0, - "step": 1900 - }, - { - "entropy": 0.42984022706747055, - "epoch": 4.699638118214716, - "grad_norm": 0.9772566556930542, - "learning_rate": 8.479551345082568e-05, - "loss": 0.356609001159668, - "mean_token_accuracy": 0.8717925781011582, - "num_tokens": 4674468.0, - "step": 1950 - }, - { - "entropy": 0.4474307084083557, - "epoch": 4.820265379975875, - "grad_norm": 0.9547265768051147, - "learning_rate": 8.209627046567966e-05, - "loss": 0.3654902267456055, - "mean_token_accuracy": 0.8724055212736129, - "num_tokens": 4788027.0, - "step": 2000 - }, - { - "entropy": 0.4377285006642342, - "epoch": 4.940892641737032, - "grad_norm": 1.0390652418136597, - "learning_rate": 7.936928685763635e-05, - "loss": 0.3645761871337891, - "mean_token_accuracy": 0.8707609891891479, - "num_tokens": 4907561.0, - "step": 2050 - }, - { - "epoch": 5.0, - "eval_entropy": 0.5173388388049737, - "eval_mean_token_accuracy": 0.7918405432379647, - "eval_not_syn_loss": 0.7532826662063599, - "eval_not_syn_runtime": 53.0766, - "eval_not_syn_samples_per_second": 26.754, - "eval_not_syn_steps_per_second": 3.354, - "eval_num_tokens": 4969030.0, - "step": 2075 - }, - { - "epoch": 5.0, - "eval_entropy": 0.4831916272975086, - "eval_mean_token_accuracy": 0.8204237202580056, - "eval_num_tokens": 4969030.0, - "eval_syn_loss": 0.6971890330314636, - "eval_syn_runtime": 54.8755, - "eval_syn_samples_per_second": 25.877, - "eval_syn_steps_per_second": 3.244, - "step": 2075 - }, - { - "entropy": 0.3804885711934831, - "epoch": 5.060313630880579, - "grad_norm": 0.8717736005783081, - "learning_rate": 7.661938518021687e-05, - "loss": 0.3034685707092285, - "mean_token_accuracy": 0.8910005471923135, - "num_tokens": 5026625.0, - "step": 2100 - }, - { - "entropy": 0.3174623078107834, - "epoch": 5.180940892641737, - "grad_norm": 0.9808429479598999, - "learning_rate": 7.385142851656478e-05, - "loss": 0.23829608917236328, - "mean_token_accuracy": 0.9131860065460206, - "num_tokens": 5141010.0, - "step": 2150 - }, - { - "entropy": 0.3187542732059956, - "epoch": 5.301568154402895, - "grad_norm": 1.3739993572235107, - "learning_rate": 7.107031187929228e-05, - "loss": 0.24024734497070313, - "mean_token_accuracy": 0.910272576212883, - "num_tokens": 5261240.0, - "step": 2200 - }, - { - "entropy": 0.31624249830842016, - "epoch": 5.422195416164053, - "grad_norm": 1.1618341207504272, - "learning_rate": 6.82809535538603e-05, - "loss": 0.24144052505493163, - "mean_token_accuracy": 0.9122701978683472, - "num_tokens": 5382160.0, - "step": 2250 - }, - { - "entropy": 0.31793304443359377, - "epoch": 5.542822677925211, - "grad_norm": 1.4960980415344238, - "learning_rate": 6.548828640080172e-05, - "loss": 0.24051647186279296, - "mean_token_accuracy": 0.9116496992111206, - "num_tokens": 5502002.0, - "step": 2300 - }, - { - "entropy": 0.3180307722091675, - "epoch": 5.66344993968637, - "grad_norm": 1.103051781654358, - "learning_rate": 6.269724913216913e-05, - "loss": 0.24294727325439452, - "mean_token_accuracy": 0.9102917027473449, - "num_tokens": 5629969.0, - "step": 2350 - }, - { - "entropy": 0.32020670652389527, - "epoch": 5.784077201447527, - "grad_norm": 1.0690243244171143, - "learning_rate": 5.991277757763442e-05, - "loss": 0.24328670501708985, - "mean_token_accuracy": 0.9101310813426972, - "num_tokens": 5750596.0, - "step": 2400 - }, - { - "entropy": 0.31846788004040716, - "epoch": 5.904704463208685, - "grad_norm": 1.203026294708252, - "learning_rate": 5.7139795955685784e-05, - "loss": 0.244984130859375, - "mean_token_accuracy": 0.9100374561548233, - "num_tokens": 5868110.0, - "step": 2450 - }, - { - "epoch": 6.0, - "eval_entropy": 0.45794740654109567, - "eval_mean_token_accuracy": 0.7891001142142864, - "eval_not_syn_loss": 0.8338615894317627, - "eval_not_syn_runtime": 53.3841, - "eval_not_syn_samples_per_second": 26.6, - "eval_not_syn_steps_per_second": 3.334, - "eval_num_tokens": 5962836.0, - "step": 2490 - }, - { - "epoch": 6.0, - "eval_entropy": 0.4288232589370749, - "eval_mean_token_accuracy": 0.8114104933953017, - "eval_num_tokens": 5962836.0, - "eval_syn_loss": 0.7791845798492432, - "eval_syn_runtime": 54.7877, - "eval_syn_samples_per_second": 25.918, - "eval_syn_steps_per_second": 3.249, - "step": 2490 - }, - { - "entropy": 0.2972855128423132, - "epoch": 6.024125452352232, - "grad_norm": 1.1748684644699097, - "learning_rate": 5.438320816535859e-05, - "loss": 0.22051204681396486, - "mean_token_accuracy": 0.9190925630656156, - "num_tokens": 5990107.0, - "step": 2500 - }, - { - "entropy": 0.2216818816959858, - "epoch": 6.144752714113389, - "grad_norm": 1.0899405479431152, - "learning_rate": 5.164788911390022e-05, - "loss": 0.1487915325164795, - "mean_token_accuracy": 0.9462692165374755, - "num_tokens": 6104230.0, - "step": 2550 - }, - { - "entropy": 0.22189782992005347, - "epoch": 6.265379975874548, - "grad_norm": 1.0967661142349243, - "learning_rate": 4.8938676095705914e-05, - "loss": 0.149844388961792, - "mean_token_accuracy": 0.9452104204893113, - "num_tokens": 6223548.0, - "step": 2600 - }, - { - "entropy": 0.2155836282670498, - "epoch": 6.386007237635706, - "grad_norm": 1.1443830728530884, - "learning_rate": 4.626036023777093e-05, - "loss": 0.1472412109375, - "mean_token_accuracy": 0.946697844862938, - "num_tokens": 6349102.0, - "step": 2650 - }, - { - "entropy": 0.22189507275819778, - "epoch": 6.5066344993968634, - "grad_norm": 1.2138874530792236, - "learning_rate": 4.3617678026788244e-05, - "loss": 0.15155397415161131, - "mean_token_accuracy": 0.9441326081752777, - "num_tokens": 6468966.0, - "step": 2700 - }, - { - "entropy": 0.22405965447425844, - "epoch": 6.627261761158022, - "grad_norm": 1.2444716691970825, - "learning_rate": 4.10153029328748e-05, - "loss": 0.15516298294067382, - "mean_token_accuracy": 0.9442053025960923, - "num_tokens": 6583296.0, - "step": 2750 - }, - { - "entropy": 0.2216674068570137, - "epoch": 6.74788902291918, - "grad_norm": 1.0056185722351074, - "learning_rate": 3.845783714474033e-05, - "loss": 0.15423223495483399, - "mean_token_accuracy": 0.9445771616697312, - "num_tokens": 6698696.0, - "step": 2800 - }, - { - "entropy": 0.21455578804016112, - "epoch": 6.868516284680338, - "grad_norm": 1.0748720169067383, - "learning_rate": 3.59498034309138e-05, - "loss": 0.14870482444763183, - "mean_token_accuracy": 0.9457976347208024, - "num_tokens": 6819711.0, - "step": 2850 - }, - { - "entropy": 0.21227611482143402, - "epoch": 6.989143546441496, - "grad_norm": 0.978204607963562, - "learning_rate": 3.349563714142155e-05, - "loss": 0.14802613258361816, - "mean_token_accuracy": 0.9466447854042053, - "num_tokens": 6944941.0, - "step": 2900 - }, - { - "epoch": 7.0, - "eval_entropy": 0.38817436741978933, - "eval_mean_token_accuracy": 0.7801904045463948, - "eval_not_syn_loss": 0.9901206493377686, - "eval_not_syn_runtime": 53.1137, - "eval_not_syn_samples_per_second": 26.735, - "eval_not_syn_steps_per_second": 3.351, - "eval_num_tokens": 6956642.0, - "step": 2905 - }, - { - "epoch": 7.0, - "eval_entropy": 0.3640736076938972, - "eval_mean_token_accuracy": 0.8158511425671953, - "eval_num_tokens": 6956642.0, - "eval_syn_loss": 0.8903024792671204, - "eval_syn_runtime": 54.3768, - "eval_syn_samples_per_second": 26.114, - "eval_syn_steps_per_second": 3.273, - "step": 2905 - }, - { - "entropy": 0.16986672166321012, - "epoch": 7.108564535585042, - "grad_norm": 0.7068538069725037, - "learning_rate": 3.1099678364060876e-05, - "loss": 0.10039780616760254, - "mean_token_accuracy": 0.964904587678235, - "num_tokens": 7064818.0, - "step": 2950 - }, - { - "entropy": 0.16062987893819808, - "epoch": 7.2291917973462, - "grad_norm": 0.963703453540802, - "learning_rate": 2.8766164249141098e-05, - "loss": 0.09831766128540038, - "mean_token_accuracy": 0.9666522133350373, - "num_tokens": 7183624.0, - "step": 3000 - }, - { - "entropy": 0.15993704989552499, - "epoch": 7.349819059107358, - "grad_norm": 0.8077407479286194, - "learning_rate": 2.6499221516264785e-05, - "loss": 0.09768331527709961, - "mean_token_accuracy": 0.9664828568696976, - "num_tokens": 7304122.0, - "step": 3050 - }, - { - "entropy": 0.16501803942024706, - "epoch": 7.470446320868517, - "grad_norm": 0.8302854299545288, - "learning_rate": 2.4302859156401512e-05, - "loss": 0.10022387504577637, - "mean_token_accuracy": 0.9645011806488037, - "num_tokens": 7422217.0, - "step": 3100 - }, - { - "entropy": 0.1541054170578718, - "epoch": 7.591073582629674, - "grad_norm": 0.8342278599739075, - "learning_rate": 2.218096134215895e-05, - "loss": 0.09620431900024413, - "mean_token_accuracy": 0.9666509574651718, - "num_tokens": 7547071.0, - "step": 3150 - }, - { - "entropy": 0.16526211962103843, - "epoch": 7.711700844390832, - "grad_norm": 0.9860979914665222, - "learning_rate": 2.0137280558790573e-05, - "loss": 0.10015854835510254, - "mean_token_accuracy": 0.9650658935308456, - "num_tokens": 7663184.0, - "step": 3200 - }, - { - "entropy": 0.1566712998598814, - "epoch": 7.832328106151991, - "grad_norm": 0.6179924011230469, - "learning_rate": 1.8175430968086255e-05, - "loss": 0.09698445320129395, - "mean_token_accuracy": 0.9661821335554123, - "num_tokens": 7785293.0, - "step": 3250 - }, - { - "entropy": 0.16459605664014818, - "epoch": 7.952955367913148, - "grad_norm": 0.7428495287895203, - "learning_rate": 1.6298882016882273e-05, - "loss": 0.09972883224487304, - "mean_token_accuracy": 0.9650641173124314, - "num_tokens": 7903616.0, - "step": 3300 - }, - { - "epoch": 8.0, - "eval_entropy": 0.339001152455137, - "eval_mean_token_accuracy": 0.7834485541568713, - "eval_not_syn_loss": 1.1380223035812378, - "eval_not_syn_runtime": 52.6766, - "eval_not_syn_samples_per_second": 26.957, - "eval_not_syn_steps_per_second": 3.379, - "eval_num_tokens": 7950448.0, - "step": 3320 - }, - { - "epoch": 8.0, - "eval_entropy": 0.3175283433513695, - "eval_mean_token_accuracy": 0.8104708164595487, - "eval_num_tokens": 7950448.0, - "eval_syn_loss": 1.0303504467010498, - "eval_syn_runtime": 54.8193, - "eval_syn_samples_per_second": 25.903, - "eval_syn_steps_per_second": 3.247, - "step": 3320 - }, - { - "entropy": 0.14740703307619, - "epoch": 8.072376357056696, - "grad_norm": 0.5945252776145935, - "learning_rate": 1.4510952301493426e-05, - "loss": 0.08425969123840332, - "mean_token_accuracy": 0.9707530875398656, - "num_tokens": 8020867.0, - "step": 3350 - }, - { - "entropy": 0.13212954834103585, - "epoch": 8.193003618817853, - "grad_norm": 0.6547225713729858, - "learning_rate": 1.2814803698917796e-05, - "loss": 0.07411238670349121, - "mean_token_accuracy": 0.9743285477161407, - "num_tokens": 8143032.0, - "step": 3400 - }, - { - "entropy": 0.13668791383504866, - "epoch": 8.31363088057901, - "grad_norm": 0.49490705132484436, - "learning_rate": 1.1213435775192938e-05, - "loss": 0.07555238246917724, - "mean_token_accuracy": 0.9738145649433136, - "num_tokens": 8262875.0, - "step": 3450 - }, - { - "entropy": 0.13185277588665487, - "epoch": 8.43425814234017, - "grad_norm": 0.43988436460494995, - "learning_rate": 9.709680480792023e-06, - "loss": 0.075065279006958, - "mean_token_accuracy": 0.9734203344583512, - "num_tokens": 8383147.0, - "step": 3500 - }, - { - "entropy": 0.1305071434378624, - "epoch": 8.554885404101327, - "grad_norm": 0.46255460381507874, - "learning_rate": 8.306197142440928e-06, - "loss": 0.07382431983947754, - "mean_token_accuracy": 0.9748922157287597, - "num_tokens": 8502743.0, - "step": 3550 - }, - { - "entropy": 0.1334294395148754, - "epoch": 8.675512665862485, - "grad_norm": 0.6507574915885925, - "learning_rate": 7.0054677602129975e-06, - "loss": 0.0754416275024414, - "mean_token_accuracy": 0.9741936498880386, - "num_tokens": 8621550.0, - "step": 3600 - }, - { - "entropy": 0.13020497389137745, - "epoch": 8.796139927623642, - "grad_norm": 0.523806631565094, - "learning_rate": 5.809792618218385e-06, - "loss": 0.07472522258758545, - "mean_token_accuracy": 0.9740422391891479, - "num_tokens": 8742590.0, - "step": 3650 - }, - { - "entropy": 0.13258147858083247, - "epoch": 8.916767189384801, - "grad_norm": 0.6534594893455505, - "learning_rate": 4.721286216650351e-06, - "loss": 0.07624986171722412, - "mean_token_accuracy": 0.9732474505901336, - "num_tokens": 8862250.0, - "step": 3700 - }, - { - "epoch": 9.0, - "eval_entropy": 0.3192606071910162, - "eval_mean_token_accuracy": 0.7872959100798275, - "eval_not_syn_loss": 1.2220250368118286, - "eval_not_syn_runtime": 53.1436, - "eval_not_syn_samples_per_second": 26.72, - "eval_not_syn_steps_per_second": 3.349, - "eval_num_tokens": 8944254.0, - "step": 3735 - }, - { - "epoch": 9.0, - "eval_entropy": 0.298558424028118, - "eval_mean_token_accuracy": 0.8055475206857317, - "eval_num_tokens": 8944254.0, - "eval_syn_loss": 1.140589714050293, - "eval_syn_runtime": 54.7832, - "eval_syn_samples_per_second": 25.92, - "eval_syn_steps_per_second": 3.249, - "step": 3735 - } - ], - "logging_steps": 50, - "max_steps": 4150, - "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": 1.901312190409728e+17, - "train_batch_size": 8, - "trial_name": null, - "trial_params": null -} diff --git a/substitutivity_original_Estonian/Qwen3.5-2B-Base_substitutivity_splits_original_features_train_substitutivity_splits_original_features_test1/checkpoint-3735/training_args.bin b/substitutivity_original_Estonian/Qwen3.5-2B-Base_substitutivity_splits_original_features_train_substitutivity_splits_original_features_test1/checkpoint-3735/training_args.bin deleted file mode 100644 index 48901efb30506830185bfaafecbe236e7095a006..0000000000000000000000000000000000000000 --- a/substitutivity_original_Estonian/Qwen3.5-2B-Base_substitutivity_splits_original_features_train_substitutivity_splits_original_features_test1/checkpoint-3735/training_args.bin +++ /dev/null @@ -1,3 +0,0 @@ -version https://git-lfs.github.com/spec/v1 -oid sha256:770fc15ebc27a54d780af597d32c57ca14b3b528f14788d142bdfce09b555bef -size 6033 diff --git a/substitutivity_original_Estonian/Qwen3.5-2B-Base_substitutivity_splits_original_features_train_substitutivity_splits_original_features_test1/checkpoint-415/README.md b/substitutivity_original_Estonian/Qwen3.5-2B-Base_substitutivity_splits_original_features_train_substitutivity_splits_original_features_test1/checkpoint-415/README.md deleted file mode 100644 index ece204d005f7bf161d0e448efdd56fd297b7c8a2..0000000000000000000000000000000000000000 --- a/substitutivity_original_Estonian/Qwen3.5-2B-Base_substitutivity_splits_original_features_train_substitutivity_splits_original_features_test1/checkpoint-415/README.md +++ /dev/null @@ -1,209 +0,0 @@ ---- -base_model: Qwen/Qwen3.5-2B-Base -library_name: peft -pipeline_tag: text-generation -tags: -- base_model:adapter:Qwen/Qwen3.5-2B-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/substitutivity_original_Estonian/Qwen3.5-2B-Base_substitutivity_splits_original_features_train_substitutivity_splits_original_features_test1/checkpoint-415/adapter_config.json b/substitutivity_original_Estonian/Qwen3.5-2B-Base_substitutivity_splits_original_features_train_substitutivity_splits_original_features_test1/checkpoint-415/adapter_config.json deleted file mode 100644 index a64145ce23adcbdbe246679087a13f4cb63ff664..0000000000000000000000000000000000000000 --- a/substitutivity_original_Estonian/Qwen3.5-2B-Base_substitutivity_splits_original_features_train_substitutivity_splits_original_features_test1/checkpoint-415/adapter_config.json +++ /dev/null @@ -1,46 +0,0 @@ -{ - "alora_invocation_tokens": null, - "alpha_pattern": {}, - "arrow_config": null, - "auto_mapping": null, - "base_model_name_or_path": "Qwen/Qwen3.5-2B-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": 256, - "lora_bias": false, - "lora_dropout": 0.027050353915701653, - "megatron_config": null, - "megatron_core": "megatron.core", - "modules_to_save": null, - "peft_type": "LORA", - "peft_version": "0.18.1", - "qalora_group_size": 16, - "r": 128, - "rank_pattern": {}, - "revision": null, - "target_modules": [ - "o_proj", - "q_proj", - "k_proj", - "down_proj", - "up_proj", - "v_proj", - "gate_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/substitutivity_original_Estonian/Qwen3.5-2B-Base_substitutivity_splits_original_features_train_substitutivity_splits_original_features_test1/checkpoint-415/adapter_model.safetensors b/substitutivity_original_Estonian/Qwen3.5-2B-Base_substitutivity_splits_original_features_train_substitutivity_splits_original_features_test1/checkpoint-415/adapter_model.safetensors deleted file mode 100644 index 3eb4ea310de09a0b71903fa9c6f373189807cb06..0000000000000000000000000000000000000000 --- a/substitutivity_original_Estonian/Qwen3.5-2B-Base_substitutivity_splits_original_features_train_substitutivity_splits_original_features_test1/checkpoint-415/adapter_model.safetensors +++ /dev/null @@ -1,3 +0,0 @@ -version https://git-lfs.github.com/spec/v1 -oid sha256:78beede6004d6863ca2db97154fbdbb59e18c3e99c4a1e5dc4e4d3b293938057 -size 349201608 diff --git a/substitutivity_original_Estonian/Qwen3.5-2B-Base_substitutivity_splits_original_features_train_substitutivity_splits_original_features_test1/checkpoint-415/chat_template.jinja b/substitutivity_original_Estonian/Qwen3.5-2B-Base_substitutivity_splits_original_features_train_substitutivity_splits_original_features_test1/checkpoint-415/chat_template.jinja deleted file mode 100644 index 0ef09f214eaa6d9bca297988afc1454b5827b2c7..0000000000000000000000000000000000000000 --- a/substitutivity_original_Estonian/Qwen3.5-2B-Base_substitutivity_splits_original_features_train_substitutivity_splits_original_features_test1/checkpoint-415/chat_template.jinja +++ /dev/null @@ -1,154 +0,0 @@ -{%- set image_count = namespace(value=0) %} -{%- set video_count = namespace(value=0) %} -{%- macro render_content(content, do_vision_count, is_system_content=false) %} - {%- if content is string %} - {{- content }} - {%- elif content is iterable and content is not mapping %} - {%- for item in content %} - {%- if 'image' in item or 'image_url' in item or item.type == 'image' %} - {%- if is_system_content %} - {{- raise_exception('System message cannot contain images.') }} - {%- endif %} - {%- if do_vision_count %} - {%- set image_count.value = image_count.value + 1 %} - {%- endif %} - {%- if add_vision_id %} - {{- 'Picture ' ~ image_count.value ~ ': ' }} - {%- endif %} - {{- '<|vision_start|><|image_pad|><|vision_end|>' }} - {%- elif 'video' in item or item.type == 'video' %} - {%- if is_system_content %} - {{- raise_exception('System message cannot contain videos.') }} - {%- endif %} - {%- if do_vision_count %} - {%- set video_count.value = video_count.value + 1 %} - {%- endif %} - {%- if add_vision_id %} - {{- 'Video ' ~ video_count.value ~ ': ' }} - {%- endif %} - {{- '<|vision_start|><|video_pad|><|vision_end|>' }} - {%- elif 'text' in item %} - {{- item.text }} - {%- else %} - {{- raise_exception('Unexpected item type in content.') }} - {%- endif %} - {%- endfor %} - {%- elif content is none or content is undefined %} - {{- '' }} - {%- else %} - {{- raise_exception('Unexpected content type.') }} - {%- endif %} -{%- endmacro %} -{%- if not messages %} - {{- raise_exception('No messages provided.') }} -{%- endif %} -{%- if tools and tools is iterable and tools is not mapping %} - {{- '<|im_start|>system\n' }} - {{- "# Tools\n\nYou have access to the following functions:\n\n" }} - {%- for tool in tools %} - {{- "\n" }} - {{- tool | tojson }} - {%- endfor %} - {{- "\n" }} - {{- '\n\nIf you choose to call a function ONLY reply in the following format with NO suffix:\n\n\n\n\nvalue_1\n\n\nThis is the value for the second parameter\nthat can span\nmultiple lines\n\n\n\n\n\nReminder:\n- Function calls MUST follow the specified format: an inner block must be nested within XML tags\n- Required parameters MUST be specified\n- You may provide optional reasoning for your function call in natural language BEFORE the function call, but NOT after\n- If there is no function call available, answer the question like normal with your current knowledge and do not tell the user about function calls\n' }} - {%- if messages[0].role == 'system' %} - {%- set content = render_content(messages[0].content, false, true)|trim %} - {%- if content %} - {{- '\n\n' + content }} - {%- endif %} - {%- endif %} - {{- '<|im_end|>\n' }} -{%- else %} - {%- if messages[0].role == 'system' %} - {%- set content = render_content(messages[0].content, false, true)|trim %} - {{- '<|im_start|>system\n' + 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" %} - {%- set content = render_content(message.content, false)|trim %} - {%- if not(content.startswith('') and content.endswith('')) %} - {%- set ns.multi_step_tool = false %} - {%- set ns.last_query_index = index %} - {%- endif %} - {%- endif %} -{%- endfor %} -{%- if ns.multi_step_tool %} - {{- raise_exception('No user query found in messages.') }} -{%- endif %} -{%- for message in messages %} - {%- set content = render_content(message.content, true)|trim %} - {%- if message.role == "system" %} - {%- if not loop.first %} - {{- raise_exception('System message must be at the beginning.') }} - {%- endif %} - {%- elif message.role == "user" %} - {{- '<|im_start|>' + message.role + '\n' + content + '<|im_end|>' + '\n' }} - {%- elif message.role == "assistant" %} - {%- set reasoning_content = '' %} - {%- if message.reasoning_content is string %} - {%- set reasoning_content = message.reasoning_content %} - {%- else %} - {%- if '' in content %} - {%- set reasoning_content = content.split('')[0].rstrip('\n').split('')[-1].lstrip('\n') %} - {%- set content = content.split('')[-1].lstrip('\n') %} - {%- endif %} - {%- endif %} - {%- set reasoning_content = reasoning_content|trim %} - {%- if loop.index0 > ns.last_query_index %} - {{- '<|im_start|>' + message.role + '\n\n' + reasoning_content + '\n\n\n' + content }} - {%- else %} - {{- '<|im_start|>' + message.role + '\n' + content }} - {%- endif %} - {%- if message.tool_calls and message.tool_calls is iterable and message.tool_calls is not mapping %} - {%- for tool_call in message.tool_calls %} - {%- if tool_call.function is defined %} - {%- set tool_call = tool_call.function %} - {%- endif %} - {%- if loop.first %} - {%- if content|trim %} - {{- '\n\n\n\n' }} - {%- else %} - {{- '\n\n' }} - {%- endif %} - {%- else %} - {{- '\n\n\n' }} - {%- endif %} - {%- if tool_call.arguments is defined %} - {%- for args_name, args_value in tool_call.arguments|items %} - {{- '\n' }} - {%- set args_value = args_value | tojson | safe if args_value is mapping or (args_value is sequence and args_value is not string) else args_value | string %} - {{- args_value }} - {{- '\n\n' }} - {%- endfor %} - {%- endif %} - {{- '\n' }} - {%- endfor %} - {%- endif %} - {{- '<|im_end|>\n' }} - {%- elif message.role == "tool" %} - {%- if loop.previtem and loop.previtem.role != "tool" %} - {{- '<|im_start|>user' }} - {%- endif %} - {{- '\n\n' }} - {{- content }} - {{- '\n' }} - {%- if not loop.last and loop.nextitem.role != "tool" %} - {{- '<|im_end|>\n' }} - {%- elif loop.last %} - {{- '<|im_end|>\n' }} - {%- endif %} - {%- else %} - {{- raise_exception('Unexpected message role.') }} - {%- endif %} -{%- endfor %} -{%- if add_generation_prompt %} - {{- '<|im_start|>assistant\n' }} - {%- if enable_thinking is defined and enable_thinking is true %} - {{- '\n' }} - {%- else %} - {{- '\n\n\n\n' }} - {%- endif %} -{%- endif %} \ No newline at end of file diff --git a/substitutivity_original_Estonian/Qwen3.5-2B-Base_substitutivity_splits_original_features_train_substitutivity_splits_original_features_test1/checkpoint-415/tokenizer.json b/substitutivity_original_Estonian/Qwen3.5-2B-Base_substitutivity_splits_original_features_train_substitutivity_splits_original_features_test1/checkpoint-415/tokenizer.json deleted file mode 100644 index 67741b04f23bfdb46501f748ce27865ec82eccfb..0000000000000000000000000000000000000000 --- a/substitutivity_original_Estonian/Qwen3.5-2B-Base_substitutivity_splits_original_features_train_substitutivity_splits_original_features_test1/checkpoint-415/tokenizer.json +++ /dev/null @@ -1,3 +0,0 @@ -version https://git-lfs.github.com/spec/v1 -oid sha256:87a7830d63fcf43bf241c3c5242e96e62dd3fdc29224ca26fed8ea333db72de4 -size 19989343 diff --git a/substitutivity_original_Estonian/Qwen3.5-2B-Base_substitutivity_splits_original_features_train_substitutivity_splits_original_features_test1/checkpoint-415/tokenizer_config.json b/substitutivity_original_Estonian/Qwen3.5-2B-Base_substitutivity_splits_original_features_train_substitutivity_splits_original_features_test1/checkpoint-415/tokenizer_config.json deleted file mode 100644 index b4a37b2a6fd3ab3317cd7bac72855be1a843b2bb..0000000000000000000000000000000000000000 --- a/substitutivity_original_Estonian/Qwen3.5-2B-Base_substitutivity_splits_original_features_train_substitutivity_splits_original_features_test1/checkpoint-415/tokenizer_config.json +++ /dev/null @@ -1,31 +0,0 @@ -{ - "add_prefix_space": false, - "audio_bos_token": "<|audio_start|>", - "audio_eos_token": "<|audio_end|>", - "audio_token": "<|audio_pad|>", - "backend": "tokenizers", - "bos_token": null, - "clean_up_tokenization_spaces": false, - "eos_token": "<|endoftext|>", - "errors": "replace", - "image_token": "<|image_pad|>", - "is_local": false, - "model_max_length": 262144, - "model_specific_special_tokens": { - "audio_bos_token": "<|audio_start|>", - "audio_eos_token": "<|audio_end|>", - "audio_token": "<|audio_pad|>", - "image_token": "<|image_pad|>", - "video_token": "<|video_pad|>", - "vision_bos_token": "<|vision_start|>", - "vision_eos_token": "<|vision_end|>" - }, - "pad_token": "<|endoftext|>", - "pretokenize_regex": "(?i:'s|'t|'re|'ve|'m|'ll|'d)|[^\\r\\n\\p{L}\\p{N}]?[\\p{L}\\p{M}]+|\\p{N}| ?[^\\s\\p{L}\\p{M}\\p{N}]+[\\r\\n]*|\\s*[\\r\\n]+|\\s+(?!\\S)|\\s+", - "split_special_tokens": false, - "tokenizer_class": "TokenizersBackend", - "unk_token": null, - "video_token": "<|video_pad|>", - "vision_bos_token": "<|vision_start|>", - "vision_eos_token": "<|vision_end|>" -} diff --git a/substitutivity_original_Estonian/Qwen3.5-2B-Base_substitutivity_splits_original_features_train_substitutivity_splits_original_features_test1/checkpoint-415/trainer_state.json b/substitutivity_original_Estonian/Qwen3.5-2B-Base_substitutivity_splits_original_features_train_substitutivity_splits_original_features_test1/checkpoint-415/trainer_state.json deleted file mode 100644 index 72380c9a708258a7458c63f015239be1f3793053..0000000000000000000000000000000000000000 --- a/substitutivity_original_Estonian/Qwen3.5-2B-Base_substitutivity_splits_original_features_train_substitutivity_splits_original_features_test1/checkpoint-415/trainer_state.json +++ /dev/null @@ -1,136 +0,0 @@ -{ - "best_global_step": null, - "best_metric": null, - "best_model_checkpoint": null, - "epoch": 1.0, - "eval_steps": 500, - "global_step": 415, - "is_hyper_param_search": false, - "is_local_process_zero": true, - "is_world_process_zero": true, - "log_history": [ - { - "entropy": 2.0473356211185454, - "epoch": 0.12062726176115803, - "grad_norm": 3.0554118156433105, - "learning_rate": 1.5682344549790923e-05, - "loss": 1.910550079345703, - "mean_token_accuracy": 0.6053373458981514, - "num_tokens": 121428.0, - "step": 50 - }, - { - "entropy": 1.0371007251739501, - "epoch": 0.24125452352231605, - "grad_norm": 2.109868049621582, - "learning_rate": 3.168473694753677e-05, - "loss": 0.9585584259033203, - "mean_token_accuracy": 0.7352996730804443, - "num_tokens": 243966.0, - "step": 100 - }, - { - "entropy": 0.9105490958690643, - "epoch": 0.3618817852834741, - "grad_norm": 1.7675701379776, - "learning_rate": 4.7687129345282605e-05, - "loss": 0.8393524932861328, - "mean_token_accuracy": 0.7588736474514007, - "num_tokens": 366649.0, - "step": 150 - }, - { - "entropy": 0.8660378754138947, - "epoch": 0.4825090470446321, - "grad_norm": 1.6502987146377563, - "learning_rate": 6.368952174302845e-05, - "loss": 0.8005844116210937, - "mean_token_accuracy": 0.7663125538825989, - "num_tokens": 488367.0, - "step": 200 - }, - { - "entropy": 0.8439123743772506, - "epoch": 0.6031363088057901, - "grad_norm": 1.538365364074707, - "learning_rate": 7.969191414077429e-05, - "loss": 0.7773619842529297, - "mean_token_accuracy": 0.7718527519702911, - "num_tokens": 603131.0, - "step": 250 - }, - { - "entropy": 0.8223806774616241, - "epoch": 0.7237635705669482, - "grad_norm": 1.1559664011001587, - "learning_rate": 9.569430653852014e-05, - "loss": 0.7587767791748047, - "mean_token_accuracy": 0.7767325991392136, - "num_tokens": 726412.0, - "step": 300 - }, - { - "entropy": 0.8256935775279999, - "epoch": 0.8443908323281062, - "grad_norm": 1.2322715520858765, - "learning_rate": 0.00011169669893626597, - "loss": 0.769886245727539, - "mean_token_accuracy": 0.773840668797493, - "num_tokens": 846516.0, - "step": 350 - }, - { - "entropy": 0.8065843790769577, - "epoch": 0.9650180940892642, - "grad_norm": 0.9908286333084106, - "learning_rate": 0.00012769909133401183, - "loss": 0.7451084899902344, - "mean_token_accuracy": 0.7794792425632476, - "num_tokens": 958978.0, - "step": 400 - }, - { - "epoch": 1.0, - "eval_entropy": 0.791056221790528, - "eval_mean_token_accuracy": 0.7507295300451557, - "eval_not_syn_loss": 0.7860428094863892, - "eval_not_syn_runtime": 53.5431, - "eval_not_syn_samples_per_second": 26.521, - "eval_not_syn_steps_per_second": 3.324, - "eval_num_tokens": 993806.0, - "step": 415 - }, - { - "epoch": 1.0, - "eval_entropy": 0.7424047408478983, - "eval_mean_token_accuracy": 0.8023804037758474, - "eval_num_tokens": 993806.0, - "eval_syn_loss": 0.7273509502410889, - "eval_syn_runtime": 54.8679, - "eval_syn_samples_per_second": 25.88, - "eval_syn_steps_per_second": 3.244, - "step": 415 - } - ], - "logging_steps": 50, - "max_steps": 4150, - "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": 2.123058450594432e+16, - "train_batch_size": 8, - "trial_name": null, - "trial_params": null -} diff --git a/substitutivity_original_Estonian/Qwen3.5-2B-Base_substitutivity_splits_original_features_train_substitutivity_splits_original_features_test1/checkpoint-415/training_args.bin b/substitutivity_original_Estonian/Qwen3.5-2B-Base_substitutivity_splits_original_features_train_substitutivity_splits_original_features_test1/checkpoint-415/training_args.bin deleted file mode 100644 index 48901efb30506830185bfaafecbe236e7095a006..0000000000000000000000000000000000000000 --- a/substitutivity_original_Estonian/Qwen3.5-2B-Base_substitutivity_splits_original_features_train_substitutivity_splits_original_features_test1/checkpoint-415/training_args.bin +++ /dev/null @@ -1,3 +0,0 @@ -version https://git-lfs.github.com/spec/v1 -oid sha256:770fc15ebc27a54d780af597d32c57ca14b3b528f14788d142bdfce09b555bef -size 6033 diff --git a/substitutivity_original_Estonian/Qwen3.5-2B-Base_substitutivity_splits_original_features_train_substitutivity_splits_original_features_test1/checkpoint-4150/README.md b/substitutivity_original_Estonian/Qwen3.5-2B-Base_substitutivity_splits_original_features_train_substitutivity_splits_original_features_test1/checkpoint-4150/README.md deleted file mode 100644 index ece204d005f7bf161d0e448efdd56fd297b7c8a2..0000000000000000000000000000000000000000 --- a/substitutivity_original_Estonian/Qwen3.5-2B-Base_substitutivity_splits_original_features_train_substitutivity_splits_original_features_test1/checkpoint-4150/README.md +++ /dev/null @@ -1,209 +0,0 @@ ---- -base_model: Qwen/Qwen3.5-2B-Base -library_name: peft -pipeline_tag: text-generation -tags: -- base_model:adapter:Qwen/Qwen3.5-2B-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/substitutivity_original_Estonian/Qwen3.5-2B-Base_substitutivity_splits_original_features_train_substitutivity_splits_original_features_test1/checkpoint-4150/adapter_config.json b/substitutivity_original_Estonian/Qwen3.5-2B-Base_substitutivity_splits_original_features_train_substitutivity_splits_original_features_test1/checkpoint-4150/adapter_config.json deleted file mode 100644 index a64145ce23adcbdbe246679087a13f4cb63ff664..0000000000000000000000000000000000000000 --- a/substitutivity_original_Estonian/Qwen3.5-2B-Base_substitutivity_splits_original_features_train_substitutivity_splits_original_features_test1/checkpoint-4150/adapter_config.json +++ /dev/null @@ -1,46 +0,0 @@ -{ - "alora_invocation_tokens": null, - "alpha_pattern": {}, - "arrow_config": null, - "auto_mapping": null, - "base_model_name_or_path": "Qwen/Qwen3.5-2B-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": 256, - "lora_bias": false, - "lora_dropout": 0.027050353915701653, - "megatron_config": null, - "megatron_core": "megatron.core", - "modules_to_save": null, - "peft_type": "LORA", - "peft_version": "0.18.1", - "qalora_group_size": 16, - "r": 128, - "rank_pattern": {}, - "revision": null, - "target_modules": [ - "o_proj", - "q_proj", - "k_proj", - "down_proj", - "up_proj", - "v_proj", - "gate_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/substitutivity_original_Estonian/Qwen3.5-2B-Base_substitutivity_splits_original_features_train_substitutivity_splits_original_features_test1/checkpoint-4150/adapter_model.safetensors b/substitutivity_original_Estonian/Qwen3.5-2B-Base_substitutivity_splits_original_features_train_substitutivity_splits_original_features_test1/checkpoint-4150/adapter_model.safetensors deleted file mode 100644 index 717c80837a70125c5a38bf31a819799cd97814a2..0000000000000000000000000000000000000000 --- a/substitutivity_original_Estonian/Qwen3.5-2B-Base_substitutivity_splits_original_features_train_substitutivity_splits_original_features_test1/checkpoint-4150/adapter_model.safetensors +++ /dev/null @@ -1,3 +0,0 @@ -version https://git-lfs.github.com/spec/v1 -oid sha256:64fb9ee4b656a3c873bb1b6a6ee88076eceaa9fb49a912f2686cc66094bb37cb -size 349201608 diff --git a/substitutivity_original_Estonian/Qwen3.5-2B-Base_substitutivity_splits_original_features_train_substitutivity_splits_original_features_test1/checkpoint-4150/chat_template.jinja b/substitutivity_original_Estonian/Qwen3.5-2B-Base_substitutivity_splits_original_features_train_substitutivity_splits_original_features_test1/checkpoint-4150/chat_template.jinja deleted file mode 100644 index 0ef09f214eaa6d9bca297988afc1454b5827b2c7..0000000000000000000000000000000000000000 --- a/substitutivity_original_Estonian/Qwen3.5-2B-Base_substitutivity_splits_original_features_train_substitutivity_splits_original_features_test1/checkpoint-4150/chat_template.jinja +++ /dev/null @@ -1,154 +0,0 @@ -{%- set image_count = namespace(value=0) %} -{%- set video_count = namespace(value=0) %} -{%- macro render_content(content, do_vision_count, is_system_content=false) %} - {%- if content is string %} - {{- content }} - {%- elif content is iterable and content is not mapping %} - {%- for item in content %} - {%- if 'image' in item or 'image_url' in item or item.type == 'image' %} - {%- if is_system_content %} - {{- raise_exception('System message cannot contain images.') }} - {%- endif %} - {%- if do_vision_count %} - {%- set image_count.value = image_count.value + 1 %} - {%- endif %} - {%- if add_vision_id %} - {{- 'Picture ' ~ image_count.value ~ ': ' }} - {%- endif %} - {{- '<|vision_start|><|image_pad|><|vision_end|>' }} - {%- elif 'video' in item or item.type == 'video' %} - {%- if is_system_content %} - {{- raise_exception('System message cannot contain videos.') }} - {%- endif %} - {%- if do_vision_count %} - {%- set video_count.value = video_count.value + 1 %} - {%- endif %} - {%- if add_vision_id %} - {{- 'Video ' ~ video_count.value ~ ': ' }} - {%- endif %} - {{- '<|vision_start|><|video_pad|><|vision_end|>' }} - {%- elif 'text' in item %} - {{- item.text }} - {%- else %} - {{- raise_exception('Unexpected item type in content.') }} - {%- endif %} - {%- endfor %} - {%- elif content is none or content is undefined %} - {{- '' }} - {%- else %} - {{- raise_exception('Unexpected content type.') }} - {%- endif %} -{%- endmacro %} -{%- if not messages %} - {{- raise_exception('No messages provided.') }} -{%- endif %} -{%- if tools and tools is iterable and tools is not mapping %} - {{- '<|im_start|>system\n' }} - {{- "# Tools\n\nYou have access to the following functions:\n\n" }} - {%- for tool in tools %} - {{- "\n" }} - {{- tool | tojson }} - {%- endfor %} - {{- "\n" }} - {{- '\n\nIf you choose to call a function ONLY reply in the following format with NO suffix:\n\n\n\n\nvalue_1\n\n\nThis is the value for the second parameter\nthat can span\nmultiple lines\n\n\n\n\n\nReminder:\n- Function calls MUST follow the specified format: an inner block must be nested within XML tags\n- Required parameters MUST be specified\n- You may provide optional reasoning for your function call in natural language BEFORE the function call, but NOT after\n- If there is no function call available, answer the question like normal with your current knowledge and do not tell the user about function calls\n' }} - {%- if messages[0].role == 'system' %} - {%- set content = render_content(messages[0].content, false, true)|trim %} - {%- if content %} - {{- '\n\n' + content }} - {%- endif %} - {%- endif %} - {{- '<|im_end|>\n' }} -{%- else %} - {%- if messages[0].role == 'system' %} - {%- set content = render_content(messages[0].content, false, true)|trim %} - {{- '<|im_start|>system\n' + 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" %} - {%- set content = render_content(message.content, false)|trim %} - {%- if not(content.startswith('') and content.endswith('')) %} - {%- set ns.multi_step_tool = false %} - {%- set ns.last_query_index = index %} - {%- endif %} - {%- endif %} -{%- endfor %} -{%- if ns.multi_step_tool %} - {{- raise_exception('No user query found in messages.') }} -{%- endif %} -{%- for message in messages %} - {%- set content = render_content(message.content, true)|trim %} - {%- if message.role == "system" %} - {%- if not loop.first %} - {{- raise_exception('System message must be at the beginning.') }} - {%- endif %} - {%- elif message.role == "user" %} - {{- '<|im_start|>' + message.role + '\n' + content + '<|im_end|>' + '\n' }} - {%- elif message.role == "assistant" %} - {%- set reasoning_content = '' %} - {%- if message.reasoning_content is string %} - {%- set reasoning_content = message.reasoning_content %} - {%- else %} - {%- if '' in content %} - {%- set reasoning_content = content.split('')[0].rstrip('\n').split('')[-1].lstrip('\n') %} - {%- set content = content.split('')[-1].lstrip('\n') %} - {%- endif %} - {%- endif %} - {%- set reasoning_content = reasoning_content|trim %} - {%- if loop.index0 > ns.last_query_index %} - {{- '<|im_start|>' + message.role + '\n\n' + reasoning_content + '\n\n\n' + content }} - {%- else %} - {{- '<|im_start|>' + message.role + '\n' + content }} - {%- endif %} - {%- if message.tool_calls and message.tool_calls is iterable and message.tool_calls is not mapping %} - {%- for tool_call in message.tool_calls %} - {%- if tool_call.function is defined %} - {%- set tool_call = tool_call.function %} - {%- endif %} - {%- if loop.first %} - {%- if content|trim %} - {{- '\n\n\n\n' }} - {%- else %} - {{- '\n\n' }} - {%- endif %} - {%- else %} - {{- '\n\n\n' }} - {%- endif %} - {%- if tool_call.arguments is defined %} - {%- for args_name, args_value in tool_call.arguments|items %} - {{- '\n' }} - {%- set args_value = args_value | tojson | safe if args_value is mapping or (args_value is sequence and args_value is not string) else args_value | string %} - {{- args_value }} - {{- '\n\n' }} - {%- endfor %} - {%- endif %} - {{- '\n' }} - {%- endfor %} - {%- endif %} - {{- '<|im_end|>\n' }} - {%- elif message.role == "tool" %} - {%- if loop.previtem and loop.previtem.role != "tool" %} - {{- '<|im_start|>user' }} - {%- endif %} - {{- '\n\n' }} - {{- content }} - {{- '\n' }} - {%- if not loop.last and loop.nextitem.role != "tool" %} - {{- '<|im_end|>\n' }} - {%- elif loop.last %} - {{- '<|im_end|>\n' }} - {%- endif %} - {%- else %} - {{- raise_exception('Unexpected message role.') }} - {%- endif %} -{%- endfor %} -{%- if add_generation_prompt %} - {{- '<|im_start|>assistant\n' }} - {%- if enable_thinking is defined and enable_thinking is true %} - {{- '\n' }} - {%- else %} - {{- '\n\n\n\n' }} - {%- endif %} -{%- endif %} \ No newline at end of file diff --git a/substitutivity_original_Estonian/Qwen3.5-2B-Base_substitutivity_splits_original_features_train_substitutivity_splits_original_features_test1/checkpoint-4150/tokenizer.json b/substitutivity_original_Estonian/Qwen3.5-2B-Base_substitutivity_splits_original_features_train_substitutivity_splits_original_features_test1/checkpoint-4150/tokenizer.json deleted file mode 100644 index 67741b04f23bfdb46501f748ce27865ec82eccfb..0000000000000000000000000000000000000000 --- a/substitutivity_original_Estonian/Qwen3.5-2B-Base_substitutivity_splits_original_features_train_substitutivity_splits_original_features_test1/checkpoint-4150/tokenizer.json +++ /dev/null @@ -1,3 +0,0 @@ -version https://git-lfs.github.com/spec/v1 -oid sha256:87a7830d63fcf43bf241c3c5242e96e62dd3fdc29224ca26fed8ea333db72de4 -size 19989343 diff --git a/substitutivity_original_Estonian/Qwen3.5-2B-Base_substitutivity_splits_original_features_train_substitutivity_splits_original_features_test1/checkpoint-4150/tokenizer_config.json b/substitutivity_original_Estonian/Qwen3.5-2B-Base_substitutivity_splits_original_features_train_substitutivity_splits_original_features_test1/checkpoint-4150/tokenizer_config.json deleted file mode 100644 index b4a37b2a6fd3ab3317cd7bac72855be1a843b2bb..0000000000000000000000000000000000000000 --- a/substitutivity_original_Estonian/Qwen3.5-2B-Base_substitutivity_splits_original_features_train_substitutivity_splits_original_features_test1/checkpoint-4150/tokenizer_config.json +++ /dev/null @@ -1,31 +0,0 @@ -{ - "add_prefix_space": false, - "audio_bos_token": "<|audio_start|>", - "audio_eos_token": "<|audio_end|>", - "audio_token": "<|audio_pad|>", - "backend": "tokenizers", - "bos_token": null, - "clean_up_tokenization_spaces": false, - "eos_token": "<|endoftext|>", - "errors": "replace", - "image_token": "<|image_pad|>", - "is_local": false, - "model_max_length": 262144, - "model_specific_special_tokens": { - "audio_bos_token": "<|audio_start|>", - "audio_eos_token": "<|audio_end|>", - "audio_token": "<|audio_pad|>", - "image_token": "<|image_pad|>", - "video_token": "<|video_pad|>", - "vision_bos_token": "<|vision_start|>", - "vision_eos_token": "<|vision_end|>" - }, - "pad_token": "<|endoftext|>", - "pretokenize_regex": "(?i:'s|'t|'re|'ve|'m|'ll|'d)|[^\\r\\n\\p{L}\\p{N}]?[\\p{L}\\p{M}]+|\\p{N}| ?[^\\s\\p{L}\\p{M}\\p{N}]+[\\r\\n]*|\\s*[\\r\\n]+|\\s+(?!\\S)|\\s+", - "split_special_tokens": false, - "tokenizer_class": "TokenizersBackend", - "unk_token": null, - "video_token": "<|video_pad|>", - "vision_bos_token": "<|vision_start|>", - "vision_eos_token": "<|vision_end|>" -} diff --git a/substitutivity_original_Estonian/Qwen3.5-2B-Base_substitutivity_splits_original_features_train_substitutivity_splits_original_features_test1/checkpoint-4150/trainer_state.json b/substitutivity_original_Estonian/Qwen3.5-2B-Base_substitutivity_splits_original_features_train_substitutivity_splits_original_features_test1/checkpoint-4150/trainer_state.json deleted file mode 100644 index 6f7ed45b2cf55b08ce9ad0c31c82a42a820bd49c..0000000000000000000000000000000000000000 --- a/substitutivity_original_Estonian/Qwen3.5-2B-Base_substitutivity_splits_original_features_train_substitutivity_splits_original_features_test1/checkpoint-4150/trainer_state.json +++ /dev/null @@ -1,1084 +0,0 @@ -{ - "best_global_step": null, - "best_metric": null, - "best_model_checkpoint": null, - "epoch": 10.0, - "eval_steps": 500, - "global_step": 4150, - "is_hyper_param_search": false, - "is_local_process_zero": true, - "is_world_process_zero": true, - "log_history": [ - { - "entropy": 2.0473356211185454, - "epoch": 0.12062726176115803, - "grad_norm": 3.0554118156433105, - "learning_rate": 1.5682344549790923e-05, - "loss": 1.910550079345703, - "mean_token_accuracy": 0.6053373458981514, - "num_tokens": 121428.0, - "step": 50 - }, - { - "entropy": 1.0371007251739501, - "epoch": 0.24125452352231605, - "grad_norm": 2.109868049621582, - "learning_rate": 3.168473694753677e-05, - "loss": 0.9585584259033203, - "mean_token_accuracy": 0.7352996730804443, - "num_tokens": 243966.0, - "step": 100 - }, - { - "entropy": 0.9105490958690643, - "epoch": 0.3618817852834741, - "grad_norm": 1.7675701379776, - "learning_rate": 4.7687129345282605e-05, - "loss": 0.8393524932861328, - "mean_token_accuracy": 0.7588736474514007, - "num_tokens": 366649.0, - "step": 150 - }, - { - "entropy": 0.8660378754138947, - "epoch": 0.4825090470446321, - "grad_norm": 1.6502987146377563, - "learning_rate": 6.368952174302845e-05, - "loss": 0.8005844116210937, - "mean_token_accuracy": 0.7663125538825989, - "num_tokens": 488367.0, - "step": 200 - }, - { - "entropy": 0.8439123743772506, - "epoch": 0.6031363088057901, - "grad_norm": 1.538365364074707, - "learning_rate": 7.969191414077429e-05, - "loss": 0.7773619842529297, - "mean_token_accuracy": 0.7718527519702911, - "num_tokens": 603131.0, - "step": 250 - }, - { - "entropy": 0.8223806774616241, - "epoch": 0.7237635705669482, - "grad_norm": 1.1559664011001587, - "learning_rate": 9.569430653852014e-05, - "loss": 0.7587767791748047, - "mean_token_accuracy": 0.7767325991392136, - "num_tokens": 726412.0, - "step": 300 - }, - { - "entropy": 0.8256935775279999, - "epoch": 0.8443908323281062, - "grad_norm": 1.2322715520858765, - "learning_rate": 0.00011169669893626597, - "loss": 0.769886245727539, - "mean_token_accuracy": 0.773840668797493, - "num_tokens": 846516.0, - "step": 350 - }, - { - "entropy": 0.8065843790769577, - "epoch": 0.9650180940892642, - "grad_norm": 0.9908286333084106, - "learning_rate": 0.00012769909133401183, - "loss": 0.7451084899902344, - "mean_token_accuracy": 0.7794792425632476, - "num_tokens": 958978.0, - "step": 400 - }, - { - "epoch": 1.0, - "eval_entropy": 0.791056221790528, - "eval_mean_token_accuracy": 0.7507295300451557, - "eval_not_syn_loss": 0.7860428094863892, - "eval_not_syn_runtime": 53.5431, - "eval_not_syn_samples_per_second": 26.521, - "eval_not_syn_steps_per_second": 3.324, - "eval_num_tokens": 993806.0, - "step": 415 - }, - { - "epoch": 1.0, - "eval_entropy": 0.7424047408478983, - "eval_mean_token_accuracy": 0.8023804037758474, - "eval_num_tokens": 993806.0, - "eval_syn_loss": 0.7273509502410889, - "eval_syn_runtime": 54.8679, - "eval_syn_samples_per_second": 25.88, - "eval_syn_steps_per_second": 3.244, - "step": 415 - }, - { - "entropy": 0.7607260761838971, - "epoch": 1.0844390832328106, - "grad_norm": 1.0653696060180664, - "learning_rate": 0.0001327927019027069, - "loss": 0.7034613037109375, - "mean_token_accuracy": 0.7869195998317063, - "num_tokens": 1080301.0, - "step": 450 - }, - { - "entropy": 0.7370894801616669, - "epoch": 1.2050663449939687, - "grad_norm": 1.763787865638733, - "learning_rate": 0.00013265416569994628, - "loss": 0.6816973114013671, - "mean_token_accuracy": 0.7914507347345352, - "num_tokens": 1201215.0, - "step": 500 - }, - { - "entropy": 0.742829909324646, - "epoch": 1.3256936067551266, - "grad_norm": 0.97663414478302, - "learning_rate": 0.00013239847940553975, - "loss": 0.6867338562011719, - "mean_token_accuracy": 0.790328289270401, - "num_tokens": 1322895.0, - "step": 550 - }, - { - "entropy": 0.7312483406066894, - "epoch": 1.4463208685162847, - "grad_norm": 1.0874731540679932, - "learning_rate": 0.00013202609518972978, - "loss": 0.676800537109375, - "mean_token_accuracy": 0.7921512228250503, - "num_tokens": 1436749.0, - "step": 600 - }, - { - "entropy": 0.7178869074583054, - "epoch": 1.5669481302774426, - "grad_norm": 1.064300298690796, - "learning_rate": 0.0001315376715980265, - "loss": 0.6635584259033203, - "mean_token_accuracy": 0.7944776183366775, - "num_tokens": 1555400.0, - "step": 650 - }, - { - "entropy": 0.7198166918754577, - "epoch": 1.6875753920386007, - "grad_norm": 0.7779315114021301, - "learning_rate": 0.00013093407238659802, - "loss": 0.6660567474365234, - "mean_token_accuracy": 0.7968300533294678, - "num_tokens": 1676892.0, - "step": 700 - }, - { - "entropy": 0.7170061504840851, - "epoch": 1.8082026537997589, - "grad_norm": 1.089854121208191, - "learning_rate": 0.00013021636499475496, - "loss": 0.6632771301269531, - "mean_token_accuracy": 0.7963915091753005, - "num_tokens": 1802689.0, - "step": 750 - }, - { - "entropy": 0.6953610718250275, - "epoch": 1.9288299155609168, - "grad_norm": 0.7996730208396912, - "learning_rate": 0.0001293858186572295, - "loss": 0.6421419525146485, - "mean_token_accuracy": 0.8002577543258667, - "num_tokens": 1920804.0, - "step": 800 - }, - { - "epoch": 2.0, - "eval_entropy": 0.7369416523515508, - "eval_mean_token_accuracy": 0.7872022366925572, - "eval_not_syn_loss": 0.713637113571167, - "eval_not_syn_runtime": 53.2265, - "eval_not_syn_samples_per_second": 26.678, - "eval_not_syn_steps_per_second": 3.344, - "eval_num_tokens": 1987612.0, - "step": 830 - }, - { - "epoch": 2.0, - "eval_entropy": 0.6877032756135705, - "eval_mean_token_accuracy": 0.7997466372639945, - "eval_num_tokens": 1987612.0, - "eval_syn_loss": 0.6614266633987427, - "eval_syn_runtime": 54.5178, - "eval_syn_samples_per_second": 26.047, - "eval_syn_steps_per_second": 3.265, - "step": 830 - }, - { - "entropy": 0.6644553433765065, - "epoch": 2.0482509047044632, - "grad_norm": 0.7562468647956848, - "learning_rate": 0.00012844390215958852, - "loss": 0.6112185287475586, - "mean_token_accuracy": 0.806649933559726, - "num_tokens": 2039660.0, - "step": 850 - }, - { - "entropy": 0.6146580249071121, - "epoch": 2.168878166465621, - "grad_norm": 1.098607063293457, - "learning_rate": 0.00012739228124074913, - "loss": 0.5553749084472657, - "mean_token_accuracy": 0.8187588346004486, - "num_tokens": 2161772.0, - "step": 900 - }, - { - "entropy": 0.6187627577781677, - "epoch": 2.2895054282267795, - "grad_norm": 1.0338040590286255, - "learning_rate": 0.00012623281564719095, - "loss": 0.5596194076538086, - "mean_token_accuracy": 0.8182566928863525, - "num_tokens": 2280833.0, - "step": 950 - }, - { - "entropy": 0.6232365870475769, - "epoch": 2.4101326899879374, - "grad_norm": 1.0010799169540405, - "learning_rate": 0.00012496755584407422, - "loss": 0.5667143249511719, - "mean_token_accuracy": 0.8175061982870102, - "num_tokens": 2398206.0, - "step": 1000 - }, - { - "entropy": 0.6151268655061721, - "epoch": 2.5307599517490953, - "grad_norm": 1.0065034627914429, - "learning_rate": 0.0001235987393890801, - "loss": 0.5598890686035156, - "mean_token_accuracy": 0.8193449324369431, - "num_tokens": 2518888.0, - "step": 1050 - }, - { - "entropy": 0.6173626267910004, - "epoch": 2.651387213510253, - "grad_norm": 0.7958805561065674, - "learning_rate": 0.00012212878697538592, - "loss": 0.5649292755126953, - "mean_token_accuracy": 0.8172902607917786, - "num_tokens": 2639763.0, - "step": 1100 - }, - { - "entropy": 0.6264930808544159, - "epoch": 2.772014475271411, - "grad_norm": 0.7718372344970703, - "learning_rate": 0.00012056029815077323, - "loss": 0.5716455078125, - "mean_token_accuracy": 0.8170186239480972, - "num_tokens": 2755540.0, - "step": 1150 - }, - { - "entropy": 0.6203387409448624, - "epoch": 2.8926417370325694, - "grad_norm": 0.7914765477180481, - "learning_rate": 0.000118896046720439, - "loss": 0.5653961944580078, - "mean_token_accuracy": 0.8170280295610428, - "num_tokens": 2874118.0, - "step": 1200 - }, - { - "epoch": 3.0, - "eval_entropy": 0.6751241854737314, - "eval_mean_token_accuracy": 0.792226163858778, - "eval_not_syn_loss": 0.6902772784233093, - "eval_not_syn_runtime": 53.2557, - "eval_not_syn_samples_per_second": 26.664, - "eval_not_syn_steps_per_second": 3.342, - "eval_num_tokens": 2981418.0, - "step": 1245 - }, - { - "epoch": 3.0, - "eval_entropy": 0.6308753868837035, - "eval_mean_token_accuracy": 0.8131611759073278, - "eval_num_tokens": 2981418.0, - "eval_syn_loss": 0.6440441012382507, - "eval_syn_runtime": 54.7064, - "eval_syn_samples_per_second": 25.957, - "eval_syn_steps_per_second": 3.254, - "step": 1245 - }, - { - "entropy": 0.6077835421369533, - "epoch": 3.012062726176116, - "grad_norm": 1.124720573425293, - "learning_rate": 0.00011713897584164003, - "loss": 0.5557136917114258, - "mean_token_accuracy": 0.8203362726201915, - "num_tokens": 2994602.0, - "step": 1250 - }, - { - "entropy": 0.507642827630043, - "epoch": 3.132689987937274, - "grad_norm": 0.9039916396141052, - "learning_rate": 0.00011529219281884579, - "loss": 0.4467372512817383, - "mean_token_accuracy": 0.8464260965585708, - "num_tokens": 3121094.0, - "step": 1300 - }, - { - "entropy": 0.5137191081047058, - "epoch": 3.2533172496984317, - "grad_norm": 0.8580691814422607, - "learning_rate": 0.00011335896360860367, - "loss": 0.4538694763183594, - "mean_token_accuracy": 0.8448719161748887, - "num_tokens": 3248038.0, - "step": 1350 - }, - { - "entropy": 0.5293738862872124, - "epoch": 3.37394451145959, - "grad_norm": 1.0177273750305176, - "learning_rate": 0.00011134270704383492, - "loss": 0.4639348220825195, - "mean_token_accuracy": 0.8428975892066956, - "num_tokens": 3363487.0, - "step": 1400 - }, - { - "entropy": 0.5218872320652008, - "epoch": 3.494571773220748, - "grad_norm": 1.0530673265457153, - "learning_rate": 0.00010924698878777538, - "loss": 0.4584346389770508, - "mean_token_accuracy": 0.8432288640737533, - "num_tokens": 3480513.0, - "step": 1450 - }, - { - "entropy": 0.5320372316241264, - "epoch": 3.615199034981906, - "grad_norm": 1.0291194915771484, - "learning_rate": 0.0001070755150282529, - "loss": 0.4711898422241211, - "mean_token_accuracy": 0.842066935300827, - "num_tokens": 3597183.0, - "step": 1500 - }, - { - "entropy": 0.5315000694990158, - "epoch": 3.7358262967430638, - "grad_norm": 0.9788838624954224, - "learning_rate": 0.00010483212592345332, - "loss": 0.4729101943969727, - "mean_token_accuracy": 0.8404586082696914, - "num_tokens": 3714173.0, - "step": 1550 - }, - { - "entropy": 0.5331643190979958, - "epoch": 3.856453558504222, - "grad_norm": 1.1248903274536133, - "learning_rate": 0.00010252078881076532, - "loss": 0.47405517578125, - "mean_token_accuracy": 0.8397526168823242, - "num_tokens": 3832652.0, - "step": 1600 - }, - { - "entropy": 0.5343865439295769, - "epoch": 3.97708082026538, - "grad_norm": 0.9008272886276245, - "learning_rate": 0.00010014559119071452, - "loss": 0.47172779083251953, - "mean_token_accuracy": 0.8404951041936874, - "num_tokens": 3955167.0, - "step": 1650 - }, - { - "epoch": 4.0, - "eval_entropy": 0.5736920103598177, - "eval_mean_token_accuracy": 0.7981750027517255, - "eval_not_syn_loss": 0.707631528377533, - "eval_not_syn_runtime": 52.2858, - "eval_not_syn_samples_per_second": 27.158, - "eval_not_syn_steps_per_second": 3.404, - "eval_num_tokens": 3975224.0, - "step": 1660 - }, - { - "epoch": 4.0, - "eval_entropy": 0.5363499959198277, - "eval_mean_token_accuracy": 0.8063885539435269, - "eval_num_tokens": 3975224.0, - "eval_syn_loss": 0.6610181331634521, - "eval_syn_runtime": 54.2167, - "eval_syn_samples_per_second": 26.191, - "eval_syn_steps_per_second": 3.283, - "step": 1660 - }, - { - "entropy": 0.4512602482179199, - "epoch": 4.0965018094089265, - "grad_norm": 0.8079906105995178, - "learning_rate": 9.771073349839416e-05, - "loss": 0.3778452682495117, - "mean_token_accuracy": 0.8684436221315404, - "num_tokens": 4072361.0, - "step": 1700 - }, - { - "entropy": 0.4228873434662819, - "epoch": 4.217129071170085, - "grad_norm": 1.1206746101379395, - "learning_rate": 9.52205216751758e-05, - "loss": 0.34671127319335937, - "mean_token_accuracy": 0.8763950079679489, - "num_tokens": 4194978.0, - "step": 1750 - }, - { - "entropy": 0.42808119297027586, - "epoch": 4.337756332931242, - "grad_norm": 1.1268125772476196, - "learning_rate": 9.26793595538368e-05, - "loss": 0.3484600830078125, - "mean_token_accuracy": 0.8753323918581009, - "num_tokens": 4315548.0, - "step": 1800 - }, - { - "entropy": 0.42506767451763156, - "epoch": 4.458383594692401, - "grad_norm": 1.1051807403564453, - "learning_rate": 9.009174107057088e-05, - "loss": 0.35434780120849607, - "mean_token_accuracy": 0.8734411865472793, - "num_tokens": 4431469.0, - "step": 1850 - }, - { - "entropy": 0.43314908266067503, - "epoch": 4.579010856453559, - "grad_norm": 0.9782640337944031, - "learning_rate": 8.746224231765484e-05, - "loss": 0.36225860595703124, - "mean_token_accuracy": 0.8715140378475189, - "num_tokens": 4551803.0, - "step": 1900 - }, - { - "entropy": 0.42984022706747055, - "epoch": 4.699638118214716, - "grad_norm": 0.9772566556930542, - "learning_rate": 8.479551345082568e-05, - "loss": 0.356609001159668, - "mean_token_accuracy": 0.8717925781011582, - "num_tokens": 4674468.0, - "step": 1950 - }, - { - "entropy": 0.4474307084083557, - "epoch": 4.820265379975875, - "grad_norm": 0.9547265768051147, - "learning_rate": 8.209627046567966e-05, - "loss": 0.3654902267456055, - "mean_token_accuracy": 0.8724055212736129, - "num_tokens": 4788027.0, - "step": 2000 - }, - { - "entropy": 0.4377285006642342, - "epoch": 4.940892641737032, - "grad_norm": 1.0390652418136597, - "learning_rate": 7.936928685763635e-05, - "loss": 0.3645761871337891, - "mean_token_accuracy": 0.8707609891891479, - "num_tokens": 4907561.0, - "step": 2050 - }, - { - "epoch": 5.0, - "eval_entropy": 0.5173388388049737, - "eval_mean_token_accuracy": 0.7918405432379647, - "eval_not_syn_loss": 0.7532826662063599, - "eval_not_syn_runtime": 53.0766, - "eval_not_syn_samples_per_second": 26.754, - "eval_not_syn_steps_per_second": 3.354, - "eval_num_tokens": 4969030.0, - "step": 2075 - }, - { - "epoch": 5.0, - "eval_entropy": 0.4831916272975086, - "eval_mean_token_accuracy": 0.8204237202580056, - "eval_num_tokens": 4969030.0, - "eval_syn_loss": 0.6971890330314636, - "eval_syn_runtime": 54.8755, - "eval_syn_samples_per_second": 25.877, - "eval_syn_steps_per_second": 3.244, - "step": 2075 - }, - { - "entropy": 0.3804885711934831, - "epoch": 5.060313630880579, - "grad_norm": 0.8717736005783081, - "learning_rate": 7.661938518021687e-05, - "loss": 0.3034685707092285, - "mean_token_accuracy": 0.8910005471923135, - "num_tokens": 5026625.0, - "step": 2100 - }, - { - "entropy": 0.3174623078107834, - "epoch": 5.180940892641737, - "grad_norm": 0.9808429479598999, - "learning_rate": 7.385142851656478e-05, - "loss": 0.23829608917236328, - "mean_token_accuracy": 0.9131860065460206, - "num_tokens": 5141010.0, - "step": 2150 - }, - { - "entropy": 0.3187542732059956, - "epoch": 5.301568154402895, - "grad_norm": 1.3739993572235107, - "learning_rate": 7.107031187929228e-05, - "loss": 0.24024734497070313, - "mean_token_accuracy": 0.910272576212883, - "num_tokens": 5261240.0, - "step": 2200 - }, - { - "entropy": 0.31624249830842016, - "epoch": 5.422195416164053, - "grad_norm": 1.1618341207504272, - "learning_rate": 6.82809535538603e-05, - "loss": 0.24144052505493163, - "mean_token_accuracy": 0.9122701978683472, - "num_tokens": 5382160.0, - "step": 2250 - }, - { - "entropy": 0.31793304443359377, - "epoch": 5.542822677925211, - "grad_norm": 1.4960980415344238, - "learning_rate": 6.548828640080172e-05, - "loss": 0.24051647186279296, - "mean_token_accuracy": 0.9116496992111206, - "num_tokens": 5502002.0, - "step": 2300 - }, - { - "entropy": 0.3180307722091675, - "epoch": 5.66344993968637, - "grad_norm": 1.103051781654358, - "learning_rate": 6.269724913216913e-05, - "loss": 0.24294727325439452, - "mean_token_accuracy": 0.9102917027473449, - "num_tokens": 5629969.0, - "step": 2350 - }, - { - "entropy": 0.32020670652389527, - "epoch": 5.784077201447527, - "grad_norm": 1.0690243244171143, - "learning_rate": 5.991277757763442e-05, - "loss": 0.24328670501708985, - "mean_token_accuracy": 0.9101310813426972, - "num_tokens": 5750596.0, - "step": 2400 - }, - { - "entropy": 0.31846788004040716, - "epoch": 5.904704463208685, - "grad_norm": 1.203026294708252, - "learning_rate": 5.7139795955685784e-05, - "loss": 0.244984130859375, - "mean_token_accuracy": 0.9100374561548233, - "num_tokens": 5868110.0, - "step": 2450 - }, - { - "epoch": 6.0, - "eval_entropy": 0.45794740654109567, - "eval_mean_token_accuracy": 0.7891001142142864, - "eval_not_syn_loss": 0.8338615894317627, - "eval_not_syn_runtime": 53.3841, - "eval_not_syn_samples_per_second": 26.6, - "eval_not_syn_steps_per_second": 3.334, - "eval_num_tokens": 5962836.0, - "step": 2490 - }, - { - "epoch": 6.0, - "eval_entropy": 0.4288232589370749, - "eval_mean_token_accuracy": 0.8114104933953017, - "eval_num_tokens": 5962836.0, - "eval_syn_loss": 0.7791845798492432, - "eval_syn_runtime": 54.7877, - "eval_syn_samples_per_second": 25.918, - "eval_syn_steps_per_second": 3.249, - "step": 2490 - }, - { - "entropy": 0.2972855128423132, - "epoch": 6.024125452352232, - "grad_norm": 1.1748684644699097, - "learning_rate": 5.438320816535859e-05, - "loss": 0.22051204681396486, - "mean_token_accuracy": 0.9190925630656156, - "num_tokens": 5990107.0, - "step": 2500 - }, - { - "entropy": 0.2216818816959858, - "epoch": 6.144752714113389, - "grad_norm": 1.0899405479431152, - "learning_rate": 5.164788911390022e-05, - "loss": 0.1487915325164795, - "mean_token_accuracy": 0.9462692165374755, - "num_tokens": 6104230.0, - "step": 2550 - }, - { - "entropy": 0.22189782992005347, - "epoch": 6.265379975874548, - "grad_norm": 1.0967661142349243, - "learning_rate": 4.8938676095705914e-05, - "loss": 0.149844388961792, - "mean_token_accuracy": 0.9452104204893113, - "num_tokens": 6223548.0, - "step": 2600 - }, - { - "entropy": 0.2155836282670498, - "epoch": 6.386007237635706, - "grad_norm": 1.1443830728530884, - "learning_rate": 4.626036023777093e-05, - "loss": 0.1472412109375, - "mean_token_accuracy": 0.946697844862938, - "num_tokens": 6349102.0, - "step": 2650 - }, - { - "entropy": 0.22189507275819778, - "epoch": 6.5066344993968634, - "grad_norm": 1.2138874530792236, - "learning_rate": 4.3617678026788244e-05, - "loss": 0.15155397415161131, - "mean_token_accuracy": 0.9441326081752777, - "num_tokens": 6468966.0, - "step": 2700 - }, - { - "entropy": 0.22405965447425844, - "epoch": 6.627261761158022, - "grad_norm": 1.2444716691970825, - "learning_rate": 4.10153029328748e-05, - "loss": 0.15516298294067382, - "mean_token_accuracy": 0.9442053025960923, - "num_tokens": 6583296.0, - "step": 2750 - }, - { - "entropy": 0.2216674068570137, - "epoch": 6.74788902291918, - "grad_norm": 1.0056185722351074, - "learning_rate": 3.845783714474033e-05, - "loss": 0.15423223495483399, - "mean_token_accuracy": 0.9445771616697312, - "num_tokens": 6698696.0, - "step": 2800 - }, - { - "entropy": 0.21455578804016112, - "epoch": 6.868516284680338, - "grad_norm": 1.0748720169067383, - "learning_rate": 3.59498034309138e-05, - "loss": 0.14870482444763183, - "mean_token_accuracy": 0.9457976347208024, - "num_tokens": 6819711.0, - "step": 2850 - }, - { - "entropy": 0.21227611482143402, - "epoch": 6.989143546441496, - "grad_norm": 0.978204607963562, - "learning_rate": 3.349563714142155e-05, - "loss": 0.14802613258361816, - "mean_token_accuracy": 0.9466447854042053, - "num_tokens": 6944941.0, - "step": 2900 - }, - { - "epoch": 7.0, - "eval_entropy": 0.38817436741978933, - "eval_mean_token_accuracy": 0.7801904045463948, - "eval_not_syn_loss": 0.9901206493377686, - "eval_not_syn_runtime": 53.1137, - "eval_not_syn_samples_per_second": 26.735, - "eval_not_syn_steps_per_second": 3.351, - "eval_num_tokens": 6956642.0, - "step": 2905 - }, - { - "epoch": 7.0, - "eval_entropy": 0.3640736076938972, - "eval_mean_token_accuracy": 0.8158511425671953, - "eval_num_tokens": 6956642.0, - "eval_syn_loss": 0.8903024792671204, - "eval_syn_runtime": 54.3768, - "eval_syn_samples_per_second": 26.114, - "eval_syn_steps_per_second": 3.273, - "step": 2905 - }, - { - "entropy": 0.16986672166321012, - "epoch": 7.108564535585042, - "grad_norm": 0.7068538069725037, - "learning_rate": 3.1099678364060876e-05, - "loss": 0.10039780616760254, - "mean_token_accuracy": 0.964904587678235, - "num_tokens": 7064818.0, - "step": 2950 - }, - { - "entropy": 0.16062987893819808, - "epoch": 7.2291917973462, - "grad_norm": 0.963703453540802, - "learning_rate": 2.8766164249141098e-05, - "loss": 0.09831766128540038, - "mean_token_accuracy": 0.9666522133350373, - "num_tokens": 7183624.0, - "step": 3000 - }, - { - "entropy": 0.15993704989552499, - "epoch": 7.349819059107358, - "grad_norm": 0.8077407479286194, - "learning_rate": 2.6499221516264785e-05, - "loss": 0.09768331527709961, - "mean_token_accuracy": 0.9664828568696976, - "num_tokens": 7304122.0, - "step": 3050 - }, - { - "entropy": 0.16501803942024706, - "epoch": 7.470446320868517, - "grad_norm": 0.8302854299545288, - "learning_rate": 2.4302859156401512e-05, - "loss": 0.10022387504577637, - "mean_token_accuracy": 0.9645011806488037, - "num_tokens": 7422217.0, - "step": 3100 - }, - { - "entropy": 0.1541054170578718, - "epoch": 7.591073582629674, - "grad_norm": 0.8342278599739075, - "learning_rate": 2.218096134215895e-05, - "loss": 0.09620431900024413, - "mean_token_accuracy": 0.9666509574651718, - "num_tokens": 7547071.0, - "step": 3150 - }, - { - "entropy": 0.16526211962103843, - "epoch": 7.711700844390832, - "grad_norm": 0.9860979914665222, - "learning_rate": 2.0137280558790573e-05, - "loss": 0.10015854835510254, - "mean_token_accuracy": 0.9650658935308456, - "num_tokens": 7663184.0, - "step": 3200 - }, - { - "entropy": 0.1566712998598814, - "epoch": 7.832328106151991, - "grad_norm": 0.6179924011230469, - "learning_rate": 1.8175430968086255e-05, - "loss": 0.09698445320129395, - "mean_token_accuracy": 0.9661821335554123, - "num_tokens": 7785293.0, - "step": 3250 - }, - { - "entropy": 0.16459605664014818, - "epoch": 7.952955367913148, - "grad_norm": 0.7428495287895203, - "learning_rate": 1.6298882016882273e-05, - "loss": 0.09972883224487304, - "mean_token_accuracy": 0.9650641173124314, - "num_tokens": 7903616.0, - "step": 3300 - }, - { - "epoch": 8.0, - "eval_entropy": 0.339001152455137, - "eval_mean_token_accuracy": 0.7834485541568713, - "eval_not_syn_loss": 1.1380223035812378, - "eval_not_syn_runtime": 52.6766, - "eval_not_syn_samples_per_second": 26.957, - "eval_not_syn_steps_per_second": 3.379, - "eval_num_tokens": 7950448.0, - "step": 3320 - }, - { - "epoch": 8.0, - "eval_entropy": 0.3175283433513695, - "eval_mean_token_accuracy": 0.8104708164595487, - "eval_num_tokens": 7950448.0, - "eval_syn_loss": 1.0303504467010498, - "eval_syn_runtime": 54.8193, - "eval_syn_samples_per_second": 25.903, - "eval_syn_steps_per_second": 3.247, - "step": 3320 - }, - { - "entropy": 0.14740703307619, - "epoch": 8.072376357056696, - "grad_norm": 0.5945252776145935, - "learning_rate": 1.4510952301493426e-05, - "loss": 0.08425969123840332, - "mean_token_accuracy": 0.9707530875398656, - "num_tokens": 8020867.0, - "step": 3350 - }, - { - "entropy": 0.13212954834103585, - "epoch": 8.193003618817853, - "grad_norm": 0.6547225713729858, - "learning_rate": 1.2814803698917796e-05, - "loss": 0.07411238670349121, - "mean_token_accuracy": 0.9743285477161407, - "num_tokens": 8143032.0, - "step": 3400 - }, - { - "entropy": 0.13668791383504866, - "epoch": 8.31363088057901, - "grad_norm": 0.49490705132484436, - "learning_rate": 1.1213435775192938e-05, - "loss": 0.07555238246917724, - "mean_token_accuracy": 0.9738145649433136, - "num_tokens": 8262875.0, - "step": 3450 - }, - { - "entropy": 0.13185277588665487, - "epoch": 8.43425814234017, - "grad_norm": 0.43988436460494995, - "learning_rate": 9.709680480792023e-06, - "loss": 0.075065279006958, - "mean_token_accuracy": 0.9734203344583512, - "num_tokens": 8383147.0, - "step": 3500 - }, - { - "entropy": 0.1305071434378624, - "epoch": 8.554885404101327, - "grad_norm": 0.46255460381507874, - "learning_rate": 8.306197142440928e-06, - "loss": 0.07382431983947754, - "mean_token_accuracy": 0.9748922157287597, - "num_tokens": 8502743.0, - "step": 3550 - }, - { - "entropy": 0.1334294395148754, - "epoch": 8.675512665862485, - "grad_norm": 0.6507574915885925, - "learning_rate": 7.0054677602129975e-06, - "loss": 0.0754416275024414, - "mean_token_accuracy": 0.9741936498880386, - "num_tokens": 8621550.0, - "step": 3600 - }, - { - "entropy": 0.13020497389137745, - "epoch": 8.796139927623642, - "grad_norm": 0.523806631565094, - "learning_rate": 5.809792618218385e-06, - "loss": 0.07472522258758545, - "mean_token_accuracy": 0.9740422391891479, - "num_tokens": 8742590.0, - "step": 3650 - }, - { - "entropy": 0.13258147858083247, - "epoch": 8.916767189384801, - "grad_norm": 0.6534594893455505, - "learning_rate": 4.721286216650351e-06, - "loss": 0.07624986171722412, - "mean_token_accuracy": 0.9732474505901336, - "num_tokens": 8862250.0, - "step": 3700 - }, - { - "epoch": 9.0, - "eval_entropy": 0.3192606071910162, - "eval_mean_token_accuracy": 0.7872959100798275, - "eval_not_syn_loss": 1.2220250368118286, - "eval_not_syn_runtime": 53.1436, - "eval_not_syn_samples_per_second": 26.72, - "eval_not_syn_steps_per_second": 3.349, - "eval_num_tokens": 8944254.0, - "step": 3735 - }, - { - "epoch": 9.0, - "eval_entropy": 0.298558424028118, - "eval_mean_token_accuracy": 0.8055475206857317, - "eval_num_tokens": 8944254.0, - "eval_syn_loss": 1.140589714050293, - "eval_syn_runtime": 54.7832, - "eval_syn_samples_per_second": 25.92, - "eval_syn_steps_per_second": 3.249, - "step": 3735 - }, - { - "entropy": 0.12874245658667402, - "epoch": 9.036188178528347, - "grad_norm": 0.2500600814819336, - "learning_rate": 3.7418735323822404e-06, - "loss": 0.07274995326995849, - "mean_token_accuracy": 0.9751881041912117, - "num_tokens": 8981377.0, - "step": 3750 - }, - { - "entropy": 0.12800926234573126, - "epoch": 9.156815440289506, - "grad_norm": 0.39855095744132996, - "learning_rate": 2.873286614728655e-06, - "loss": 0.06881739616394043, - "mean_token_accuracy": 0.9762847191095352, - "num_tokens": 9099674.0, - "step": 3800 - }, - { - "entropy": 0.12072551108896733, - "epoch": 9.277442702050664, - "grad_norm": 0.3722686171531677, - "learning_rate": 2.117061522390382e-06, - "loss": 0.06577431678771972, - "mean_token_accuracy": 0.9772092348337174, - "num_tokens": 9222172.0, - "step": 3850 - }, - { - "entropy": 0.12452757865190506, - "epoch": 9.398069963811821, - "grad_norm": 0.4538884460926056, - "learning_rate": 1.474535607000592e-06, - "loss": 0.06838587284088135, - "mean_token_accuracy": 0.9763706302642823, - "num_tokens": 9341003.0, - "step": 3900 - }, - { - "entropy": 0.12404791057109833, - "epoch": 9.51869722557298, - "grad_norm": 0.614382266998291, - "learning_rate": 9.468451480757799e-07, - "loss": 0.06680603504180908, - "mean_token_accuracy": 0.9762758719921112, - "num_tokens": 9461190.0, - "step": 3950 - }, - { - "entropy": 0.12405681416392327, - "epoch": 9.639324487334138, - "grad_norm": 0.4734494388103485, - "learning_rate": 5.349233435542355e-07, - "loss": 0.06736649513244629, - "mean_token_accuracy": 0.9763290709257126, - "num_tokens": 9577775.0, - "step": 4000 - }, - { - "entropy": 0.12378192521631717, - "epoch": 9.759951749095295, - "grad_norm": 0.5742450952529907, - "learning_rate": 2.394986594754693e-07, - "loss": 0.06865021228790283, - "mean_token_accuracy": 0.9762201535701752, - "num_tokens": 9696918.0, - "step": 4050 - }, - { - "entropy": 0.12134758442640305, - "epoch": 9.880579010856454, - "grad_norm": 0.37256020307540894, - "learning_rate": 6.109354171933916e-08, - "loss": 0.06521551609039307, - "mean_token_accuracy": 0.9767780256271362, - "num_tokens": 9821164.0, - "step": 4100 - }, - { - "entropy": 0.12435369434380772, - "epoch": 10.0, - "grad_norm": 0.4080880880355835, - "learning_rate": 2.349208286529078e-11, - "loss": 0.06751453399658203, - "mean_token_accuracy": 0.9757449506509184, - "num_tokens": 9938060.0, - "step": 4150 - }, - { - "epoch": 10.0, - "eval_entropy": 0.3112048939540145, - "eval_mean_token_accuracy": 0.7877589882759566, - "eval_not_syn_loss": 1.2889702320098877, - "eval_not_syn_runtime": 53.09, - "eval_not_syn_samples_per_second": 26.747, - "eval_not_syn_steps_per_second": 3.353, - "eval_num_tokens": 9938060.0, - "step": 4150 - }, - { - "epoch": 10.0, - "eval_entropy": 0.29035538543811007, - "eval_mean_token_accuracy": 0.8043871950567438, - "eval_num_tokens": 9938060.0, - "eval_syn_loss": 1.207026720046997, - "eval_syn_runtime": 54.4951, - "eval_syn_samples_per_second": 26.057, - "eval_syn_steps_per_second": 3.266, - "step": 4150 - } - ], - "logging_steps": 50, - "max_steps": 4150, - "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": true - }, - "attributes": {} - } - }, - "total_flos": 2.1117503009615808e+17, - "train_batch_size": 8, - "trial_name": null, - "trial_params": null -} diff --git a/substitutivity_original_Estonian/Qwen3.5-2B-Base_substitutivity_splits_original_features_train_substitutivity_splits_original_features_test1/checkpoint-4150/training_args.bin b/substitutivity_original_Estonian/Qwen3.5-2B-Base_substitutivity_splits_original_features_train_substitutivity_splits_original_features_test1/checkpoint-4150/training_args.bin deleted file mode 100644 index 48901efb30506830185bfaafecbe236e7095a006..0000000000000000000000000000000000000000 --- a/substitutivity_original_Estonian/Qwen3.5-2B-Base_substitutivity_splits_original_features_train_substitutivity_splits_original_features_test1/checkpoint-4150/training_args.bin +++ /dev/null @@ -1,3 +0,0 @@ -version https://git-lfs.github.com/spec/v1 -oid sha256:770fc15ebc27a54d780af597d32c57ca14b3b528f14788d142bdfce09b555bef -size 6033 diff --git a/substitutivity_original_Estonian/Qwen3.5-2B-Base_substitutivity_splits_original_features_train_substitutivity_splits_original_features_test1/checkpoint-830/README.md b/substitutivity_original_Estonian/Qwen3.5-2B-Base_substitutivity_splits_original_features_train_substitutivity_splits_original_features_test1/checkpoint-830/README.md deleted file mode 100644 index ece204d005f7bf161d0e448efdd56fd297b7c8a2..0000000000000000000000000000000000000000 --- a/substitutivity_original_Estonian/Qwen3.5-2B-Base_substitutivity_splits_original_features_train_substitutivity_splits_original_features_test1/checkpoint-830/README.md +++ /dev/null @@ -1,209 +0,0 @@ ---- -base_model: Qwen/Qwen3.5-2B-Base -library_name: peft -pipeline_tag: text-generation -tags: -- base_model:adapter:Qwen/Qwen3.5-2B-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/substitutivity_original_Estonian/Qwen3.5-2B-Base_substitutivity_splits_original_features_train_substitutivity_splits_original_features_test1/checkpoint-830/adapter_config.json b/substitutivity_original_Estonian/Qwen3.5-2B-Base_substitutivity_splits_original_features_train_substitutivity_splits_original_features_test1/checkpoint-830/adapter_config.json deleted file mode 100644 index a64145ce23adcbdbe246679087a13f4cb63ff664..0000000000000000000000000000000000000000 --- a/substitutivity_original_Estonian/Qwen3.5-2B-Base_substitutivity_splits_original_features_train_substitutivity_splits_original_features_test1/checkpoint-830/adapter_config.json +++ /dev/null @@ -1,46 +0,0 @@ -{ - "alora_invocation_tokens": null, - "alpha_pattern": {}, - "arrow_config": null, - "auto_mapping": null, - "base_model_name_or_path": "Qwen/Qwen3.5-2B-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": 256, - "lora_bias": false, - "lora_dropout": 0.027050353915701653, - "megatron_config": null, - "megatron_core": "megatron.core", - "modules_to_save": null, - "peft_type": "LORA", - "peft_version": "0.18.1", - "qalora_group_size": 16, - "r": 128, - "rank_pattern": {}, - "revision": null, - "target_modules": [ - "o_proj", - "q_proj", - "k_proj", - "down_proj", - "up_proj", - "v_proj", - "gate_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/substitutivity_original_Estonian/Qwen3.5-2B-Base_substitutivity_splits_original_features_train_substitutivity_splits_original_features_test1/checkpoint-830/adapter_model.safetensors b/substitutivity_original_Estonian/Qwen3.5-2B-Base_substitutivity_splits_original_features_train_substitutivity_splits_original_features_test1/checkpoint-830/adapter_model.safetensors deleted file mode 100644 index 0bd535abd5ebc95c9475ba30e093909b452dce5e..0000000000000000000000000000000000000000 --- a/substitutivity_original_Estonian/Qwen3.5-2B-Base_substitutivity_splits_original_features_train_substitutivity_splits_original_features_test1/checkpoint-830/adapter_model.safetensors +++ /dev/null @@ -1,3 +0,0 @@ -version https://git-lfs.github.com/spec/v1 -oid sha256:9785e362703be77ae42b63401590c3da9b7f01fad3d09b86fd95fe21f880388b -size 349201608 diff --git a/substitutivity_original_Estonian/Qwen3.5-2B-Base_substitutivity_splits_original_features_train_substitutivity_splits_original_features_test1/checkpoint-830/chat_template.jinja b/substitutivity_original_Estonian/Qwen3.5-2B-Base_substitutivity_splits_original_features_train_substitutivity_splits_original_features_test1/checkpoint-830/chat_template.jinja deleted file mode 100644 index 0ef09f214eaa6d9bca297988afc1454b5827b2c7..0000000000000000000000000000000000000000 --- a/substitutivity_original_Estonian/Qwen3.5-2B-Base_substitutivity_splits_original_features_train_substitutivity_splits_original_features_test1/checkpoint-830/chat_template.jinja +++ /dev/null @@ -1,154 +0,0 @@ -{%- set image_count = namespace(value=0) %} -{%- set video_count = namespace(value=0) %} -{%- macro render_content(content, do_vision_count, is_system_content=false) %} - {%- if content is string %} - {{- content }} - {%- elif content is iterable and content is not mapping %} - {%- for item in content %} - {%- if 'image' in item or 'image_url' in item or item.type == 'image' %} - {%- if is_system_content %} - {{- raise_exception('System message cannot contain images.') }} - {%- endif %} - {%- if do_vision_count %} - {%- set image_count.value = image_count.value + 1 %} - {%- endif %} - {%- if add_vision_id %} - {{- 'Picture ' ~ image_count.value ~ ': ' }} - {%- endif %} - {{- '<|vision_start|><|image_pad|><|vision_end|>' }} - {%- elif 'video' in item or item.type == 'video' %} - {%- if is_system_content %} - {{- raise_exception('System message cannot contain videos.') }} - {%- endif %} - {%- if do_vision_count %} - {%- set video_count.value = video_count.value + 1 %} - {%- endif %} - {%- if add_vision_id %} - {{- 'Video ' ~ video_count.value ~ ': ' }} - {%- endif %} - {{- '<|vision_start|><|video_pad|><|vision_end|>' }} - {%- elif 'text' in item %} - {{- item.text }} - {%- else %} - {{- raise_exception('Unexpected item type in content.') }} - {%- endif %} - {%- endfor %} - {%- elif content is none or content is undefined %} - {{- '' }} - {%- else %} - {{- raise_exception('Unexpected content type.') }} - {%- endif %} -{%- endmacro %} -{%- if not messages %} - {{- raise_exception('No messages provided.') }} -{%- endif %} -{%- if tools and tools is iterable and tools is not mapping %} - {{- '<|im_start|>system\n' }} - {{- "# Tools\n\nYou have access to the following functions:\n\n" }} - {%- for tool in tools %} - {{- "\n" }} - {{- tool | tojson }} - {%- endfor %} - {{- "\n" }} - {{- '\n\nIf you choose to call a function ONLY reply in the following format with NO suffix:\n\n\n\n\nvalue_1\n\n\nThis is the value for the second parameter\nthat can span\nmultiple lines\n\n\n\n\n\nReminder:\n- Function calls MUST follow the specified format: an inner block must be nested within XML tags\n- Required parameters MUST be specified\n- You may provide optional reasoning for your function call in natural language BEFORE the function call, but NOT after\n- If there is no function call available, answer the question like normal with your current knowledge and do not tell the user about function calls\n' }} - {%- if messages[0].role == 'system' %} - {%- set content = render_content(messages[0].content, false, true)|trim %} - {%- if content %} - {{- '\n\n' + content }} - {%- endif %} - {%- endif %} - {{- '<|im_end|>\n' }} -{%- else %} - {%- if messages[0].role == 'system' %} - {%- set content = render_content(messages[0].content, false, true)|trim %} - {{- '<|im_start|>system\n' + 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" %} - {%- set content = render_content(message.content, false)|trim %} - {%- if not(content.startswith('') and content.endswith('')) %} - {%- set ns.multi_step_tool = false %} - {%- set ns.last_query_index = index %} - {%- endif %} - {%- endif %} -{%- endfor %} -{%- if ns.multi_step_tool %} - {{- raise_exception('No user query found in messages.') }} -{%- endif %} -{%- for message in messages %} - {%- set content = render_content(message.content, true)|trim %} - {%- if message.role == "system" %} - {%- if not loop.first %} - {{- raise_exception('System message must be at the beginning.') }} - {%- endif %} - {%- elif message.role == "user" %} - {{- '<|im_start|>' + message.role + '\n' + content + '<|im_end|>' + '\n' }} - {%- elif message.role == "assistant" %} - {%- set reasoning_content = '' %} - {%- if message.reasoning_content is string %} - {%- set reasoning_content = message.reasoning_content %} - {%- else %} - {%- if '' in content %} - {%- set reasoning_content = content.split('')[0].rstrip('\n').split('')[-1].lstrip('\n') %} - {%- set content = content.split('')[-1].lstrip('\n') %} - {%- endif %} - {%- endif %} - {%- set reasoning_content = reasoning_content|trim %} - {%- if loop.index0 > ns.last_query_index %} - {{- '<|im_start|>' + message.role + '\n\n' + reasoning_content + '\n\n\n' + content }} - {%- else %} - {{- '<|im_start|>' + message.role + '\n' + content }} - {%- endif %} - {%- if message.tool_calls and message.tool_calls is iterable and message.tool_calls is not mapping %} - {%- for tool_call in message.tool_calls %} - {%- if tool_call.function is defined %} - {%- set tool_call = tool_call.function %} - {%- endif %} - {%- if loop.first %} - {%- if content|trim %} - {{- '\n\n\n\n' }} - {%- else %} - {{- '\n\n' }} - {%- endif %} - {%- else %} - {{- '\n\n\n' }} - {%- endif %} - {%- if tool_call.arguments is defined %} - {%- for args_name, args_value in tool_call.arguments|items %} - {{- '\n' }} - {%- set args_value = args_value | tojson | safe if args_value is mapping or (args_value is sequence and args_value is not string) else args_value | string %} - {{- args_value }} - {{- '\n\n' }} - {%- endfor %} - {%- endif %} - {{- '\n' }} - {%- endfor %} - {%- endif %} - {{- '<|im_end|>\n' }} - {%- elif message.role == "tool" %} - {%- if loop.previtem and loop.previtem.role != "tool" %} - {{- '<|im_start|>user' }} - {%- endif %} - {{- '\n\n' }} - {{- content }} - {{- '\n' }} - {%- if not loop.last and loop.nextitem.role != "tool" %} - {{- '<|im_end|>\n' }} - {%- elif loop.last %} - {{- '<|im_end|>\n' }} - {%- endif %} - {%- else %} - {{- raise_exception('Unexpected message role.') }} - {%- endif %} -{%- endfor %} -{%- if add_generation_prompt %} - {{- '<|im_start|>assistant\n' }} - {%- if enable_thinking is defined and enable_thinking is true %} - {{- '\n' }} - {%- else %} - {{- '\n\n\n\n' }} - {%- endif %} -{%- endif %} \ No newline at end of file diff --git a/substitutivity_original_Estonian/Qwen3.5-2B-Base_substitutivity_splits_original_features_train_substitutivity_splits_original_features_test1/checkpoint-830/tokenizer.json b/substitutivity_original_Estonian/Qwen3.5-2B-Base_substitutivity_splits_original_features_train_substitutivity_splits_original_features_test1/checkpoint-830/tokenizer.json deleted file mode 100644 index 67741b04f23bfdb46501f748ce27865ec82eccfb..0000000000000000000000000000000000000000 --- a/substitutivity_original_Estonian/Qwen3.5-2B-Base_substitutivity_splits_original_features_train_substitutivity_splits_original_features_test1/checkpoint-830/tokenizer.json +++ /dev/null @@ -1,3 +0,0 @@ -version https://git-lfs.github.com/spec/v1 -oid sha256:87a7830d63fcf43bf241c3c5242e96e62dd3fdc29224ca26fed8ea333db72de4 -size 19989343 diff --git a/substitutivity_original_Estonian/Qwen3.5-2B-Base_substitutivity_splits_original_features_train_substitutivity_splits_original_features_test1/checkpoint-830/tokenizer_config.json b/substitutivity_original_Estonian/Qwen3.5-2B-Base_substitutivity_splits_original_features_train_substitutivity_splits_original_features_test1/checkpoint-830/tokenizer_config.json deleted file mode 100644 index b4a37b2a6fd3ab3317cd7bac72855be1a843b2bb..0000000000000000000000000000000000000000 --- a/substitutivity_original_Estonian/Qwen3.5-2B-Base_substitutivity_splits_original_features_train_substitutivity_splits_original_features_test1/checkpoint-830/tokenizer_config.json +++ /dev/null @@ -1,31 +0,0 @@ -{ - "add_prefix_space": false, - "audio_bos_token": "<|audio_start|>", - "audio_eos_token": "<|audio_end|>", - "audio_token": "<|audio_pad|>", - "backend": "tokenizers", - "bos_token": null, - "clean_up_tokenization_spaces": false, - "eos_token": "<|endoftext|>", - "errors": "replace", - "image_token": "<|image_pad|>", - "is_local": false, - "model_max_length": 262144, - "model_specific_special_tokens": { - "audio_bos_token": "<|audio_start|>", - "audio_eos_token": "<|audio_end|>", - "audio_token": "<|audio_pad|>", - "image_token": "<|image_pad|>", - "video_token": "<|video_pad|>", - "vision_bos_token": "<|vision_start|>", - "vision_eos_token": "<|vision_end|>" - }, - "pad_token": "<|endoftext|>", - "pretokenize_regex": "(?i:'s|'t|'re|'ve|'m|'ll|'d)|[^\\r\\n\\p{L}\\p{N}]?[\\p{L}\\p{M}]+|\\p{N}| ?[^\\s\\p{L}\\p{M}\\p{N}]+[\\r\\n]*|\\s*[\\r\\n]+|\\s+(?!\\S)|\\s+", - "split_special_tokens": false, - "tokenizer_class": "TokenizersBackend", - "unk_token": null, - "video_token": "<|video_pad|>", - "vision_bos_token": "<|vision_start|>", - "vision_eos_token": "<|vision_end|>" -} diff --git a/substitutivity_original_Estonian/Qwen3.5-2B-Base_substitutivity_splits_original_features_train_substitutivity_splits_original_features_test1/checkpoint-830/trainer_state.json b/substitutivity_original_Estonian/Qwen3.5-2B-Base_substitutivity_splits_original_features_train_substitutivity_splits_original_features_test1/checkpoint-830/trainer_state.json deleted file mode 100644 index e8b59c8656e9c5c923ab4d49a49f2b690c22d91f..0000000000000000000000000000000000000000 --- a/substitutivity_original_Estonian/Qwen3.5-2B-Base_substitutivity_splits_original_features_train_substitutivity_splits_original_features_test1/checkpoint-830/trainer_state.json +++ /dev/null @@ -1,238 +0,0 @@ -{ - "best_global_step": null, - "best_metric": null, - "best_model_checkpoint": null, - "epoch": 2.0, - "eval_steps": 500, - "global_step": 830, - "is_hyper_param_search": false, - "is_local_process_zero": true, - "is_world_process_zero": true, - "log_history": [ - { - "entropy": 2.0473356211185454, - "epoch": 0.12062726176115803, - "grad_norm": 3.0554118156433105, - "learning_rate": 1.5682344549790923e-05, - "loss": 1.910550079345703, - "mean_token_accuracy": 0.6053373458981514, - "num_tokens": 121428.0, - "step": 50 - }, - { - "entropy": 1.0371007251739501, - "epoch": 0.24125452352231605, - "grad_norm": 2.109868049621582, - "learning_rate": 3.168473694753677e-05, - "loss": 0.9585584259033203, - "mean_token_accuracy": 0.7352996730804443, - "num_tokens": 243966.0, - "step": 100 - }, - { - "entropy": 0.9105490958690643, - "epoch": 0.3618817852834741, - "grad_norm": 1.7675701379776, - "learning_rate": 4.7687129345282605e-05, - "loss": 0.8393524932861328, - "mean_token_accuracy": 0.7588736474514007, - "num_tokens": 366649.0, - "step": 150 - }, - { - "entropy": 0.8660378754138947, - "epoch": 0.4825090470446321, - "grad_norm": 1.6502987146377563, - "learning_rate": 6.368952174302845e-05, - "loss": 0.8005844116210937, - "mean_token_accuracy": 0.7663125538825989, - "num_tokens": 488367.0, - "step": 200 - }, - { - "entropy": 0.8439123743772506, - "epoch": 0.6031363088057901, - "grad_norm": 1.538365364074707, - "learning_rate": 7.969191414077429e-05, - "loss": 0.7773619842529297, - "mean_token_accuracy": 0.7718527519702911, - "num_tokens": 603131.0, - "step": 250 - }, - { - "entropy": 0.8223806774616241, - "epoch": 0.7237635705669482, - "grad_norm": 1.1559664011001587, - "learning_rate": 9.569430653852014e-05, - "loss": 0.7587767791748047, - "mean_token_accuracy": 0.7767325991392136, - "num_tokens": 726412.0, - "step": 300 - }, - { - "entropy": 0.8256935775279999, - "epoch": 0.8443908323281062, - "grad_norm": 1.2322715520858765, - "learning_rate": 0.00011169669893626597, - "loss": 0.769886245727539, - "mean_token_accuracy": 0.773840668797493, - "num_tokens": 846516.0, - "step": 350 - }, - { - "entropy": 0.8065843790769577, - "epoch": 0.9650180940892642, - "grad_norm": 0.9908286333084106, - "learning_rate": 0.00012769909133401183, - "loss": 0.7451084899902344, - "mean_token_accuracy": 0.7794792425632476, - "num_tokens": 958978.0, - "step": 400 - }, - { - "epoch": 1.0, - "eval_entropy": 0.791056221790528, - "eval_mean_token_accuracy": 0.7507295300451557, - "eval_not_syn_loss": 0.7860428094863892, - "eval_not_syn_runtime": 53.5431, - "eval_not_syn_samples_per_second": 26.521, - "eval_not_syn_steps_per_second": 3.324, - "eval_num_tokens": 993806.0, - "step": 415 - }, - { - "epoch": 1.0, - "eval_entropy": 0.7424047408478983, - "eval_mean_token_accuracy": 0.8023804037758474, - "eval_num_tokens": 993806.0, - "eval_syn_loss": 0.7273509502410889, - "eval_syn_runtime": 54.8679, - "eval_syn_samples_per_second": 25.88, - "eval_syn_steps_per_second": 3.244, - "step": 415 - }, - { - "entropy": 0.7607260761838971, - "epoch": 1.0844390832328106, - "grad_norm": 1.0653696060180664, - "learning_rate": 0.0001327927019027069, - "loss": 0.7034613037109375, - "mean_token_accuracy": 0.7869195998317063, - "num_tokens": 1080301.0, - "step": 450 - }, - { - "entropy": 0.7370894801616669, - "epoch": 1.2050663449939687, - "grad_norm": 1.763787865638733, - "learning_rate": 0.00013265416569994628, - "loss": 0.6816973114013671, - "mean_token_accuracy": 0.7914507347345352, - "num_tokens": 1201215.0, - "step": 500 - }, - { - "entropy": 0.742829909324646, - "epoch": 1.3256936067551266, - "grad_norm": 0.97663414478302, - "learning_rate": 0.00013239847940553975, - "loss": 0.6867338562011719, - "mean_token_accuracy": 0.790328289270401, - "num_tokens": 1322895.0, - "step": 550 - }, - { - "entropy": 0.7312483406066894, - "epoch": 1.4463208685162847, - "grad_norm": 1.0874731540679932, - "learning_rate": 0.00013202609518972978, - "loss": 0.676800537109375, - "mean_token_accuracy": 0.7921512228250503, - "num_tokens": 1436749.0, - "step": 600 - }, - { - "entropy": 0.7178869074583054, - "epoch": 1.5669481302774426, - "grad_norm": 1.064300298690796, - "learning_rate": 0.0001315376715980265, - "loss": 0.6635584259033203, - "mean_token_accuracy": 0.7944776183366775, - "num_tokens": 1555400.0, - "step": 650 - }, - { - "entropy": 0.7198166918754577, - "epoch": 1.6875753920386007, - "grad_norm": 0.7779315114021301, - "learning_rate": 0.00013093407238659802, - "loss": 0.6660567474365234, - "mean_token_accuracy": 0.7968300533294678, - "num_tokens": 1676892.0, - "step": 700 - }, - { - "entropy": 0.7170061504840851, - "epoch": 1.8082026537997589, - "grad_norm": 1.089854121208191, - "learning_rate": 0.00013021636499475496, - "loss": 0.6632771301269531, - "mean_token_accuracy": 0.7963915091753005, - "num_tokens": 1802689.0, - "step": 750 - }, - { - "entropy": 0.6953610718250275, - "epoch": 1.9288299155609168, - "grad_norm": 0.7996730208396912, - "learning_rate": 0.0001293858186572295, - "loss": 0.6421419525146485, - "mean_token_accuracy": 0.8002577543258667, - "num_tokens": 1920804.0, - "step": 800 - }, - { - "epoch": 2.0, - "eval_entropy": 0.7369416523515508, - "eval_mean_token_accuracy": 0.7872022366925572, - "eval_not_syn_loss": 0.713637113571167, - "eval_not_syn_runtime": 53.2265, - "eval_not_syn_samples_per_second": 26.678, - "eval_not_syn_steps_per_second": 3.344, - "eval_num_tokens": 1987612.0, - "step": 830 - }, - { - "epoch": 2.0, - "eval_entropy": 0.6877032756135705, - "eval_mean_token_accuracy": 0.7997466372639945, - "eval_num_tokens": 1987612.0, - "eval_syn_loss": 0.6614266633987427, - "eval_syn_runtime": 54.5178, - "eval_syn_samples_per_second": 26.047, - "eval_syn_steps_per_second": 3.265, - "step": 830 - } - ], - "logging_steps": 50, - "max_steps": 4150, - "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": 4.229782001727744e+16, - "train_batch_size": 8, - "trial_name": null, - "trial_params": null -} diff --git a/substitutivity_original_Estonian/Qwen3.5-2B-Base_substitutivity_splits_original_features_train_substitutivity_splits_original_features_test1/checkpoint-830/training_args.bin b/substitutivity_original_Estonian/Qwen3.5-2B-Base_substitutivity_splits_original_features_train_substitutivity_splits_original_features_test1/checkpoint-830/training_args.bin deleted file mode 100644 index 48901efb30506830185bfaafecbe236e7095a006..0000000000000000000000000000000000000000 --- a/substitutivity_original_Estonian/Qwen3.5-2B-Base_substitutivity_splits_original_features_train_substitutivity_splits_original_features_test1/checkpoint-830/training_args.bin +++ /dev/null @@ -1,3 +0,0 @@ -version https://git-lfs.github.com/spec/v1 -oid sha256:770fc15ebc27a54d780af597d32c57ca14b3b528f14788d142bdfce09b555bef -size 6033 diff --git a/substitutivity_original_Estonian/Qwen3.5-2B-Base_substitutivity_splits_original_features_train_substitutivity_splits_original_features_test2/README.md b/substitutivity_original_Estonian/Qwen3.5-2B-Base_substitutivity_splits_original_features_train_substitutivity_splits_original_features_test2/README.md deleted file mode 100644 index f9cf3d4247cecf81035b9c45e802c4ecd3ef91d5..0000000000000000000000000000000000000000 --- a/substitutivity_original_Estonian/Qwen3.5-2B-Base_substitutivity_splits_original_features_train_substitutivity_splits_original_features_test2/README.md +++ /dev/null @@ -1,58 +0,0 @@ ---- -base_model: Qwen/Qwen3.5-2B-Base -library_name: transformers -model_name: Qwen3.5-2B-Base_substitutivity_splits_original_features_train_substitutivity_splits_original_features_test2 -tags: -- generated_from_trainer -- sft -- trl -licence: license ---- - -# Model Card for Qwen3.5-2B-Base_substitutivity_splits_original_features_train_substitutivity_splits_original_features_test2 - -This model is a fine-tuned version of [Qwen/Qwen3.5-2B-Base](https://huggingface.co/Qwen/Qwen3.5-2B-Base). -It has been trained using [TRL](https://github.com/huggingface/trl). - -## Quick start - -```python -from transformers import pipeline - -question = "If you had a time machine, but could only go to the past or the future once and never return, which would you choose and why?" -generator = pipeline("text-generation", model="None", device="cuda") -output = generator([{"role": "user", "content": question}], max_new_tokens=128, return_full_text=False)[0] -print(output["generated_text"]) -``` - -## Training procedure - -[Visualize in Weights & Biases](https://wandb.ai/katriin-kukk/Cross_lingual_morphological_generalization/runs/38ls9xu6) - - - -This model was trained with SFT. - -### Framework versions - -- TRL: 0.29.0 -- Transformers: 5.5.4 -- Pytorch: 2.10.0 -- Datasets: 4.6.1 -- Tokenizers: 0.22.2 - -## Citations - - - -Cite TRL as: - -```bibtex -@software{vonwerra2020trl, - title = {{TRL: Transformers Reinforcement Learning}}, - author = {von Werra, Leandro and Belkada, Younes and Tunstall, Lewis and Beeching, Edward and Thrush, Tristan and Lambert, Nathan and Huang, Shengyi and Rasul, Kashif and Gallouédec, Quentin}, - license = {Apache-2.0}, - url = {https://github.com/huggingface/trl}, - year = {2020} -} -``` \ No newline at end of file diff --git a/substitutivity_original_Estonian/Qwen3.5-2B-Base_substitutivity_splits_original_features_train_substitutivity_splits_original_features_test2/checkpoint-1245/README.md b/substitutivity_original_Estonian/Qwen3.5-2B-Base_substitutivity_splits_original_features_train_substitutivity_splits_original_features_test2/checkpoint-1245/README.md deleted file mode 100644 index ece204d005f7bf161d0e448efdd56fd297b7c8a2..0000000000000000000000000000000000000000 --- a/substitutivity_original_Estonian/Qwen3.5-2B-Base_substitutivity_splits_original_features_train_substitutivity_splits_original_features_test2/checkpoint-1245/README.md +++ /dev/null @@ -1,209 +0,0 @@ ---- -base_model: Qwen/Qwen3.5-2B-Base -library_name: peft -pipeline_tag: text-generation -tags: -- base_model:adapter:Qwen/Qwen3.5-2B-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/substitutivity_original_Estonian/Qwen3.5-2B-Base_substitutivity_splits_original_features_train_substitutivity_splits_original_features_test2/checkpoint-1245/adapter_config.json b/substitutivity_original_Estonian/Qwen3.5-2B-Base_substitutivity_splits_original_features_train_substitutivity_splits_original_features_test2/checkpoint-1245/adapter_config.json deleted file mode 100644 index 6a095bb3bfe249ce8e15df6ae999e88fa1e17162..0000000000000000000000000000000000000000 --- a/substitutivity_original_Estonian/Qwen3.5-2B-Base_substitutivity_splits_original_features_train_substitutivity_splits_original_features_test2/checkpoint-1245/adapter_config.json +++ /dev/null @@ -1,46 +0,0 @@ -{ - "alora_invocation_tokens": null, - "alpha_pattern": {}, - "arrow_config": null, - "auto_mapping": null, - "base_model_name_or_path": "Qwen/Qwen3.5-2B-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": 0.05731002492184622, - "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": [ - "o_proj", - "q_proj", - "k_proj", - "down_proj", - "up_proj", - "v_proj", - "gate_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/substitutivity_original_Estonian/Qwen3.5-2B-Base_substitutivity_splits_original_features_train_substitutivity_splits_original_features_test2/checkpoint-1245/adapter_model.safetensors b/substitutivity_original_Estonian/Qwen3.5-2B-Base_substitutivity_splits_original_features_train_substitutivity_splits_original_features_test2/checkpoint-1245/adapter_model.safetensors deleted file mode 100644 index 76af8087a7015776bef122b7685aacb1c8e57d89..0000000000000000000000000000000000000000 --- a/substitutivity_original_Estonian/Qwen3.5-2B-Base_substitutivity_splits_original_features_train_substitutivity_splits_original_features_test2/checkpoint-1245/adapter_model.safetensors +++ /dev/null @@ -1,3 +0,0 @@ -version https://git-lfs.github.com/spec/v1 -oid sha256:087943ecb86e205c42cf5c713cebad8dbf7fbcd9684932dc51a7306cc95b4a5f -size 174613392 diff --git a/substitutivity_original_Estonian/Qwen3.5-2B-Base_substitutivity_splits_original_features_train_substitutivity_splits_original_features_test2/checkpoint-1245/chat_template.jinja b/substitutivity_original_Estonian/Qwen3.5-2B-Base_substitutivity_splits_original_features_train_substitutivity_splits_original_features_test2/checkpoint-1245/chat_template.jinja deleted file mode 100644 index 0ef09f214eaa6d9bca297988afc1454b5827b2c7..0000000000000000000000000000000000000000 --- a/substitutivity_original_Estonian/Qwen3.5-2B-Base_substitutivity_splits_original_features_train_substitutivity_splits_original_features_test2/checkpoint-1245/chat_template.jinja +++ /dev/null @@ -1,154 +0,0 @@ -{%- set image_count = namespace(value=0) %} -{%- set video_count = namespace(value=0) %} -{%- macro render_content(content, do_vision_count, is_system_content=false) %} - {%- if content is string %} - {{- content }} - {%- elif content is iterable and content is not mapping %} - {%- for item in content %} - {%- if 'image' in item or 'image_url' in item or item.type == 'image' %} - {%- if is_system_content %} - {{- raise_exception('System message cannot contain images.') }} - {%- endif %} - {%- if do_vision_count %} - {%- set image_count.value = image_count.value + 1 %} - {%- endif %} - {%- if add_vision_id %} - {{- 'Picture ' ~ image_count.value ~ ': ' }} - {%- endif %} - {{- '<|vision_start|><|image_pad|><|vision_end|>' }} - {%- elif 'video' in item or item.type == 'video' %} - {%- if is_system_content %} - {{- raise_exception('System message cannot contain videos.') }} - {%- endif %} - {%- if do_vision_count %} - {%- set video_count.value = video_count.value + 1 %} - {%- endif %} - {%- if add_vision_id %} - {{- 'Video ' ~ video_count.value ~ ': ' }} - {%- endif %} - {{- '<|vision_start|><|video_pad|><|vision_end|>' }} - {%- elif 'text' in item %} - {{- item.text }} - {%- else %} - {{- raise_exception('Unexpected item type in content.') }} - {%- endif %} - {%- endfor %} - {%- elif content is none or content is undefined %} - {{- '' }} - {%- else %} - {{- raise_exception('Unexpected content type.') }} - {%- endif %} -{%- endmacro %} -{%- if not messages %} - {{- raise_exception('No messages provided.') }} -{%- endif %} -{%- if tools and tools is iterable and tools is not mapping %} - {{- '<|im_start|>system\n' }} - {{- "# Tools\n\nYou have access to the following functions:\n\n" }} - {%- for tool in tools %} - {{- "\n" }} - {{- tool | tojson }} - {%- endfor %} - {{- "\n" }} - {{- '\n\nIf you choose to call a function ONLY reply in the following format with NO suffix:\n\n\n\n\nvalue_1\n\n\nThis is the value for the second parameter\nthat can span\nmultiple lines\n\n\n\n\n\nReminder:\n- Function calls MUST follow the specified format: an inner block must be nested within XML tags\n- Required parameters MUST be specified\n- You may provide optional reasoning for your function call in natural language BEFORE the function call, but NOT after\n- If there is no function call available, answer the question like normal with your current knowledge and do not tell the user about function calls\n' }} - {%- if messages[0].role == 'system' %} - {%- set content = render_content(messages[0].content, false, true)|trim %} - {%- if content %} - {{- '\n\n' + content }} - {%- endif %} - {%- endif %} - {{- '<|im_end|>\n' }} -{%- else %} - {%- if messages[0].role == 'system' %} - {%- set content = render_content(messages[0].content, false, true)|trim %} - {{- '<|im_start|>system\n' + 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" %} - {%- set content = render_content(message.content, false)|trim %} - {%- if not(content.startswith('') and content.endswith('')) %} - {%- set ns.multi_step_tool = false %} - {%- set ns.last_query_index = index %} - {%- endif %} - {%- endif %} -{%- endfor %} -{%- if ns.multi_step_tool %} - {{- raise_exception('No user query found in messages.') }} -{%- endif %} -{%- for message in messages %} - {%- set content = render_content(message.content, true)|trim %} - {%- if message.role == "system" %} - {%- if not loop.first %} - {{- raise_exception('System message must be at the beginning.') }} - {%- endif %} - {%- elif message.role == "user" %} - {{- '<|im_start|>' + message.role + '\n' + content + '<|im_end|>' + '\n' }} - {%- elif message.role == "assistant" %} - {%- set reasoning_content = '' %} - {%- if message.reasoning_content is string %} - {%- set reasoning_content = message.reasoning_content %} - {%- else %} - {%- if '' in content %} - {%- set reasoning_content = content.split('')[0].rstrip('\n').split('')[-1].lstrip('\n') %} - {%- set content = content.split('')[-1].lstrip('\n') %} - {%- endif %} - {%- endif %} - {%- set reasoning_content = reasoning_content|trim %} - {%- if loop.index0 > ns.last_query_index %} - {{- '<|im_start|>' + message.role + '\n\n' + reasoning_content + '\n\n\n' + content }} - {%- else %} - {{- '<|im_start|>' + message.role + '\n' + content }} - {%- endif %} - {%- if message.tool_calls and message.tool_calls is iterable and message.tool_calls is not mapping %} - {%- for tool_call in message.tool_calls %} - {%- if tool_call.function is defined %} - {%- set tool_call = tool_call.function %} - {%- endif %} - {%- if loop.first %} - {%- if content|trim %} - {{- '\n\n\n\n' }} - {%- else %} - {{- '\n\n' }} - {%- endif %} - {%- else %} - {{- '\n\n\n' }} - {%- endif %} - {%- if tool_call.arguments is defined %} - {%- for args_name, args_value in tool_call.arguments|items %} - {{- '\n' }} - {%- set args_value = args_value | tojson | safe if args_value is mapping or (args_value is sequence and args_value is not string) else args_value | string %} - {{- args_value }} - {{- '\n\n' }} - {%- endfor %} - {%- endif %} - {{- '\n' }} - {%- endfor %} - {%- endif %} - {{- '<|im_end|>\n' }} - {%- elif message.role == "tool" %} - {%- if loop.previtem and loop.previtem.role != "tool" %} - {{- '<|im_start|>user' }} - {%- endif %} - {{- '\n\n' }} - {{- content }} - {{- '\n' }} - {%- if not loop.last and loop.nextitem.role != "tool" %} - {{- '<|im_end|>\n' }} - {%- elif loop.last %} - {{- '<|im_end|>\n' }} - {%- endif %} - {%- else %} - {{- raise_exception('Unexpected message role.') }} - {%- endif %} -{%- endfor %} -{%- if add_generation_prompt %} - {{- '<|im_start|>assistant\n' }} - {%- if enable_thinking is defined and enable_thinking is true %} - {{- '\n' }} - {%- else %} - {{- '\n\n\n\n' }} - {%- endif %} -{%- endif %} \ No newline at end of file diff --git a/substitutivity_original_Estonian/Qwen3.5-2B-Base_substitutivity_splits_original_features_train_substitutivity_splits_original_features_test2/checkpoint-1245/tokenizer.json b/substitutivity_original_Estonian/Qwen3.5-2B-Base_substitutivity_splits_original_features_train_substitutivity_splits_original_features_test2/checkpoint-1245/tokenizer.json deleted file mode 100644 index 67741b04f23bfdb46501f748ce27865ec82eccfb..0000000000000000000000000000000000000000 --- a/substitutivity_original_Estonian/Qwen3.5-2B-Base_substitutivity_splits_original_features_train_substitutivity_splits_original_features_test2/checkpoint-1245/tokenizer.json +++ /dev/null @@ -1,3 +0,0 @@ -version https://git-lfs.github.com/spec/v1 -oid sha256:87a7830d63fcf43bf241c3c5242e96e62dd3fdc29224ca26fed8ea333db72de4 -size 19989343 diff --git a/substitutivity_original_Estonian/Qwen3.5-2B-Base_substitutivity_splits_original_features_train_substitutivity_splits_original_features_test2/checkpoint-1245/tokenizer_config.json b/substitutivity_original_Estonian/Qwen3.5-2B-Base_substitutivity_splits_original_features_train_substitutivity_splits_original_features_test2/checkpoint-1245/tokenizer_config.json deleted file mode 100644 index b4a37b2a6fd3ab3317cd7bac72855be1a843b2bb..0000000000000000000000000000000000000000 --- a/substitutivity_original_Estonian/Qwen3.5-2B-Base_substitutivity_splits_original_features_train_substitutivity_splits_original_features_test2/checkpoint-1245/tokenizer_config.json +++ /dev/null @@ -1,31 +0,0 @@ -{ - "add_prefix_space": false, - "audio_bos_token": "<|audio_start|>", - "audio_eos_token": "<|audio_end|>", - "audio_token": "<|audio_pad|>", - "backend": "tokenizers", - "bos_token": null, - "clean_up_tokenization_spaces": false, - "eos_token": "<|endoftext|>", - "errors": "replace", - "image_token": "<|image_pad|>", - "is_local": false, - "model_max_length": 262144, - "model_specific_special_tokens": { - "audio_bos_token": "<|audio_start|>", - "audio_eos_token": "<|audio_end|>", - "audio_token": "<|audio_pad|>", - "image_token": "<|image_pad|>", - "video_token": "<|video_pad|>", - "vision_bos_token": "<|vision_start|>", - "vision_eos_token": "<|vision_end|>" - }, - "pad_token": "<|endoftext|>", - "pretokenize_regex": "(?i:'s|'t|'re|'ve|'m|'ll|'d)|[^\\r\\n\\p{L}\\p{N}]?[\\p{L}\\p{M}]+|\\p{N}| ?[^\\s\\p{L}\\p{M}\\p{N}]+[\\r\\n]*|\\s*[\\r\\n]+|\\s+(?!\\S)|\\s+", - "split_special_tokens": false, - "tokenizer_class": "TokenizersBackend", - "unk_token": null, - "video_token": "<|video_pad|>", - "vision_bos_token": "<|vision_start|>", - "vision_eos_token": "<|vision_end|>" -} diff --git a/substitutivity_original_Estonian/Qwen3.5-2B-Base_substitutivity_splits_original_features_train_substitutivity_splits_original_features_test2/checkpoint-1245/trainer_state.json b/substitutivity_original_Estonian/Qwen3.5-2B-Base_substitutivity_splits_original_features_train_substitutivity_splits_original_features_test2/checkpoint-1245/trainer_state.json deleted file mode 100644 index fd108af7abd0b69531af4d747419e0d07288f1d4..0000000000000000000000000000000000000000 --- a/substitutivity_original_Estonian/Qwen3.5-2B-Base_substitutivity_splits_original_features_train_substitutivity_splits_original_features_test2/checkpoint-1245/trainer_state.json +++ /dev/null @@ -1,340 +0,0 @@ -{ - "best_global_step": null, - "best_metric": null, - "best_model_checkpoint": null, - "epoch": 3.0, - "eval_steps": 500, - "global_step": 1245, - "is_hyper_param_search": false, - "is_local_process_zero": true, - "is_world_process_zero": true, - "log_history": [ - { - "entropy": 2.2933380508422854, - "epoch": 0.12062726176115803, - "grad_norm": 2.8075337409973145, - "learning_rate": 1.5531017374629602e-05, - "loss": 2.1429234313964844, - "mean_token_accuracy": 0.5779684272408485, - "num_tokens": 121428.0, - "step": 50 - }, - { - "entropy": 1.1315062403678895, - "epoch": 0.24125452352231605, - "grad_norm": 1.9415371417999268, - "learning_rate": 3.1378994287516954e-05, - "loss": 1.0461698913574218, - "mean_token_accuracy": 0.7166991305351257, - "num_tokens": 243966.0, - "step": 100 - }, - { - "entropy": 0.9537658470869065, - "epoch": 0.3618817852834741, - "grad_norm": 1.8353157043457031, - "learning_rate": 4.72269712004043e-05, - "loss": 0.8774301147460938, - "mean_token_accuracy": 0.7506789201498032, - "num_tokens": 366649.0, - "step": 150 - }, - { - "entropy": 0.8955666983127594, - "epoch": 0.4825090470446321, - "grad_norm": 1.5419880151748657, - "learning_rate": 6.307494811329165e-05, - "loss": 0.8234302520751953, - "mean_token_accuracy": 0.7616239994764328, - "num_tokens": 488367.0, - "step": 200 - }, - { - "entropy": 0.8689469587802887, - "epoch": 0.6031363088057901, - "grad_norm": 1.5661484003067017, - "learning_rate": 7.8922925026179e-05, - "loss": 0.7961890411376953, - "mean_token_accuracy": 0.767715499997139, - "num_tokens": 603131.0, - "step": 250 - }, - { - "entropy": 0.8425652319192887, - "epoch": 0.7237635705669482, - "grad_norm": 1.5694905519485474, - "learning_rate": 9.477090193906636e-05, - "loss": 0.7718594360351563, - "mean_token_accuracy": 0.7736722755432129, - "num_tokens": 726412.0, - "step": 300 - }, - { - "entropy": 0.843294187784195, - "epoch": 0.8443908323281062, - "grad_norm": 1.1118028163909912, - "learning_rate": 0.00011061887885195371, - "loss": 0.7801297760009765, - "mean_token_accuracy": 0.7714760398864746, - "num_tokens": 846516.0, - "step": 350 - }, - { - "entropy": 0.8213214933872223, - "epoch": 0.9650180940892642, - "grad_norm": 0.9202918410301208, - "learning_rate": 0.00012646685576484106, - "loss": 0.7514453887939453, - "mean_token_accuracy": 0.7774044269323349, - "num_tokens": 958978.0, - "step": 400 - }, - { - "epoch": 1.0, - "eval_entropy": 0.818668900915746, - "eval_mean_token_accuracy": 0.7533496866065464, - "eval_not_syn_loss": 0.7798981070518494, - "eval_not_syn_runtime": 53.8744, - "eval_not_syn_samples_per_second": 26.358, - "eval_not_syn_steps_per_second": 3.304, - "eval_num_tokens": 993806.0, - "step": 415 - }, - { - "epoch": 1.0, - "eval_entropy": 0.7627468125873738, - "eval_mean_token_accuracy": 0.8024111574285486, - "eval_num_tokens": 993806.0, - "eval_syn_loss": 0.7218216061592102, - "eval_syn_runtime": 55.3239, - "eval_syn_samples_per_second": 25.667, - "eval_syn_steps_per_second": 3.217, - "step": 415 - }, - { - "entropy": 0.7802230679627621, - "epoch": 1.0844390832328106, - "grad_norm": 1.0145268440246582, - "learning_rate": 0.00013151131541121804, - "loss": 0.7111608123779297, - "mean_token_accuracy": 0.7845737542768921, - "num_tokens": 1080301.0, - "step": 450 - }, - { - "entropy": 0.7558828103542328, - "epoch": 1.2050663449939687, - "grad_norm": 1.0477591753005981, - "learning_rate": 0.0001313741160170038, - "loss": 0.6933037567138672, - "mean_token_accuracy": 0.7889221370220184, - "num_tokens": 1201215.0, - "step": 500 - }, - { - "entropy": 0.7593123841285706, - "epoch": 1.3256936067551266, - "grad_norm": 0.899161159992218, - "learning_rate": 0.00013112089697387701, - "loss": 0.6962493133544921, - "mean_token_accuracy": 0.7877359414100646, - "num_tokens": 1322895.0, - "step": 550 - }, - { - "entropy": 0.7460986822843552, - "epoch": 1.4463208685162847, - "grad_norm": 1.1061116456985474, - "learning_rate": 0.00013075210608885215, - "loss": 0.6817412567138672, - "mean_token_accuracy": 0.7904874366521836, - "num_tokens": 1436749.0, - "step": 600 - }, - { - "entropy": 0.735516265630722, - "epoch": 1.5669481302774426, - "grad_norm": 0.9925405979156494, - "learning_rate": 0.00013026839555278796, - "loss": 0.6690705871582031, - "mean_token_accuracy": 0.7930250608921051, - "num_tokens": 1555400.0, - "step": 650 - }, - { - "entropy": 0.7302519899606704, - "epoch": 1.6875753920386007, - "grad_norm": 0.800025463104248, - "learning_rate": 0.0001296706207870159, - "loss": 0.6714833831787109, - "mean_token_accuracy": 0.7946307575702667, - "num_tokens": 1676892.0, - "step": 700 - }, - { - "entropy": 0.7321263110637665, - "epoch": 1.8082026537997589, - "grad_norm": 1.0500682592391968, - "learning_rate": 0.00012895983893056423, - "loss": 0.6672868347167968, - "mean_token_accuracy": 0.7947645837068558, - "num_tokens": 1802689.0, - "step": 750 - }, - { - "entropy": 0.7075755137205124, - "epoch": 1.9288299155609168, - "grad_norm": 0.9248411655426025, - "learning_rate": 0.00012813730697065298, - "loss": 0.6466568756103516, - "mean_token_accuracy": 0.798747689127922, - "num_tokens": 1920804.0, - "step": 800 - }, - { - "epoch": 2.0, - "eval_entropy": 0.7426781811740961, - "eval_mean_token_accuracy": 0.7693507239389955, - "eval_not_syn_loss": 0.7139139175415039, - "eval_not_syn_runtime": 54.4286, - "eval_not_syn_samples_per_second": 26.089, - "eval_not_syn_steps_per_second": 3.27, - "eval_num_tokens": 1987612.0, - "step": 830 - }, - { - "epoch": 2.0, - "eval_entropy": 0.6937930352902144, - "eval_mean_token_accuracy": 0.8180048633157537, - "eval_num_tokens": 1987612.0, - "eval_syn_loss": 0.6489660143852234, - "eval_syn_runtime": 56.1525, - "eval_syn_samples_per_second": 25.288, - "eval_syn_steps_per_second": 3.17, - "step": 830 - }, - { - "entropy": 0.6853217440422135, - "epoch": 2.0482509047044632, - "grad_norm": 0.6932648420333862, - "learning_rate": 0.00012720447951976603, - "loss": 0.6212613296508789, - "mean_token_accuracy": 0.8047762579388089, - "num_tokens": 2039660.0, - "step": 850 - }, - { - "entropy": 0.6459032052755356, - "epoch": 2.168878166465621, - "grad_norm": 1.0688549280166626, - "learning_rate": 0.00012616300624323124, - "loss": 0.5740525817871094, - "mean_token_accuracy": 0.8148404061794281, - "num_tokens": 2161772.0, - "step": 900 - }, - { - "entropy": 0.6463843190670013, - "epoch": 2.2895054282267795, - "grad_norm": 0.847271203994751, - "learning_rate": 0.0001250147289418581, - "loss": 0.577717170715332, - "mean_token_accuracy": 0.814143658876419, - "num_tokens": 2280833.0, - "step": 950 - }, - { - "entropy": 0.6461306965351105, - "epoch": 2.4101326899879374, - "grad_norm": 0.9998957514762878, - "learning_rate": 0.00012376167829479217, - "loss": 0.5813011932373047, - "mean_token_accuracy": 0.8144401633739471, - "num_tokens": 2398206.0, - "step": 1000 - }, - { - "entropy": 0.640914659500122, - "epoch": 2.5307599517490953, - "grad_norm": 0.8692949414253235, - "learning_rate": 0.0001224060702683459, - "loss": 0.5755012512207032, - "mean_token_accuracy": 0.8168018108606339, - "num_tokens": 2518888.0, - "step": 1050 - }, - { - "entropy": 0.6421274638175964, - "epoch": 2.651387213510253, - "grad_norm": 0.8920196294784546, - "learning_rate": 0.00012095030219715738, - "loss": 0.5785483932495117, - "mean_token_accuracy": 0.8137972563505173, - "num_tokens": 2639763.0, - "step": 1100 - }, - { - "entropy": 0.6466812592744827, - "epoch": 2.772014475271411, - "grad_norm": 0.6876946687698364, - "learning_rate": 0.00011939694854460696, - "loss": 0.5816070938110351, - "mean_token_accuracy": 0.8145407867431641, - "num_tokens": 2755540.0, - "step": 1150 - }, - { - "entropy": 0.6378493067622185, - "epoch": 2.8926417370325694, - "grad_norm": 0.7532753944396973, - "learning_rate": 0.00011774875634998911, - "loss": 0.5754346084594727, - "mean_token_accuracy": 0.8148861181735992, - "num_tokens": 2874118.0, - "step": 1200 - }, - { - "epoch": 3.0, - "eval_entropy": 0.6912222209941136, - "eval_mean_token_accuracy": 0.7868467750174276, - "eval_not_syn_loss": 0.6782422661781311, - "eval_not_syn_runtime": 53.2627, - "eval_not_syn_samples_per_second": 26.66, - "eval_not_syn_steps_per_second": 3.342, - "eval_num_tokens": 2981418.0, - "step": 1245 - }, - { - "epoch": 3.0, - "eval_entropy": 0.6467212934172555, - "eval_mean_token_accuracy": 0.8215091817834405, - "eval_num_tokens": 2981418.0, - "eval_syn_loss": 0.6301999092102051, - "eval_syn_runtime": 55.0115, - "eval_syn_samples_per_second": 25.813, - "eval_syn_steps_per_second": 3.236, - "step": 1245 - } - ], - "logging_steps": 50, - "max_steps": 4150, - "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": 6.152012505473472e+16, - "train_batch_size": 8, - "trial_name": null, - "trial_params": null -} diff --git a/substitutivity_original_Estonian/Qwen3.5-2B-Base_substitutivity_splits_original_features_train_substitutivity_splits_original_features_test2/checkpoint-1245/training_args.bin b/substitutivity_original_Estonian/Qwen3.5-2B-Base_substitutivity_splits_original_features_train_substitutivity_splits_original_features_test2/checkpoint-1245/training_args.bin deleted file mode 100644 index f3d8098f65400413dd700f32ada5649470871017..0000000000000000000000000000000000000000 --- a/substitutivity_original_Estonian/Qwen3.5-2B-Base_substitutivity_splits_original_features_train_substitutivity_splits_original_features_test2/checkpoint-1245/training_args.bin +++ /dev/null @@ -1,3 +0,0 @@ -version https://git-lfs.github.com/spec/v1 -oid sha256:617af0f41414927baaaf158949585bb75eb4b05ff78baa5216c860ef0b431009 -size 6033 diff --git a/substitutivity_original_Estonian/Qwen3.5-2B-Base_substitutivity_splits_original_features_train_substitutivity_splits_original_features_test2/checkpoint-1660/README.md b/substitutivity_original_Estonian/Qwen3.5-2B-Base_substitutivity_splits_original_features_train_substitutivity_splits_original_features_test2/checkpoint-1660/README.md deleted file mode 100644 index ece204d005f7bf161d0e448efdd56fd297b7c8a2..0000000000000000000000000000000000000000 --- a/substitutivity_original_Estonian/Qwen3.5-2B-Base_substitutivity_splits_original_features_train_substitutivity_splits_original_features_test2/checkpoint-1660/README.md +++ /dev/null @@ -1,209 +0,0 @@ ---- -base_model: Qwen/Qwen3.5-2B-Base -library_name: peft -pipeline_tag: text-generation -tags: -- base_model:adapter:Qwen/Qwen3.5-2B-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/substitutivity_original_Estonian/Qwen3.5-2B-Base_substitutivity_splits_original_features_train_substitutivity_splits_original_features_test2/checkpoint-1660/adapter_config.json b/substitutivity_original_Estonian/Qwen3.5-2B-Base_substitutivity_splits_original_features_train_substitutivity_splits_original_features_test2/checkpoint-1660/adapter_config.json deleted file mode 100644 index 6a095bb3bfe249ce8e15df6ae999e88fa1e17162..0000000000000000000000000000000000000000 --- a/substitutivity_original_Estonian/Qwen3.5-2B-Base_substitutivity_splits_original_features_train_substitutivity_splits_original_features_test2/checkpoint-1660/adapter_config.json +++ /dev/null @@ -1,46 +0,0 @@ -{ - "alora_invocation_tokens": null, - "alpha_pattern": {}, - "arrow_config": null, - "auto_mapping": null, - "base_model_name_or_path": "Qwen/Qwen3.5-2B-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": 0.05731002492184622, - "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": [ - "o_proj", - "q_proj", - "k_proj", - "down_proj", - "up_proj", - "v_proj", - "gate_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/substitutivity_original_Estonian/Qwen3.5-2B-Base_substitutivity_splits_original_features_train_substitutivity_splits_original_features_test2/checkpoint-1660/adapter_model.safetensors b/substitutivity_original_Estonian/Qwen3.5-2B-Base_substitutivity_splits_original_features_train_substitutivity_splits_original_features_test2/checkpoint-1660/adapter_model.safetensors deleted file mode 100644 index 024d54e227ef8990aeaf7e7c2320bba30ebb1734..0000000000000000000000000000000000000000 --- a/substitutivity_original_Estonian/Qwen3.5-2B-Base_substitutivity_splits_original_features_train_substitutivity_splits_original_features_test2/checkpoint-1660/adapter_model.safetensors +++ /dev/null @@ -1,3 +0,0 @@ -version https://git-lfs.github.com/spec/v1 -oid sha256:ddec79cfc1192cb25e8cf814c414f7c3fd24cd921b4cf609a2fde870affd40c5 -size 174613392 diff --git a/substitutivity_original_Estonian/Qwen3.5-2B-Base_substitutivity_splits_original_features_train_substitutivity_splits_original_features_test2/checkpoint-1660/chat_template.jinja b/substitutivity_original_Estonian/Qwen3.5-2B-Base_substitutivity_splits_original_features_train_substitutivity_splits_original_features_test2/checkpoint-1660/chat_template.jinja deleted file mode 100644 index 0ef09f214eaa6d9bca297988afc1454b5827b2c7..0000000000000000000000000000000000000000 --- a/substitutivity_original_Estonian/Qwen3.5-2B-Base_substitutivity_splits_original_features_train_substitutivity_splits_original_features_test2/checkpoint-1660/chat_template.jinja +++ /dev/null @@ -1,154 +0,0 @@ -{%- set image_count = namespace(value=0) %} -{%- set video_count = namespace(value=0) %} -{%- macro render_content(content, do_vision_count, is_system_content=false) %} - {%- if content is string %} - {{- content }} - {%- elif content is iterable and content is not mapping %} - {%- for item in content %} - {%- if 'image' in item or 'image_url' in item or item.type == 'image' %} - {%- if is_system_content %} - {{- raise_exception('System message cannot contain images.') }} - {%- endif %} - {%- if do_vision_count %} - {%- set image_count.value = image_count.value + 1 %} - {%- endif %} - {%- if add_vision_id %} - {{- 'Picture ' ~ image_count.value ~ ': ' }} - {%- endif %} - {{- '<|vision_start|><|image_pad|><|vision_end|>' }} - {%- elif 'video' in item or item.type == 'video' %} - {%- if is_system_content %} - {{- raise_exception('System message cannot contain videos.') }} - {%- endif %} - {%- if do_vision_count %} - {%- set video_count.value = video_count.value + 1 %} - {%- endif %} - {%- if add_vision_id %} - {{- 'Video ' ~ video_count.value ~ ': ' }} - {%- endif %} - {{- '<|vision_start|><|video_pad|><|vision_end|>' }} - {%- elif 'text' in item %} - {{- item.text }} - {%- else %} - {{- raise_exception('Unexpected item type in content.') }} - {%- endif %} - {%- endfor %} - {%- elif content is none or content is undefined %} - {{- '' }} - {%- else %} - {{- raise_exception('Unexpected content type.') }} - {%- endif %} -{%- endmacro %} -{%- if not messages %} - {{- raise_exception('No messages provided.') }} -{%- endif %} -{%- if tools and tools is iterable and tools is not mapping %} - {{- '<|im_start|>system\n' }} - {{- "# Tools\n\nYou have access to the following functions:\n\n" }} - {%- for tool in tools %} - {{- "\n" }} - {{- tool | tojson }} - {%- endfor %} - {{- "\n" }} - {{- '\n\nIf you choose to call a function ONLY reply in the following format with NO suffix:\n\n\n\n\nvalue_1\n\n\nThis is the value for the second parameter\nthat can span\nmultiple lines\n\n\n\n\n\nReminder:\n- Function calls MUST follow the specified format: an inner block must be nested within XML tags\n- Required parameters MUST be specified\n- You may provide optional reasoning for your function call in natural language BEFORE the function call, but NOT after\n- If there is no function call available, answer the question like normal with your current knowledge and do not tell the user about function calls\n' }} - {%- if messages[0].role == 'system' %} - {%- set content = render_content(messages[0].content, false, true)|trim %} - {%- if content %} - {{- '\n\n' + content }} - {%- endif %} - {%- endif %} - {{- '<|im_end|>\n' }} -{%- else %} - {%- if messages[0].role == 'system' %} - {%- set content = render_content(messages[0].content, false, true)|trim %} - {{- '<|im_start|>system\n' + 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" %} - {%- set content = render_content(message.content, false)|trim %} - {%- if not(content.startswith('') and content.endswith('')) %} - {%- set ns.multi_step_tool = false %} - {%- set ns.last_query_index = index %} - {%- endif %} - {%- endif %} -{%- endfor %} -{%- if ns.multi_step_tool %} - {{- raise_exception('No user query found in messages.') }} -{%- endif %} -{%- for message in messages %} - {%- set content = render_content(message.content, true)|trim %} - {%- if message.role == "system" %} - {%- if not loop.first %} - {{- raise_exception('System message must be at the beginning.') }} - {%- endif %} - {%- elif message.role == "user" %} - {{- '<|im_start|>' + message.role + '\n' + content + '<|im_end|>' + '\n' }} - {%- elif message.role == "assistant" %} - {%- set reasoning_content = '' %} - {%- if message.reasoning_content is string %} - {%- set reasoning_content = message.reasoning_content %} - {%- else %} - {%- if '' in content %} - {%- set reasoning_content = content.split('')[0].rstrip('\n').split('')[-1].lstrip('\n') %} - {%- set content = content.split('')[-1].lstrip('\n') %} - {%- endif %} - {%- endif %} - {%- set reasoning_content = reasoning_content|trim %} - {%- if loop.index0 > ns.last_query_index %} - {{- '<|im_start|>' + message.role + '\n\n' + reasoning_content + '\n\n\n' + content }} - {%- else %} - {{- '<|im_start|>' + message.role + '\n' + content }} - {%- endif %} - {%- if message.tool_calls and message.tool_calls is iterable and message.tool_calls is not mapping %} - {%- for tool_call in message.tool_calls %} - {%- if tool_call.function is defined %} - {%- set tool_call = tool_call.function %} - {%- endif %} - {%- if loop.first %} - {%- if content|trim %} - {{- '\n\n\n\n' }} - {%- else %} - {{- '\n\n' }} - {%- endif %} - {%- else %} - {{- '\n\n\n' }} - {%- endif %} - {%- if tool_call.arguments is defined %} - {%- for args_name, args_value in tool_call.arguments|items %} - {{- '\n' }} - {%- set args_value = args_value | tojson | safe if args_value is mapping or (args_value is sequence and args_value is not string) else args_value | string %} - {{- args_value }} - {{- '\n\n' }} - {%- endfor %} - {%- endif %} - {{- '\n' }} - {%- endfor %} - {%- endif %} - {{- '<|im_end|>\n' }} - {%- elif message.role == "tool" %} - {%- if loop.previtem and loop.previtem.role != "tool" %} - {{- '<|im_start|>user' }} - {%- endif %} - {{- '\n\n' }} - {{- content }} - {{- '\n' }} - {%- if not loop.last and loop.nextitem.role != "tool" %} - {{- '<|im_end|>\n' }} - {%- elif loop.last %} - {{- '<|im_end|>\n' }} - {%- endif %} - {%- else %} - {{- raise_exception('Unexpected message role.') }} - {%- endif %} -{%- endfor %} -{%- if add_generation_prompt %} - {{- '<|im_start|>assistant\n' }} - {%- if enable_thinking is defined and enable_thinking is true %} - {{- '\n' }} - {%- else %} - {{- '\n\n\n\n' }} - {%- endif %} -{%- endif %} \ No newline at end of file diff --git a/substitutivity_original_Estonian/Qwen3.5-2B-Base_substitutivity_splits_original_features_train_substitutivity_splits_original_features_test2/checkpoint-1660/tokenizer.json b/substitutivity_original_Estonian/Qwen3.5-2B-Base_substitutivity_splits_original_features_train_substitutivity_splits_original_features_test2/checkpoint-1660/tokenizer.json deleted file mode 100644 index 67741b04f23bfdb46501f748ce27865ec82eccfb..0000000000000000000000000000000000000000 --- a/substitutivity_original_Estonian/Qwen3.5-2B-Base_substitutivity_splits_original_features_train_substitutivity_splits_original_features_test2/checkpoint-1660/tokenizer.json +++ /dev/null @@ -1,3 +0,0 @@ -version https://git-lfs.github.com/spec/v1 -oid sha256:87a7830d63fcf43bf241c3c5242e96e62dd3fdc29224ca26fed8ea333db72de4 -size 19989343 diff --git a/substitutivity_original_Estonian/Qwen3.5-2B-Base_substitutivity_splits_original_features_train_substitutivity_splits_original_features_test2/checkpoint-1660/tokenizer_config.json b/substitutivity_original_Estonian/Qwen3.5-2B-Base_substitutivity_splits_original_features_train_substitutivity_splits_original_features_test2/checkpoint-1660/tokenizer_config.json deleted file mode 100644 index b4a37b2a6fd3ab3317cd7bac72855be1a843b2bb..0000000000000000000000000000000000000000 --- a/substitutivity_original_Estonian/Qwen3.5-2B-Base_substitutivity_splits_original_features_train_substitutivity_splits_original_features_test2/checkpoint-1660/tokenizer_config.json +++ /dev/null @@ -1,31 +0,0 @@ -{ - "add_prefix_space": false, - "audio_bos_token": "<|audio_start|>", - "audio_eos_token": "<|audio_end|>", - "audio_token": "<|audio_pad|>", - "backend": "tokenizers", - "bos_token": null, - "clean_up_tokenization_spaces": false, - "eos_token": "<|endoftext|>", - "errors": "replace", - "image_token": "<|image_pad|>", - "is_local": false, - "model_max_length": 262144, - "model_specific_special_tokens": { - "audio_bos_token": "<|audio_start|>", - "audio_eos_token": "<|audio_end|>", - "audio_token": "<|audio_pad|>", - "image_token": "<|image_pad|>", - "video_token": "<|video_pad|>", - "vision_bos_token": "<|vision_start|>", - "vision_eos_token": "<|vision_end|>" - }, - "pad_token": "<|endoftext|>", - "pretokenize_regex": "(?i:'s|'t|'re|'ve|'m|'ll|'d)|[^\\r\\n\\p{L}\\p{N}]?[\\p{L}\\p{M}]+|\\p{N}| ?[^\\s\\p{L}\\p{M}\\p{N}]+[\\r\\n]*|\\s*[\\r\\n]+|\\s+(?!\\S)|\\s+", - "split_special_tokens": false, - "tokenizer_class": "TokenizersBackend", - "unk_token": null, - "video_token": "<|video_pad|>", - "vision_bos_token": "<|vision_start|>", - "vision_eos_token": "<|vision_end|>" -} diff --git a/substitutivity_original_Estonian/Qwen3.5-2B-Base_substitutivity_splits_original_features_train_substitutivity_splits_original_features_test2/checkpoint-1660/trainer_state.json b/substitutivity_original_Estonian/Qwen3.5-2B-Base_substitutivity_splits_original_features_train_substitutivity_splits_original_features_test2/checkpoint-1660/trainer_state.json deleted file mode 100644 index 76f23c336a209772d26827e04d0f96f758000a03..0000000000000000000000000000000000000000 --- a/substitutivity_original_Estonian/Qwen3.5-2B-Base_substitutivity_splits_original_features_train_substitutivity_splits_original_features_test2/checkpoint-1660/trainer_state.json +++ /dev/null @@ -1,452 +0,0 @@ -{ - "best_global_step": null, - "best_metric": null, - "best_model_checkpoint": null, - "epoch": 4.0, - "eval_steps": 500, - "global_step": 1660, - "is_hyper_param_search": false, - "is_local_process_zero": true, - "is_world_process_zero": true, - "log_history": [ - { - "entropy": 2.2933380508422854, - "epoch": 0.12062726176115803, - "grad_norm": 2.8075337409973145, - "learning_rate": 1.5531017374629602e-05, - "loss": 2.1429234313964844, - "mean_token_accuracy": 0.5779684272408485, - "num_tokens": 121428.0, - "step": 50 - }, - { - "entropy": 1.1315062403678895, - "epoch": 0.24125452352231605, - "grad_norm": 1.9415371417999268, - "learning_rate": 3.1378994287516954e-05, - "loss": 1.0461698913574218, - "mean_token_accuracy": 0.7166991305351257, - "num_tokens": 243966.0, - "step": 100 - }, - { - "entropy": 0.9537658470869065, - "epoch": 0.3618817852834741, - "grad_norm": 1.8353157043457031, - "learning_rate": 4.72269712004043e-05, - "loss": 0.8774301147460938, - "mean_token_accuracy": 0.7506789201498032, - "num_tokens": 366649.0, - "step": 150 - }, - { - "entropy": 0.8955666983127594, - "epoch": 0.4825090470446321, - "grad_norm": 1.5419880151748657, - "learning_rate": 6.307494811329165e-05, - "loss": 0.8234302520751953, - "mean_token_accuracy": 0.7616239994764328, - "num_tokens": 488367.0, - "step": 200 - }, - { - "entropy": 0.8689469587802887, - "epoch": 0.6031363088057901, - "grad_norm": 1.5661484003067017, - "learning_rate": 7.8922925026179e-05, - "loss": 0.7961890411376953, - "mean_token_accuracy": 0.767715499997139, - "num_tokens": 603131.0, - "step": 250 - }, - { - "entropy": 0.8425652319192887, - "epoch": 0.7237635705669482, - "grad_norm": 1.5694905519485474, - "learning_rate": 9.477090193906636e-05, - "loss": 0.7718594360351563, - "mean_token_accuracy": 0.7736722755432129, - "num_tokens": 726412.0, - "step": 300 - }, - { - "entropy": 0.843294187784195, - "epoch": 0.8443908323281062, - "grad_norm": 1.1118028163909912, - "learning_rate": 0.00011061887885195371, - "loss": 0.7801297760009765, - "mean_token_accuracy": 0.7714760398864746, - "num_tokens": 846516.0, - "step": 350 - }, - { - "entropy": 0.8213214933872223, - "epoch": 0.9650180940892642, - "grad_norm": 0.9202918410301208, - "learning_rate": 0.00012646685576484106, - "loss": 0.7514453887939453, - "mean_token_accuracy": 0.7774044269323349, - "num_tokens": 958978.0, - "step": 400 - }, - { - "epoch": 1.0, - "eval_entropy": 0.818668900915746, - "eval_mean_token_accuracy": 0.7533496866065464, - "eval_not_syn_loss": 0.7798981070518494, - "eval_not_syn_runtime": 53.8744, - "eval_not_syn_samples_per_second": 26.358, - "eval_not_syn_steps_per_second": 3.304, - "eval_num_tokens": 993806.0, - "step": 415 - }, - { - "epoch": 1.0, - "eval_entropy": 0.7627468125873738, - "eval_mean_token_accuracy": 0.8024111574285486, - "eval_num_tokens": 993806.0, - "eval_syn_loss": 0.7218216061592102, - "eval_syn_runtime": 55.3239, - "eval_syn_samples_per_second": 25.667, - "eval_syn_steps_per_second": 3.217, - "step": 415 - }, - { - "entropy": 0.7802230679627621, - "epoch": 1.0844390832328106, - "grad_norm": 1.0145268440246582, - "learning_rate": 0.00013151131541121804, - "loss": 0.7111608123779297, - "mean_token_accuracy": 0.7845737542768921, - "num_tokens": 1080301.0, - "step": 450 - }, - { - "entropy": 0.7558828103542328, - "epoch": 1.2050663449939687, - "grad_norm": 1.0477591753005981, - "learning_rate": 0.0001313741160170038, - "loss": 0.6933037567138672, - "mean_token_accuracy": 0.7889221370220184, - "num_tokens": 1201215.0, - "step": 500 - }, - { - "entropy": 0.7593123841285706, - "epoch": 1.3256936067551266, - "grad_norm": 0.899161159992218, - "learning_rate": 0.00013112089697387701, - "loss": 0.6962493133544921, - "mean_token_accuracy": 0.7877359414100646, - "num_tokens": 1322895.0, - "step": 550 - }, - { - "entropy": 0.7460986822843552, - "epoch": 1.4463208685162847, - "grad_norm": 1.1061116456985474, - "learning_rate": 0.00013075210608885215, - "loss": 0.6817412567138672, - "mean_token_accuracy": 0.7904874366521836, - "num_tokens": 1436749.0, - "step": 600 - }, - { - "entropy": 0.735516265630722, - "epoch": 1.5669481302774426, - "grad_norm": 0.9925405979156494, - "learning_rate": 0.00013026839555278796, - "loss": 0.6690705871582031, - "mean_token_accuracy": 0.7930250608921051, - "num_tokens": 1555400.0, - "step": 650 - }, - { - "entropy": 0.7302519899606704, - "epoch": 1.6875753920386007, - "grad_norm": 0.800025463104248, - "learning_rate": 0.0001296706207870159, - "loss": 0.6714833831787109, - "mean_token_accuracy": 0.7946307575702667, - "num_tokens": 1676892.0, - "step": 700 - }, - { - "entropy": 0.7321263110637665, - "epoch": 1.8082026537997589, - "grad_norm": 1.0500682592391968, - "learning_rate": 0.00012895983893056423, - "loss": 0.6672868347167968, - "mean_token_accuracy": 0.7947645837068558, - "num_tokens": 1802689.0, - "step": 750 - }, - { - "entropy": 0.7075755137205124, - "epoch": 1.9288299155609168, - "grad_norm": 0.9248411655426025, - "learning_rate": 0.00012813730697065298, - "loss": 0.6466568756103516, - "mean_token_accuracy": 0.798747689127922, - "num_tokens": 1920804.0, - "step": 800 - }, - { - "epoch": 2.0, - "eval_entropy": 0.7426781811740961, - "eval_mean_token_accuracy": 0.7693507239389955, - "eval_not_syn_loss": 0.7139139175415039, - "eval_not_syn_runtime": 54.4286, - "eval_not_syn_samples_per_second": 26.089, - "eval_not_syn_steps_per_second": 3.27, - "eval_num_tokens": 1987612.0, - "step": 830 - }, - { - "epoch": 2.0, - "eval_entropy": 0.6937930352902144, - "eval_mean_token_accuracy": 0.8180048633157537, - "eval_num_tokens": 1987612.0, - "eval_syn_loss": 0.6489660143852234, - "eval_syn_runtime": 56.1525, - "eval_syn_samples_per_second": 25.288, - "eval_syn_steps_per_second": 3.17, - "step": 830 - }, - { - "entropy": 0.6853217440422135, - "epoch": 2.0482509047044632, - "grad_norm": 0.6932648420333862, - "learning_rate": 0.00012720447951976603, - "loss": 0.6212613296508789, - "mean_token_accuracy": 0.8047762579388089, - "num_tokens": 2039660.0, - "step": 850 - }, - { - "entropy": 0.6459032052755356, - "epoch": 2.168878166465621, - "grad_norm": 1.0688549280166626, - "learning_rate": 0.00012616300624323124, - "loss": 0.5740525817871094, - "mean_token_accuracy": 0.8148404061794281, - "num_tokens": 2161772.0, - "step": 900 - }, - { - "entropy": 0.6463843190670013, - "epoch": 2.2895054282267795, - "grad_norm": 0.847271203994751, - "learning_rate": 0.0001250147289418581, - "loss": 0.577717170715332, - "mean_token_accuracy": 0.814143658876419, - "num_tokens": 2280833.0, - "step": 950 - }, - { - "entropy": 0.6461306965351105, - "epoch": 2.4101326899879374, - "grad_norm": 0.9998957514762878, - "learning_rate": 0.00012376167829479217, - "loss": 0.5813011932373047, - "mean_token_accuracy": 0.8144401633739471, - "num_tokens": 2398206.0, - "step": 1000 - }, - { - "entropy": 0.640914659500122, - "epoch": 2.5307599517490953, - "grad_norm": 0.8692949414253235, - "learning_rate": 0.0001224060702683459, - "loss": 0.5755012512207032, - "mean_token_accuracy": 0.8168018108606339, - "num_tokens": 2518888.0, - "step": 1050 - }, - { - "entropy": 0.6421274638175964, - "epoch": 2.651387213510253, - "grad_norm": 0.8920196294784546, - "learning_rate": 0.00012095030219715738, - "loss": 0.5785483932495117, - "mean_token_accuracy": 0.8137972563505173, - "num_tokens": 2639763.0, - "step": 1100 - }, - { - "entropy": 0.6466812592744827, - "epoch": 2.772014475271411, - "grad_norm": 0.6876946687698364, - "learning_rate": 0.00011939694854460696, - "loss": 0.5816070938110351, - "mean_token_accuracy": 0.8145407867431641, - "num_tokens": 2755540.0, - "step": 1150 - }, - { - "entropy": 0.6378493067622185, - "epoch": 2.8926417370325694, - "grad_norm": 0.7532753944396973, - "learning_rate": 0.00011774875634998911, - "loss": 0.5754346084594727, - "mean_token_accuracy": 0.8148861181735992, - "num_tokens": 2874118.0, - "step": 1200 - }, - { - "epoch": 3.0, - "eval_entropy": 0.6912222209941136, - "eval_mean_token_accuracy": 0.7868467750174276, - "eval_not_syn_loss": 0.6782422661781311, - "eval_not_syn_runtime": 53.2627, - "eval_not_syn_samples_per_second": 26.66, - "eval_not_syn_steps_per_second": 3.342, - "eval_num_tokens": 2981418.0, - "step": 1245 - }, - { - "epoch": 3.0, - "eval_entropy": 0.6467212934172555, - "eval_mean_token_accuracy": 0.8215091817834405, - "eval_num_tokens": 2981418.0, - "eval_syn_loss": 0.6301999092102051, - "eval_syn_runtime": 55.0115, - "eval_syn_samples_per_second": 25.813, - "eval_syn_steps_per_second": 3.236, - "step": 1245 - }, - { - "entropy": 0.6316030452949832, - "epoch": 3.012062726176116, - "grad_norm": 0.7761592864990234, - "learning_rate": 0.00011600864037049124, - "loss": 0.5697880935668945, - "mean_token_accuracy": 0.8159356400220081, - "num_tokens": 2994602.0, - "step": 1250 - }, - { - "entropy": 0.5467158386111259, - "epoch": 3.132689987937274, - "grad_norm": 0.8169087767601013, - "learning_rate": 0.00011417967792657078, - "loss": 0.47844127655029295, - "mean_token_accuracy": 0.8378052765130997, - "num_tokens": 3121094.0, - "step": 1300 - }, - { - "entropy": 0.5454566389322281, - "epoch": 3.2533172496984317, - "grad_norm": 0.8219636678695679, - "learning_rate": 0.0001122651034598459, - "loss": 0.4839338302612305, - "mean_token_accuracy": 0.8385071897506714, - "num_tokens": 3248038.0, - "step": 1350 - }, - { - "entropy": 0.5598723018169403, - "epoch": 3.37394451145959, - "grad_norm": 0.9956360459327698, - "learning_rate": 0.00011026830281312423, - "loss": 0.492743034362793, - "mean_token_accuracy": 0.8357370454072952, - "num_tokens": 3363487.0, - "step": 1400 - }, - { - "entropy": 0.5479365903139114, - "epoch": 3.494571773220748, - "grad_norm": 0.8586823344230652, - "learning_rate": 0.00010819280724268524, - "loss": 0.4810820770263672, - "mean_token_accuracy": 0.837928666472435, - "num_tokens": 3480513.0, - "step": 1450 - }, - { - "entropy": 0.5619524419307709, - "epoch": 3.615199034981906, - "grad_norm": 1.0894954204559326, - "learning_rate": 0.0001060422871734048, - "loss": 0.4974747085571289, - "mean_token_accuracy": 0.8349784809350967, - "num_tokens": 3597183.0, - "step": 1500 - }, - { - "entropy": 0.5588027596473694, - "epoch": 3.7358262967430638, - "grad_norm": 0.9905257225036621, - "learning_rate": 0.00010382054570776654, - "loss": 0.49687202453613283, - "mean_token_accuracy": 0.834458841085434, - "num_tokens": 3714173.0, - "step": 1550 - }, - { - "entropy": 0.5654991599917412, - "epoch": 3.856453558504222, - "grad_norm": 1.0116610527038574, - "learning_rate": 0.00010153151190023791, - "loss": 0.5016670989990234, - "mean_token_accuracy": 0.8326652735471726, - "num_tokens": 3832652.0, - "step": 1600 - }, - { - "entropy": 0.5605382058024406, - "epoch": 3.97708082026538, - "grad_norm": 0.8011100888252258, - "learning_rate": 9.917923380890624e-05, - "loss": 0.49551734924316404, - "mean_token_accuracy": 0.8343535929918289, - "num_tokens": 3955167.0, - "step": 1650 - }, - { - "epoch": 4.0, - "eval_entropy": 0.5819476859623127, - "eval_mean_token_accuracy": 0.7963069781158747, - "eval_not_syn_loss": 0.6924306750297546, - "eval_not_syn_runtime": 54.027, - "eval_not_syn_samples_per_second": 26.283, - "eval_not_syn_steps_per_second": 3.295, - "eval_num_tokens": 3975224.0, - "step": 1660 - }, - { - "epoch": 4.0, - "eval_entropy": 0.545470636212424, - "eval_mean_token_accuracy": 0.8127643666910321, - "eval_num_tokens": 3975224.0, - "eval_syn_loss": 0.649576723575592, - "eval_syn_runtime": 55.2296, - "eval_syn_samples_per_second": 25.711, - "eval_syn_steps_per_second": 3.223, - "step": 1660 - } - ], - "logging_steps": 50, - "max_steps": 4150, - "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": 8.227704885453773e+16, - "train_batch_size": 8, - "trial_name": null, - "trial_params": null -} diff --git a/substitutivity_original_Estonian/Qwen3.5-2B-Base_substitutivity_splits_original_features_train_substitutivity_splits_original_features_test2/checkpoint-1660/training_args.bin b/substitutivity_original_Estonian/Qwen3.5-2B-Base_substitutivity_splits_original_features_train_substitutivity_splits_original_features_test2/checkpoint-1660/training_args.bin deleted file mode 100644 index f3d8098f65400413dd700f32ada5649470871017..0000000000000000000000000000000000000000 --- a/substitutivity_original_Estonian/Qwen3.5-2B-Base_substitutivity_splits_original_features_train_substitutivity_splits_original_features_test2/checkpoint-1660/training_args.bin +++ /dev/null @@ -1,3 +0,0 @@ -version https://git-lfs.github.com/spec/v1 -oid sha256:617af0f41414927baaaf158949585bb75eb4b05ff78baa5216c860ef0b431009 -size 6033 diff --git a/substitutivity_original_Estonian/Qwen3.5-2B-Base_substitutivity_splits_original_features_train_substitutivity_splits_original_features_test2/checkpoint-2075/README.md b/substitutivity_original_Estonian/Qwen3.5-2B-Base_substitutivity_splits_original_features_train_substitutivity_splits_original_features_test2/checkpoint-2075/README.md deleted file mode 100644 index ece204d005f7bf161d0e448efdd56fd297b7c8a2..0000000000000000000000000000000000000000 --- a/substitutivity_original_Estonian/Qwen3.5-2B-Base_substitutivity_splits_original_features_train_substitutivity_splits_original_features_test2/checkpoint-2075/README.md +++ /dev/null @@ -1,209 +0,0 @@ ---- -base_model: Qwen/Qwen3.5-2B-Base -library_name: peft -pipeline_tag: text-generation -tags: -- base_model:adapter:Qwen/Qwen3.5-2B-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/substitutivity_original_Estonian/Qwen3.5-2B-Base_substitutivity_splits_original_features_train_substitutivity_splits_original_features_test2/checkpoint-2075/adapter_config.json b/substitutivity_original_Estonian/Qwen3.5-2B-Base_substitutivity_splits_original_features_train_substitutivity_splits_original_features_test2/checkpoint-2075/adapter_config.json deleted file mode 100644 index 6a095bb3bfe249ce8e15df6ae999e88fa1e17162..0000000000000000000000000000000000000000 --- a/substitutivity_original_Estonian/Qwen3.5-2B-Base_substitutivity_splits_original_features_train_substitutivity_splits_original_features_test2/checkpoint-2075/adapter_config.json +++ /dev/null @@ -1,46 +0,0 @@ -{ - "alora_invocation_tokens": null, - "alpha_pattern": {}, - "arrow_config": null, - "auto_mapping": null, - "base_model_name_or_path": "Qwen/Qwen3.5-2B-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": 0.05731002492184622, - "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": [ - "o_proj", - "q_proj", - "k_proj", - "down_proj", - "up_proj", - "v_proj", - "gate_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/substitutivity_original_Estonian/Qwen3.5-2B-Base_substitutivity_splits_original_features_train_substitutivity_splits_original_features_test2/checkpoint-2075/adapter_model.safetensors b/substitutivity_original_Estonian/Qwen3.5-2B-Base_substitutivity_splits_original_features_train_substitutivity_splits_original_features_test2/checkpoint-2075/adapter_model.safetensors deleted file mode 100644 index 8c7f8b38efbb94adec04bc64084f0fa13aadf22a..0000000000000000000000000000000000000000 --- a/substitutivity_original_Estonian/Qwen3.5-2B-Base_substitutivity_splits_original_features_train_substitutivity_splits_original_features_test2/checkpoint-2075/adapter_model.safetensors +++ /dev/null @@ -1,3 +0,0 @@ -version https://git-lfs.github.com/spec/v1 -oid sha256:522c1d59ca4c88adf116d836caf13c0738d160018ecd90a455898b22928f2d8e -size 174613392 diff --git a/substitutivity_original_Estonian/Qwen3.5-2B-Base_substitutivity_splits_original_features_train_substitutivity_splits_original_features_test2/checkpoint-2075/chat_template.jinja b/substitutivity_original_Estonian/Qwen3.5-2B-Base_substitutivity_splits_original_features_train_substitutivity_splits_original_features_test2/checkpoint-2075/chat_template.jinja deleted file mode 100644 index 0ef09f214eaa6d9bca297988afc1454b5827b2c7..0000000000000000000000000000000000000000 --- a/substitutivity_original_Estonian/Qwen3.5-2B-Base_substitutivity_splits_original_features_train_substitutivity_splits_original_features_test2/checkpoint-2075/chat_template.jinja +++ /dev/null @@ -1,154 +0,0 @@ -{%- set image_count = namespace(value=0) %} -{%- set video_count = namespace(value=0) %} -{%- macro render_content(content, do_vision_count, is_system_content=false) %} - {%- if content is string %} - {{- content }} - {%- elif content is iterable and content is not mapping %} - {%- for item in content %} - {%- if 'image' in item or 'image_url' in item or item.type == 'image' %} - {%- if is_system_content %} - {{- raise_exception('System message cannot contain images.') }} - {%- endif %} - {%- if do_vision_count %} - {%- set image_count.value = image_count.value + 1 %} - {%- endif %} - {%- if add_vision_id %} - {{- 'Picture ' ~ image_count.value ~ ': ' }} - {%- endif %} - {{- '<|vision_start|><|image_pad|><|vision_end|>' }} - {%- elif 'video' in item or item.type == 'video' %} - {%- if is_system_content %} - {{- raise_exception('System message cannot contain videos.') }} - {%- endif %} - {%- if do_vision_count %} - {%- set video_count.value = video_count.value + 1 %} - {%- endif %} - {%- if add_vision_id %} - {{- 'Video ' ~ video_count.value ~ ': ' }} - {%- endif %} - {{- '<|vision_start|><|video_pad|><|vision_end|>' }} - {%- elif 'text' in item %} - {{- item.text }} - {%- else %} - {{- raise_exception('Unexpected item type in content.') }} - {%- endif %} - {%- endfor %} - {%- elif content is none or content is undefined %} - {{- '' }} - {%- else %} - {{- raise_exception('Unexpected content type.') }} - {%- endif %} -{%- endmacro %} -{%- if not messages %} - {{- raise_exception('No messages provided.') }} -{%- endif %} -{%- if tools and tools is iterable and tools is not mapping %} - {{- '<|im_start|>system\n' }} - {{- "# Tools\n\nYou have access to the following functions:\n\n" }} - {%- for tool in tools %} - {{- "\n" }} - {{- tool | tojson }} - {%- endfor %} - {{- "\n" }} - {{- '\n\nIf you choose to call a function ONLY reply in the following format with NO suffix:\n\n\n\n\nvalue_1\n\n\nThis is the value for the second parameter\nthat can span\nmultiple lines\n\n\n\n\n\nReminder:\n- Function calls MUST follow the specified format: an inner block must be nested within XML tags\n- Required parameters MUST be specified\n- You may provide optional reasoning for your function call in natural language BEFORE the function call, but NOT after\n- If there is no function call available, answer the question like normal with your current knowledge and do not tell the user about function calls\n' }} - {%- if messages[0].role == 'system' %} - {%- set content = render_content(messages[0].content, false, true)|trim %} - {%- if content %} - {{- '\n\n' + content }} - {%- endif %} - {%- endif %} - {{- '<|im_end|>\n' }} -{%- else %} - {%- if messages[0].role == 'system' %} - {%- set content = render_content(messages[0].content, false, true)|trim %} - {{- '<|im_start|>system\n' + 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" %} - {%- set content = render_content(message.content, false)|trim %} - {%- if not(content.startswith('') and content.endswith('')) %} - {%- set ns.multi_step_tool = false %} - {%- set ns.last_query_index = index %} - {%- endif %} - {%- endif %} -{%- endfor %} -{%- if ns.multi_step_tool %} - {{- raise_exception('No user query found in messages.') }} -{%- endif %} -{%- for message in messages %} - {%- set content = render_content(message.content, true)|trim %} - {%- if message.role == "system" %} - {%- if not loop.first %} - {{- raise_exception('System message must be at the beginning.') }} - {%- endif %} - {%- elif message.role == "user" %} - {{- '<|im_start|>' + message.role + '\n' + content + '<|im_end|>' + '\n' }} - {%- elif message.role == "assistant" %} - {%- set reasoning_content = '' %} - {%- if message.reasoning_content is string %} - {%- set reasoning_content = message.reasoning_content %} - {%- else %} - {%- if '' in content %} - {%- set reasoning_content = content.split('')[0].rstrip('\n').split('')[-1].lstrip('\n') %} - {%- set content = content.split('')[-1].lstrip('\n') %} - {%- endif %} - {%- endif %} - {%- set reasoning_content = reasoning_content|trim %} - {%- if loop.index0 > ns.last_query_index %} - {{- '<|im_start|>' + message.role + '\n\n' + reasoning_content + '\n\n\n' + content }} - {%- else %} - {{- '<|im_start|>' + message.role + '\n' + content }} - {%- endif %} - {%- if message.tool_calls and message.tool_calls is iterable and message.tool_calls is not mapping %} - {%- for tool_call in message.tool_calls %} - {%- if tool_call.function is defined %} - {%- set tool_call = tool_call.function %} - {%- endif %} - {%- if loop.first %} - {%- if content|trim %} - {{- '\n\n\n\n' }} - {%- else %} - {{- '\n\n' }} - {%- endif %} - {%- else %} - {{- '\n\n\n' }} - {%- endif %} - {%- if tool_call.arguments is defined %} - {%- for args_name, args_value in tool_call.arguments|items %} - {{- '\n' }} - {%- set args_value = args_value | tojson | safe if args_value is mapping or (args_value is sequence and args_value is not string) else args_value | string %} - {{- args_value }} - {{- '\n\n' }} - {%- endfor %} - {%- endif %} - {{- '\n' }} - {%- endfor %} - {%- endif %} - {{- '<|im_end|>\n' }} - {%- elif message.role == "tool" %} - {%- if loop.previtem and loop.previtem.role != "tool" %} - {{- '<|im_start|>user' }} - {%- endif %} - {{- '\n\n' }} - {{- content }} - {{- '\n' }} - {%- if not loop.last and loop.nextitem.role != "tool" %} - {{- '<|im_end|>\n' }} - {%- elif loop.last %} - {{- '<|im_end|>\n' }} - {%- endif %} - {%- else %} - {{- raise_exception('Unexpected message role.') }} - {%- endif %} -{%- endfor %} -{%- if add_generation_prompt %} - {{- '<|im_start|>assistant\n' }} - {%- if enable_thinking is defined and enable_thinking is true %} - {{- '\n' }} - {%- else %} - {{- '\n\n\n\n' }} - {%- endif %} -{%- endif %} \ No newline at end of file diff --git a/substitutivity_original_Estonian/Qwen3.5-2B-Base_substitutivity_splits_original_features_train_substitutivity_splits_original_features_test2/checkpoint-2075/tokenizer.json b/substitutivity_original_Estonian/Qwen3.5-2B-Base_substitutivity_splits_original_features_train_substitutivity_splits_original_features_test2/checkpoint-2075/tokenizer.json deleted file mode 100644 index 67741b04f23bfdb46501f748ce27865ec82eccfb..0000000000000000000000000000000000000000 --- a/substitutivity_original_Estonian/Qwen3.5-2B-Base_substitutivity_splits_original_features_train_substitutivity_splits_original_features_test2/checkpoint-2075/tokenizer.json +++ /dev/null @@ -1,3 +0,0 @@ -version https://git-lfs.github.com/spec/v1 -oid sha256:87a7830d63fcf43bf241c3c5242e96e62dd3fdc29224ca26fed8ea333db72de4 -size 19989343 diff --git a/substitutivity_original_Estonian/Qwen3.5-2B-Base_substitutivity_splits_original_features_train_substitutivity_splits_original_features_test2/checkpoint-2075/tokenizer_config.json b/substitutivity_original_Estonian/Qwen3.5-2B-Base_substitutivity_splits_original_features_train_substitutivity_splits_original_features_test2/checkpoint-2075/tokenizer_config.json deleted file mode 100644 index b4a37b2a6fd3ab3317cd7bac72855be1a843b2bb..0000000000000000000000000000000000000000 --- a/substitutivity_original_Estonian/Qwen3.5-2B-Base_substitutivity_splits_original_features_train_substitutivity_splits_original_features_test2/checkpoint-2075/tokenizer_config.json +++ /dev/null @@ -1,31 +0,0 @@ -{ - "add_prefix_space": false, - "audio_bos_token": "<|audio_start|>", - "audio_eos_token": "<|audio_end|>", - "audio_token": "<|audio_pad|>", - "backend": "tokenizers", - "bos_token": null, - "clean_up_tokenization_spaces": false, - "eos_token": "<|endoftext|>", - "errors": "replace", - "image_token": "<|image_pad|>", - "is_local": false, - "model_max_length": 262144, - "model_specific_special_tokens": { - "audio_bos_token": "<|audio_start|>", - "audio_eos_token": "<|audio_end|>", - "audio_token": "<|audio_pad|>", - "image_token": "<|image_pad|>", - "video_token": "<|video_pad|>", - "vision_bos_token": "<|vision_start|>", - "vision_eos_token": "<|vision_end|>" - }, - "pad_token": "<|endoftext|>", - "pretokenize_regex": "(?i:'s|'t|'re|'ve|'m|'ll|'d)|[^\\r\\n\\p{L}\\p{N}]?[\\p{L}\\p{M}]+|\\p{N}| ?[^\\s\\p{L}\\p{M}\\p{N}]+[\\r\\n]*|\\s*[\\r\\n]+|\\s+(?!\\S)|\\s+", - "split_special_tokens": false, - "tokenizer_class": "TokenizersBackend", - "unk_token": null, - "video_token": "<|video_pad|>", - "vision_bos_token": "<|vision_start|>", - "vision_eos_token": "<|vision_end|>" -} diff --git a/substitutivity_original_Estonian/Qwen3.5-2B-Base_substitutivity_splits_original_features_train_substitutivity_splits_original_features_test2/checkpoint-2075/trainer_state.json b/substitutivity_original_Estonian/Qwen3.5-2B-Base_substitutivity_splits_original_features_train_substitutivity_splits_original_features_test2/checkpoint-2075/trainer_state.json deleted file mode 100644 index 922fd6506727b5cfa553c590a0d9c1bb9a40c989..0000000000000000000000000000000000000000 --- a/substitutivity_original_Estonian/Qwen3.5-2B-Base_substitutivity_splits_original_features_train_substitutivity_splits_original_features_test2/checkpoint-2075/trainer_state.json +++ /dev/null @@ -1,554 +0,0 @@ -{ - "best_global_step": null, - "best_metric": null, - "best_model_checkpoint": null, - "epoch": 5.0, - "eval_steps": 500, - "global_step": 2075, - "is_hyper_param_search": false, - "is_local_process_zero": true, - "is_world_process_zero": true, - "log_history": [ - { - "entropy": 2.2933380508422854, - "epoch": 0.12062726176115803, - "grad_norm": 2.8075337409973145, - "learning_rate": 1.5531017374629602e-05, - "loss": 2.1429234313964844, - "mean_token_accuracy": 0.5779684272408485, - "num_tokens": 121428.0, - "step": 50 - }, - { - "entropy": 1.1315062403678895, - "epoch": 0.24125452352231605, - "grad_norm": 1.9415371417999268, - "learning_rate": 3.1378994287516954e-05, - "loss": 1.0461698913574218, - "mean_token_accuracy": 0.7166991305351257, - "num_tokens": 243966.0, - "step": 100 - }, - { - "entropy": 0.9537658470869065, - "epoch": 0.3618817852834741, - "grad_norm": 1.8353157043457031, - "learning_rate": 4.72269712004043e-05, - "loss": 0.8774301147460938, - "mean_token_accuracy": 0.7506789201498032, - "num_tokens": 366649.0, - "step": 150 - }, - { - "entropy": 0.8955666983127594, - "epoch": 0.4825090470446321, - "grad_norm": 1.5419880151748657, - "learning_rate": 6.307494811329165e-05, - "loss": 0.8234302520751953, - "mean_token_accuracy": 0.7616239994764328, - "num_tokens": 488367.0, - "step": 200 - }, - { - "entropy": 0.8689469587802887, - "epoch": 0.6031363088057901, - "grad_norm": 1.5661484003067017, - "learning_rate": 7.8922925026179e-05, - "loss": 0.7961890411376953, - "mean_token_accuracy": 0.767715499997139, - "num_tokens": 603131.0, - "step": 250 - }, - { - "entropy": 0.8425652319192887, - "epoch": 0.7237635705669482, - "grad_norm": 1.5694905519485474, - "learning_rate": 9.477090193906636e-05, - "loss": 0.7718594360351563, - "mean_token_accuracy": 0.7736722755432129, - "num_tokens": 726412.0, - "step": 300 - }, - { - "entropy": 0.843294187784195, - "epoch": 0.8443908323281062, - "grad_norm": 1.1118028163909912, - "learning_rate": 0.00011061887885195371, - "loss": 0.7801297760009765, - "mean_token_accuracy": 0.7714760398864746, - "num_tokens": 846516.0, - "step": 350 - }, - { - "entropy": 0.8213214933872223, - "epoch": 0.9650180940892642, - "grad_norm": 0.9202918410301208, - "learning_rate": 0.00012646685576484106, - "loss": 0.7514453887939453, - "mean_token_accuracy": 0.7774044269323349, - "num_tokens": 958978.0, - "step": 400 - }, - { - "epoch": 1.0, - "eval_entropy": 0.818668900915746, - "eval_mean_token_accuracy": 0.7533496866065464, - "eval_not_syn_loss": 0.7798981070518494, - "eval_not_syn_runtime": 53.8744, - "eval_not_syn_samples_per_second": 26.358, - "eval_not_syn_steps_per_second": 3.304, - "eval_num_tokens": 993806.0, - "step": 415 - }, - { - "epoch": 1.0, - "eval_entropy": 0.7627468125873738, - "eval_mean_token_accuracy": 0.8024111574285486, - "eval_num_tokens": 993806.0, - "eval_syn_loss": 0.7218216061592102, - "eval_syn_runtime": 55.3239, - "eval_syn_samples_per_second": 25.667, - "eval_syn_steps_per_second": 3.217, - "step": 415 - }, - { - "entropy": 0.7802230679627621, - "epoch": 1.0844390832328106, - "grad_norm": 1.0145268440246582, - "learning_rate": 0.00013151131541121804, - "loss": 0.7111608123779297, - "mean_token_accuracy": 0.7845737542768921, - "num_tokens": 1080301.0, - "step": 450 - }, - { - "entropy": 0.7558828103542328, - "epoch": 1.2050663449939687, - "grad_norm": 1.0477591753005981, - "learning_rate": 0.0001313741160170038, - "loss": 0.6933037567138672, - "mean_token_accuracy": 0.7889221370220184, - "num_tokens": 1201215.0, - "step": 500 - }, - { - "entropy": 0.7593123841285706, - "epoch": 1.3256936067551266, - "grad_norm": 0.899161159992218, - "learning_rate": 0.00013112089697387701, - "loss": 0.6962493133544921, - "mean_token_accuracy": 0.7877359414100646, - "num_tokens": 1322895.0, - "step": 550 - }, - { - "entropy": 0.7460986822843552, - "epoch": 1.4463208685162847, - "grad_norm": 1.1061116456985474, - "learning_rate": 0.00013075210608885215, - "loss": 0.6817412567138672, - "mean_token_accuracy": 0.7904874366521836, - "num_tokens": 1436749.0, - "step": 600 - }, - { - "entropy": 0.735516265630722, - "epoch": 1.5669481302774426, - "grad_norm": 0.9925405979156494, - "learning_rate": 0.00013026839555278796, - "loss": 0.6690705871582031, - "mean_token_accuracy": 0.7930250608921051, - "num_tokens": 1555400.0, - "step": 650 - }, - { - "entropy": 0.7302519899606704, - "epoch": 1.6875753920386007, - "grad_norm": 0.800025463104248, - "learning_rate": 0.0001296706207870159, - "loss": 0.6714833831787109, - "mean_token_accuracy": 0.7946307575702667, - "num_tokens": 1676892.0, - "step": 700 - }, - { - "entropy": 0.7321263110637665, - "epoch": 1.8082026537997589, - "grad_norm": 1.0500682592391968, - "learning_rate": 0.00012895983893056423, - "loss": 0.6672868347167968, - "mean_token_accuracy": 0.7947645837068558, - "num_tokens": 1802689.0, - "step": 750 - }, - { - "entropy": 0.7075755137205124, - "epoch": 1.9288299155609168, - "grad_norm": 0.9248411655426025, - "learning_rate": 0.00012813730697065298, - "loss": 0.6466568756103516, - "mean_token_accuracy": 0.798747689127922, - "num_tokens": 1920804.0, - "step": 800 - }, - { - "epoch": 2.0, - "eval_entropy": 0.7426781811740961, - "eval_mean_token_accuracy": 0.7693507239389955, - "eval_not_syn_loss": 0.7139139175415039, - "eval_not_syn_runtime": 54.4286, - "eval_not_syn_samples_per_second": 26.089, - "eval_not_syn_steps_per_second": 3.27, - "eval_num_tokens": 1987612.0, - "step": 830 - }, - { - "epoch": 2.0, - "eval_entropy": 0.6937930352902144, - "eval_mean_token_accuracy": 0.8180048633157537, - "eval_num_tokens": 1987612.0, - "eval_syn_loss": 0.6489660143852234, - "eval_syn_runtime": 56.1525, - "eval_syn_samples_per_second": 25.288, - "eval_syn_steps_per_second": 3.17, - "step": 830 - }, - { - "entropy": 0.6853217440422135, - "epoch": 2.0482509047044632, - "grad_norm": 0.6932648420333862, - "learning_rate": 0.00012720447951976603, - "loss": 0.6212613296508789, - "mean_token_accuracy": 0.8047762579388089, - "num_tokens": 2039660.0, - "step": 850 - }, - { - "entropy": 0.6459032052755356, - "epoch": 2.168878166465621, - "grad_norm": 1.0688549280166626, - "learning_rate": 0.00012616300624323124, - "loss": 0.5740525817871094, - "mean_token_accuracy": 0.8148404061794281, - "num_tokens": 2161772.0, - "step": 900 - }, - { - "entropy": 0.6463843190670013, - "epoch": 2.2895054282267795, - "grad_norm": 0.847271203994751, - "learning_rate": 0.0001250147289418581, - "loss": 0.577717170715332, - "mean_token_accuracy": 0.814143658876419, - "num_tokens": 2280833.0, - "step": 950 - }, - { - "entropy": 0.6461306965351105, - "epoch": 2.4101326899879374, - "grad_norm": 0.9998957514762878, - "learning_rate": 0.00012376167829479217, - "loss": 0.5813011932373047, - "mean_token_accuracy": 0.8144401633739471, - "num_tokens": 2398206.0, - "step": 1000 - }, - { - "entropy": 0.640914659500122, - "epoch": 2.5307599517490953, - "grad_norm": 0.8692949414253235, - "learning_rate": 0.0001224060702683459, - "loss": 0.5755012512207032, - "mean_token_accuracy": 0.8168018108606339, - "num_tokens": 2518888.0, - "step": 1050 - }, - { - "entropy": 0.6421274638175964, - "epoch": 2.651387213510253, - "grad_norm": 0.8920196294784546, - "learning_rate": 0.00012095030219715738, - "loss": 0.5785483932495117, - "mean_token_accuracy": 0.8137972563505173, - "num_tokens": 2639763.0, - "step": 1100 - }, - { - "entropy": 0.6466812592744827, - "epoch": 2.772014475271411, - "grad_norm": 0.6876946687698364, - "learning_rate": 0.00011939694854460696, - "loss": 0.5816070938110351, - "mean_token_accuracy": 0.8145407867431641, - "num_tokens": 2755540.0, - "step": 1150 - }, - { - "entropy": 0.6378493067622185, - "epoch": 2.8926417370325694, - "grad_norm": 0.7532753944396973, - "learning_rate": 0.00011774875634998911, - "loss": 0.5754346084594727, - "mean_token_accuracy": 0.8148861181735992, - "num_tokens": 2874118.0, - "step": 1200 - }, - { - "epoch": 3.0, - "eval_entropy": 0.6912222209941136, - "eval_mean_token_accuracy": 0.7868467750174276, - "eval_not_syn_loss": 0.6782422661781311, - "eval_not_syn_runtime": 53.2627, - "eval_not_syn_samples_per_second": 26.66, - "eval_not_syn_steps_per_second": 3.342, - "eval_num_tokens": 2981418.0, - "step": 1245 - }, - { - "epoch": 3.0, - "eval_entropy": 0.6467212934172555, - "eval_mean_token_accuracy": 0.8215091817834405, - "eval_num_tokens": 2981418.0, - "eval_syn_loss": 0.6301999092102051, - "eval_syn_runtime": 55.0115, - "eval_syn_samples_per_second": 25.813, - "eval_syn_steps_per_second": 3.236, - "step": 1245 - }, - { - "entropy": 0.6316030452949832, - "epoch": 3.012062726176116, - "grad_norm": 0.7761592864990234, - "learning_rate": 0.00011600864037049124, - "loss": 0.5697880935668945, - "mean_token_accuracy": 0.8159356400220081, - "num_tokens": 2994602.0, - "step": 1250 - }, - { - "entropy": 0.5467158386111259, - "epoch": 3.132689987937274, - "grad_norm": 0.8169087767601013, - "learning_rate": 0.00011417967792657078, - "loss": 0.47844127655029295, - "mean_token_accuracy": 0.8378052765130997, - "num_tokens": 3121094.0, - "step": 1300 - }, - { - "entropy": 0.5454566389322281, - "epoch": 3.2533172496984317, - "grad_norm": 0.8219636678695679, - "learning_rate": 0.0001122651034598459, - "loss": 0.4839338302612305, - "mean_token_accuracy": 0.8385071897506714, - "num_tokens": 3248038.0, - "step": 1350 - }, - { - "entropy": 0.5598723018169403, - "epoch": 3.37394451145959, - "grad_norm": 0.9956360459327698, - "learning_rate": 0.00011026830281312423, - "loss": 0.492743034362793, - "mean_token_accuracy": 0.8357370454072952, - "num_tokens": 3363487.0, - "step": 1400 - }, - { - "entropy": 0.5479365903139114, - "epoch": 3.494571773220748, - "grad_norm": 0.8586823344230652, - "learning_rate": 0.00010819280724268524, - "loss": 0.4810820770263672, - "mean_token_accuracy": 0.837928666472435, - "num_tokens": 3480513.0, - "step": 1450 - }, - { - "entropy": 0.5619524419307709, - "epoch": 3.615199034981906, - "grad_norm": 1.0894954204559326, - "learning_rate": 0.0001060422871734048, - "loss": 0.4974747085571289, - "mean_token_accuracy": 0.8349784809350967, - "num_tokens": 3597183.0, - "step": 1500 - }, - { - "entropy": 0.5588027596473694, - "epoch": 3.7358262967430638, - "grad_norm": 0.9905257225036621, - "learning_rate": 0.00010382054570776654, - "loss": 0.49687202453613283, - "mean_token_accuracy": 0.834458841085434, - "num_tokens": 3714173.0, - "step": 1550 - }, - { - "entropy": 0.5654991599917412, - "epoch": 3.856453558504222, - "grad_norm": 1.0116610527038574, - "learning_rate": 0.00010153151190023791, - "loss": 0.5016670989990234, - "mean_token_accuracy": 0.8326652735471726, - "num_tokens": 3832652.0, - "step": 1600 - }, - { - "entropy": 0.5605382058024406, - "epoch": 3.97708082026538, - "grad_norm": 0.8011100888252258, - "learning_rate": 9.917923380890624e-05, - "loss": 0.49551734924316404, - "mean_token_accuracy": 0.8343535929918289, - "num_tokens": 3955167.0, - "step": 1650 - }, - { - "epoch": 4.0, - "eval_entropy": 0.5819476859623127, - "eval_mean_token_accuracy": 0.7963069781158747, - "eval_not_syn_loss": 0.6924306750297546, - "eval_not_syn_runtime": 54.027, - "eval_not_syn_samples_per_second": 26.283, - "eval_not_syn_steps_per_second": 3.295, - "eval_num_tokens": 3975224.0, - "step": 1660 - }, - { - "epoch": 4.0, - "eval_entropy": 0.545470636212424, - "eval_mean_token_accuracy": 0.8127643666910321, - "eval_num_tokens": 3975224.0, - "eval_syn_loss": 0.649576723575592, - "eval_syn_runtime": 55.2296, - "eval_syn_samples_per_second": 25.711, - "eval_syn_steps_per_second": 3.223, - "step": 1660 - }, - { - "entropy": 0.4920780294471317, - "epoch": 4.0965018094089265, - "grad_norm": 0.8362822532653809, - "learning_rate": 9.676787133666148e-05, - "loss": 0.4147996139526367, - "mean_token_accuracy": 0.8581844628459275, - "num_tokens": 4072361.0, - "step": 1700 - }, - { - "entropy": 0.4675139603018761, - "epoch": 4.217129071170085, - "grad_norm": 0.9920316338539124, - "learning_rate": 9.430168887458645e-05, - "loss": 0.3863528823852539, - "mean_token_accuracy": 0.865827117562294, - "num_tokens": 4194978.0, - "step": 1750 - }, - { - "entropy": 0.47163557589054106, - "epoch": 4.337756332931242, - "grad_norm": 1.0789282321929932, - "learning_rate": 9.178504776056418e-05, - "loss": 0.390842170715332, - "mean_token_accuracy": 0.8634983265399933, - "num_tokens": 4315548.0, - "step": 1800 - }, - { - "entropy": 0.47117673605680466, - "epoch": 4.458383594692401, - "grad_norm": 1.0369561910629272, - "learning_rate": 8.922239856643896e-05, - "loss": 0.39387256622314454, - "mean_token_accuracy": 0.8625931572914124, - "num_tokens": 4431469.0, - "step": 1850 - }, - { - "entropy": 0.479859294295311, - "epoch": 4.579010856453559, - "grad_norm": 1.0827305316925049, - "learning_rate": 8.661827322737095e-05, - "loss": 0.40195709228515625, - "mean_token_accuracy": 0.8612629306316376, - "num_tokens": 4551803.0, - "step": 1900 - }, - { - "entropy": 0.4729380664229393, - "epoch": 4.699638118214716, - "grad_norm": 1.0114246606826782, - "learning_rate": 8.397727702730326e-05, - "loss": 0.39648338317871096, - "mean_token_accuracy": 0.8610477095842362, - "num_tokens": 4674468.0, - "step": 1950 - }, - { - "entropy": 0.4903118047118187, - "epoch": 4.820265379975875, - "grad_norm": 1.0115467309951782, - "learning_rate": 8.130408045471496e-05, - "loss": 0.4067733001708984, - "mean_token_accuracy": 0.8597196912765503, - "num_tokens": 4788027.0, - "step": 2000 - }, - { - "entropy": 0.48099748432636263, - "epoch": 4.940892641737032, - "grad_norm": 1.0201005935668945, - "learning_rate": 7.860341094306242e-05, - "loss": 0.4030665588378906, - "mean_token_accuracy": 0.8603792876005173, - "num_tokens": 4907561.0, - "step": 2050 - }, - { - "epoch": 5.0, - "eval_entropy": 0.5492983292997553, - "eval_mean_token_accuracy": 0.796870938847574, - "eval_not_syn_loss": 0.7129728198051453, - "eval_not_syn_runtime": 53.9575, - "eval_not_syn_samples_per_second": 26.317, - "eval_not_syn_steps_per_second": 3.299, - "eval_num_tokens": 4969030.0, - "step": 2075 - }, - { - "epoch": 5.0, - "eval_entropy": 0.513823946037989, - "eval_mean_token_accuracy": 0.8174765464964877, - "eval_num_tokens": 4969030.0, - "eval_syn_loss": 0.6699439883232117, - "eval_syn_runtime": 56.0033, - "eval_syn_samples_per_second": 25.356, - "eval_syn_steps_per_second": 3.178, - "step": 2075 - } - ], - "logging_steps": 50, - "max_steps": 4150, - "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": 1.0256255729624371e+17, - "train_batch_size": 8, - "trial_name": null, - "trial_params": null -} diff --git a/substitutivity_original_Estonian/Qwen3.5-2B-Base_substitutivity_splits_original_features_train_substitutivity_splits_original_features_test2/checkpoint-2075/training_args.bin b/substitutivity_original_Estonian/Qwen3.5-2B-Base_substitutivity_splits_original_features_train_substitutivity_splits_original_features_test2/checkpoint-2075/training_args.bin deleted file mode 100644 index f3d8098f65400413dd700f32ada5649470871017..0000000000000000000000000000000000000000 --- a/substitutivity_original_Estonian/Qwen3.5-2B-Base_substitutivity_splits_original_features_train_substitutivity_splits_original_features_test2/checkpoint-2075/training_args.bin +++ /dev/null @@ -1,3 +0,0 @@ -version https://git-lfs.github.com/spec/v1 -oid sha256:617af0f41414927baaaf158949585bb75eb4b05ff78baa5216c860ef0b431009 -size 6033 diff --git a/substitutivity_original_Estonian/Qwen3.5-2B-Base_substitutivity_splits_original_features_train_substitutivity_splits_original_features_test2/checkpoint-2490/README.md b/substitutivity_original_Estonian/Qwen3.5-2B-Base_substitutivity_splits_original_features_train_substitutivity_splits_original_features_test2/checkpoint-2490/README.md deleted file mode 100644 index ece204d005f7bf161d0e448efdd56fd297b7c8a2..0000000000000000000000000000000000000000 --- a/substitutivity_original_Estonian/Qwen3.5-2B-Base_substitutivity_splits_original_features_train_substitutivity_splits_original_features_test2/checkpoint-2490/README.md +++ /dev/null @@ -1,209 +0,0 @@ ---- -base_model: Qwen/Qwen3.5-2B-Base -library_name: peft -pipeline_tag: text-generation -tags: -- base_model:adapter:Qwen/Qwen3.5-2B-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/substitutivity_original_Estonian/Qwen3.5-2B-Base_substitutivity_splits_original_features_train_substitutivity_splits_original_features_test2/checkpoint-2490/adapter_config.json b/substitutivity_original_Estonian/Qwen3.5-2B-Base_substitutivity_splits_original_features_train_substitutivity_splits_original_features_test2/checkpoint-2490/adapter_config.json deleted file mode 100644 index 6a095bb3bfe249ce8e15df6ae999e88fa1e17162..0000000000000000000000000000000000000000 --- a/substitutivity_original_Estonian/Qwen3.5-2B-Base_substitutivity_splits_original_features_train_substitutivity_splits_original_features_test2/checkpoint-2490/adapter_config.json +++ /dev/null @@ -1,46 +0,0 @@ -{ - "alora_invocation_tokens": null, - "alpha_pattern": {}, - "arrow_config": null, - "auto_mapping": null, - "base_model_name_or_path": "Qwen/Qwen3.5-2B-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": 0.05731002492184622, - "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": [ - "o_proj", - "q_proj", - "k_proj", - "down_proj", - "up_proj", - "v_proj", - "gate_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/substitutivity_original_Estonian/Qwen3.5-2B-Base_substitutivity_splits_original_features_train_substitutivity_splits_original_features_test2/checkpoint-2490/adapter_model.safetensors b/substitutivity_original_Estonian/Qwen3.5-2B-Base_substitutivity_splits_original_features_train_substitutivity_splits_original_features_test2/checkpoint-2490/adapter_model.safetensors deleted file mode 100644 index 95049cfe1ef78cc9b686ccca1d1784e3d144ff65..0000000000000000000000000000000000000000 --- a/substitutivity_original_Estonian/Qwen3.5-2B-Base_substitutivity_splits_original_features_train_substitutivity_splits_original_features_test2/checkpoint-2490/adapter_model.safetensors +++ /dev/null @@ -1,3 +0,0 @@ -version https://git-lfs.github.com/spec/v1 -oid sha256:24a3d7876617a41b54c4ac29bb609c0fe66a708065fac0b99af76c31c548b04e -size 174613392 diff --git a/substitutivity_original_Estonian/Qwen3.5-2B-Base_substitutivity_splits_original_features_train_substitutivity_splits_original_features_test2/checkpoint-2490/chat_template.jinja b/substitutivity_original_Estonian/Qwen3.5-2B-Base_substitutivity_splits_original_features_train_substitutivity_splits_original_features_test2/checkpoint-2490/chat_template.jinja deleted file mode 100644 index 0ef09f214eaa6d9bca297988afc1454b5827b2c7..0000000000000000000000000000000000000000 --- a/substitutivity_original_Estonian/Qwen3.5-2B-Base_substitutivity_splits_original_features_train_substitutivity_splits_original_features_test2/checkpoint-2490/chat_template.jinja +++ /dev/null @@ -1,154 +0,0 @@ -{%- set image_count = namespace(value=0) %} -{%- set video_count = namespace(value=0) %} -{%- macro render_content(content, do_vision_count, is_system_content=false) %} - {%- if content is string %} - {{- content }} - {%- elif content is iterable and content is not mapping %} - {%- for item in content %} - {%- if 'image' in item or 'image_url' in item or item.type == 'image' %} - {%- if is_system_content %} - {{- raise_exception('System message cannot contain images.') }} - {%- endif %} - {%- if do_vision_count %} - {%- set image_count.value = image_count.value + 1 %} - {%- endif %} - {%- if add_vision_id %} - {{- 'Picture ' ~ image_count.value ~ ': ' }} - {%- endif %} - {{- '<|vision_start|><|image_pad|><|vision_end|>' }} - {%- elif 'video' in item or item.type == 'video' %} - {%- if is_system_content %} - {{- raise_exception('System message cannot contain videos.') }} - {%- endif %} - {%- if do_vision_count %} - {%- set video_count.value = video_count.value + 1 %} - {%- endif %} - {%- if add_vision_id %} - {{- 'Video ' ~ video_count.value ~ ': ' }} - {%- endif %} - {{- '<|vision_start|><|video_pad|><|vision_end|>' }} - {%- elif 'text' in item %} - {{- item.text }} - {%- else %} - {{- raise_exception('Unexpected item type in content.') }} - {%- endif %} - {%- endfor %} - {%- elif content is none or content is undefined %} - {{- '' }} - {%- else %} - {{- raise_exception('Unexpected content type.') }} - {%- endif %} -{%- endmacro %} -{%- if not messages %} - {{- raise_exception('No messages provided.') }} -{%- endif %} -{%- if tools and tools is iterable and tools is not mapping %} - {{- '<|im_start|>system\n' }} - {{- "# Tools\n\nYou have access to the following functions:\n\n" }} - {%- for tool in tools %} - {{- "\n" }} - {{- tool | tojson }} - {%- endfor %} - {{- "\n" }} - {{- '\n\nIf you choose to call a function ONLY reply in the following format with NO suffix:\n\n\n\n\nvalue_1\n\n\nThis is the value for the second parameter\nthat can span\nmultiple lines\n\n\n\n\n\nReminder:\n- Function calls MUST follow the specified format: an inner block must be nested within XML tags\n- Required parameters MUST be specified\n- You may provide optional reasoning for your function call in natural language BEFORE the function call, but NOT after\n- If there is no function call available, answer the question like normal with your current knowledge and do not tell the user about function calls\n' }} - {%- if messages[0].role == 'system' %} - {%- set content = render_content(messages[0].content, false, true)|trim %} - {%- if content %} - {{- '\n\n' + content }} - {%- endif %} - {%- endif %} - {{- '<|im_end|>\n' }} -{%- else %} - {%- if messages[0].role == 'system' %} - {%- set content = render_content(messages[0].content, false, true)|trim %} - {{- '<|im_start|>system\n' + 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" %} - {%- set content = render_content(message.content, false)|trim %} - {%- if not(content.startswith('') and content.endswith('')) %} - {%- set ns.multi_step_tool = false %} - {%- set ns.last_query_index = index %} - {%- endif %} - {%- endif %} -{%- endfor %} -{%- if ns.multi_step_tool %} - {{- raise_exception('No user query found in messages.') }} -{%- endif %} -{%- for message in messages %} - {%- set content = render_content(message.content, true)|trim %} - {%- if message.role == "system" %} - {%- if not loop.first %} - {{- raise_exception('System message must be at the beginning.') }} - {%- endif %} - {%- elif message.role == "user" %} - {{- '<|im_start|>' + message.role + '\n' + content + '<|im_end|>' + '\n' }} - {%- elif message.role == "assistant" %} - {%- set reasoning_content = '' %} - {%- if message.reasoning_content is string %} - {%- set reasoning_content = message.reasoning_content %} - {%- else %} - {%- if '' in content %} - {%- set reasoning_content = content.split('')[0].rstrip('\n').split('')[-1].lstrip('\n') %} - {%- set content = content.split('')[-1].lstrip('\n') %} - {%- endif %} - {%- endif %} - {%- set reasoning_content = reasoning_content|trim %} - {%- if loop.index0 > ns.last_query_index %} - {{- '<|im_start|>' + message.role + '\n\n' + reasoning_content + '\n\n\n' + content }} - {%- else %} - {{- '<|im_start|>' + message.role + '\n' + content }} - {%- endif %} - {%- if message.tool_calls and message.tool_calls is iterable and message.tool_calls is not mapping %} - {%- for tool_call in message.tool_calls %} - {%- if tool_call.function is defined %} - {%- set tool_call = tool_call.function %} - {%- endif %} - {%- if loop.first %} - {%- if content|trim %} - {{- '\n\n\n\n' }} - {%- else %} - {{- '\n\n' }} - {%- endif %} - {%- else %} - {{- '\n\n\n' }} - {%- endif %} - {%- if tool_call.arguments is defined %} - {%- for args_name, args_value in tool_call.arguments|items %} - {{- '\n' }} - {%- set args_value = args_value | tojson | safe if args_value is mapping or (args_value is sequence and args_value is not string) else args_value | string %} - {{- args_value }} - {{- '\n\n' }} - {%- endfor %} - {%- endif %} - {{- '\n' }} - {%- endfor %} - {%- endif %} - {{- '<|im_end|>\n' }} - {%- elif message.role == "tool" %} - {%- if loop.previtem and loop.previtem.role != "tool" %} - {{- '<|im_start|>user' }} - {%- endif %} - {{- '\n\n' }} - {{- content }} - {{- '\n' }} - {%- if not loop.last and loop.nextitem.role != "tool" %} - {{- '<|im_end|>\n' }} - {%- elif loop.last %} - {{- '<|im_end|>\n' }} - {%- endif %} - {%- else %} - {{- raise_exception('Unexpected message role.') }} - {%- endif %} -{%- endfor %} -{%- if add_generation_prompt %} - {{- '<|im_start|>assistant\n' }} - {%- if enable_thinking is defined and enable_thinking is true %} - {{- '\n' }} - {%- else %} - {{- '\n\n\n\n' }} - {%- endif %} -{%- endif %} \ No newline at end of file diff --git a/substitutivity_original_Estonian/Qwen3.5-2B-Base_substitutivity_splits_original_features_train_substitutivity_splits_original_features_test2/checkpoint-2490/tokenizer.json b/substitutivity_original_Estonian/Qwen3.5-2B-Base_substitutivity_splits_original_features_train_substitutivity_splits_original_features_test2/checkpoint-2490/tokenizer.json deleted file mode 100644 index 67741b04f23bfdb46501f748ce27865ec82eccfb..0000000000000000000000000000000000000000 --- a/substitutivity_original_Estonian/Qwen3.5-2B-Base_substitutivity_splits_original_features_train_substitutivity_splits_original_features_test2/checkpoint-2490/tokenizer.json +++ /dev/null @@ -1,3 +0,0 @@ -version https://git-lfs.github.com/spec/v1 -oid sha256:87a7830d63fcf43bf241c3c5242e96e62dd3fdc29224ca26fed8ea333db72de4 -size 19989343 diff --git a/substitutivity_original_Estonian/Qwen3.5-2B-Base_substitutivity_splits_original_features_train_substitutivity_splits_original_features_test2/checkpoint-2490/tokenizer_config.json b/substitutivity_original_Estonian/Qwen3.5-2B-Base_substitutivity_splits_original_features_train_substitutivity_splits_original_features_test2/checkpoint-2490/tokenizer_config.json deleted file mode 100644 index b4a37b2a6fd3ab3317cd7bac72855be1a843b2bb..0000000000000000000000000000000000000000 --- a/substitutivity_original_Estonian/Qwen3.5-2B-Base_substitutivity_splits_original_features_train_substitutivity_splits_original_features_test2/checkpoint-2490/tokenizer_config.json +++ /dev/null @@ -1,31 +0,0 @@ -{ - "add_prefix_space": false, - "audio_bos_token": "<|audio_start|>", - "audio_eos_token": "<|audio_end|>", - "audio_token": "<|audio_pad|>", - "backend": "tokenizers", - "bos_token": null, - "clean_up_tokenization_spaces": false, - "eos_token": "<|endoftext|>", - "errors": "replace", - "image_token": "<|image_pad|>", - "is_local": false, - "model_max_length": 262144, - "model_specific_special_tokens": { - "audio_bos_token": "<|audio_start|>", - "audio_eos_token": "<|audio_end|>", - "audio_token": "<|audio_pad|>", - "image_token": "<|image_pad|>", - "video_token": "<|video_pad|>", - "vision_bos_token": "<|vision_start|>", - "vision_eos_token": "<|vision_end|>" - }, - "pad_token": "<|endoftext|>", - "pretokenize_regex": "(?i:'s|'t|'re|'ve|'m|'ll|'d)|[^\\r\\n\\p{L}\\p{N}]?[\\p{L}\\p{M}]+|\\p{N}| ?[^\\s\\p{L}\\p{M}\\p{N}]+[\\r\\n]*|\\s*[\\r\\n]+|\\s+(?!\\S)|\\s+", - "split_special_tokens": false, - "tokenizer_class": "TokenizersBackend", - "unk_token": null, - "video_token": "<|video_pad|>", - "vision_bos_token": "<|vision_start|>", - "vision_eos_token": "<|vision_end|>" -} diff --git a/substitutivity_original_Estonian/Qwen3.5-2B-Base_substitutivity_splits_original_features_train_substitutivity_splits_original_features_test2/checkpoint-2490/trainer_state.json b/substitutivity_original_Estonian/Qwen3.5-2B-Base_substitutivity_splits_original_features_train_substitutivity_splits_original_features_test2/checkpoint-2490/trainer_state.json deleted file mode 100644 index 237737af6b1a4fae99eba28d42f3842ca2340106..0000000000000000000000000000000000000000 --- a/substitutivity_original_Estonian/Qwen3.5-2B-Base_substitutivity_splits_original_features_train_substitutivity_splits_original_features_test2/checkpoint-2490/trainer_state.json +++ /dev/null @@ -1,656 +0,0 @@ -{ - "best_global_step": null, - "best_metric": null, - "best_model_checkpoint": null, - "epoch": 6.0, - "eval_steps": 500, - "global_step": 2490, - "is_hyper_param_search": false, - "is_local_process_zero": true, - "is_world_process_zero": true, - "log_history": [ - { - "entropy": 2.2933380508422854, - "epoch": 0.12062726176115803, - "grad_norm": 2.8075337409973145, - "learning_rate": 1.5531017374629602e-05, - "loss": 2.1429234313964844, - "mean_token_accuracy": 0.5779684272408485, - "num_tokens": 121428.0, - "step": 50 - }, - { - "entropy": 1.1315062403678895, - "epoch": 0.24125452352231605, - "grad_norm": 1.9415371417999268, - "learning_rate": 3.1378994287516954e-05, - "loss": 1.0461698913574218, - "mean_token_accuracy": 0.7166991305351257, - "num_tokens": 243966.0, - "step": 100 - }, - { - "entropy": 0.9537658470869065, - "epoch": 0.3618817852834741, - "grad_norm": 1.8353157043457031, - "learning_rate": 4.72269712004043e-05, - "loss": 0.8774301147460938, - "mean_token_accuracy": 0.7506789201498032, - "num_tokens": 366649.0, - "step": 150 - }, - { - "entropy": 0.8955666983127594, - "epoch": 0.4825090470446321, - "grad_norm": 1.5419880151748657, - "learning_rate": 6.307494811329165e-05, - "loss": 0.8234302520751953, - "mean_token_accuracy": 0.7616239994764328, - "num_tokens": 488367.0, - "step": 200 - }, - { - "entropy": 0.8689469587802887, - "epoch": 0.6031363088057901, - "grad_norm": 1.5661484003067017, - "learning_rate": 7.8922925026179e-05, - "loss": 0.7961890411376953, - "mean_token_accuracy": 0.767715499997139, - "num_tokens": 603131.0, - "step": 250 - }, - { - "entropy": 0.8425652319192887, - "epoch": 0.7237635705669482, - "grad_norm": 1.5694905519485474, - "learning_rate": 9.477090193906636e-05, - "loss": 0.7718594360351563, - "mean_token_accuracy": 0.7736722755432129, - "num_tokens": 726412.0, - "step": 300 - }, - { - "entropy": 0.843294187784195, - "epoch": 0.8443908323281062, - "grad_norm": 1.1118028163909912, - "learning_rate": 0.00011061887885195371, - "loss": 0.7801297760009765, - "mean_token_accuracy": 0.7714760398864746, - "num_tokens": 846516.0, - "step": 350 - }, - { - "entropy": 0.8213214933872223, - "epoch": 0.9650180940892642, - "grad_norm": 0.9202918410301208, - "learning_rate": 0.00012646685576484106, - "loss": 0.7514453887939453, - "mean_token_accuracy": 0.7774044269323349, - "num_tokens": 958978.0, - "step": 400 - }, - { - "epoch": 1.0, - "eval_entropy": 0.818668900915746, - "eval_mean_token_accuracy": 0.7533496866065464, - "eval_not_syn_loss": 0.7798981070518494, - "eval_not_syn_runtime": 53.8744, - "eval_not_syn_samples_per_second": 26.358, - "eval_not_syn_steps_per_second": 3.304, - "eval_num_tokens": 993806.0, - "step": 415 - }, - { - "epoch": 1.0, - "eval_entropy": 0.7627468125873738, - "eval_mean_token_accuracy": 0.8024111574285486, - "eval_num_tokens": 993806.0, - "eval_syn_loss": 0.7218216061592102, - "eval_syn_runtime": 55.3239, - "eval_syn_samples_per_second": 25.667, - "eval_syn_steps_per_second": 3.217, - "step": 415 - }, - { - "entropy": 0.7802230679627621, - "epoch": 1.0844390832328106, - "grad_norm": 1.0145268440246582, - "learning_rate": 0.00013151131541121804, - "loss": 0.7111608123779297, - "mean_token_accuracy": 0.7845737542768921, - "num_tokens": 1080301.0, - "step": 450 - }, - { - "entropy": 0.7558828103542328, - "epoch": 1.2050663449939687, - "grad_norm": 1.0477591753005981, - "learning_rate": 0.0001313741160170038, - "loss": 0.6933037567138672, - "mean_token_accuracy": 0.7889221370220184, - "num_tokens": 1201215.0, - "step": 500 - }, - { - "entropy": 0.7593123841285706, - "epoch": 1.3256936067551266, - "grad_norm": 0.899161159992218, - "learning_rate": 0.00013112089697387701, - "loss": 0.6962493133544921, - "mean_token_accuracy": 0.7877359414100646, - "num_tokens": 1322895.0, - "step": 550 - }, - { - "entropy": 0.7460986822843552, - "epoch": 1.4463208685162847, - "grad_norm": 1.1061116456985474, - "learning_rate": 0.00013075210608885215, - "loss": 0.6817412567138672, - "mean_token_accuracy": 0.7904874366521836, - "num_tokens": 1436749.0, - "step": 600 - }, - { - "entropy": 0.735516265630722, - "epoch": 1.5669481302774426, - "grad_norm": 0.9925405979156494, - "learning_rate": 0.00013026839555278796, - "loss": 0.6690705871582031, - "mean_token_accuracy": 0.7930250608921051, - "num_tokens": 1555400.0, - "step": 650 - }, - { - "entropy": 0.7302519899606704, - "epoch": 1.6875753920386007, - "grad_norm": 0.800025463104248, - "learning_rate": 0.0001296706207870159, - "loss": 0.6714833831787109, - "mean_token_accuracy": 0.7946307575702667, - "num_tokens": 1676892.0, - "step": 700 - }, - { - "entropy": 0.7321263110637665, - "epoch": 1.8082026537997589, - "grad_norm": 1.0500682592391968, - "learning_rate": 0.00012895983893056423, - "loss": 0.6672868347167968, - "mean_token_accuracy": 0.7947645837068558, - "num_tokens": 1802689.0, - "step": 750 - }, - { - "entropy": 0.7075755137205124, - "epoch": 1.9288299155609168, - "grad_norm": 0.9248411655426025, - "learning_rate": 0.00012813730697065298, - "loss": 0.6466568756103516, - "mean_token_accuracy": 0.798747689127922, - "num_tokens": 1920804.0, - "step": 800 - }, - { - "epoch": 2.0, - "eval_entropy": 0.7426781811740961, - "eval_mean_token_accuracy": 0.7693507239389955, - "eval_not_syn_loss": 0.7139139175415039, - "eval_not_syn_runtime": 54.4286, - "eval_not_syn_samples_per_second": 26.089, - "eval_not_syn_steps_per_second": 3.27, - "eval_num_tokens": 1987612.0, - "step": 830 - }, - { - "epoch": 2.0, - "eval_entropy": 0.6937930352902144, - "eval_mean_token_accuracy": 0.8180048633157537, - "eval_num_tokens": 1987612.0, - "eval_syn_loss": 0.6489660143852234, - "eval_syn_runtime": 56.1525, - "eval_syn_samples_per_second": 25.288, - "eval_syn_steps_per_second": 3.17, - "step": 830 - }, - { - "entropy": 0.6853217440422135, - "epoch": 2.0482509047044632, - "grad_norm": 0.6932648420333862, - "learning_rate": 0.00012720447951976603, - "loss": 0.6212613296508789, - "mean_token_accuracy": 0.8047762579388089, - "num_tokens": 2039660.0, - "step": 850 - }, - { - "entropy": 0.6459032052755356, - "epoch": 2.168878166465621, - "grad_norm": 1.0688549280166626, - "learning_rate": 0.00012616300624323124, - "loss": 0.5740525817871094, - "mean_token_accuracy": 0.8148404061794281, - "num_tokens": 2161772.0, - "step": 900 - }, - { - "entropy": 0.6463843190670013, - "epoch": 2.2895054282267795, - "grad_norm": 0.847271203994751, - "learning_rate": 0.0001250147289418581, - "loss": 0.577717170715332, - "mean_token_accuracy": 0.814143658876419, - "num_tokens": 2280833.0, - "step": 950 - }, - { - "entropy": 0.6461306965351105, - "epoch": 2.4101326899879374, - "grad_norm": 0.9998957514762878, - "learning_rate": 0.00012376167829479217, - "loss": 0.5813011932373047, - "mean_token_accuracy": 0.8144401633739471, - "num_tokens": 2398206.0, - "step": 1000 - }, - { - "entropy": 0.640914659500122, - "epoch": 2.5307599517490953, - "grad_norm": 0.8692949414253235, - "learning_rate": 0.0001224060702683459, - "loss": 0.5755012512207032, - "mean_token_accuracy": 0.8168018108606339, - "num_tokens": 2518888.0, - "step": 1050 - }, - { - "entropy": 0.6421274638175964, - "epoch": 2.651387213510253, - "grad_norm": 0.8920196294784546, - "learning_rate": 0.00012095030219715738, - "loss": 0.5785483932495117, - "mean_token_accuracy": 0.8137972563505173, - "num_tokens": 2639763.0, - "step": 1100 - }, - { - "entropy": 0.6466812592744827, - "epoch": 2.772014475271411, - "grad_norm": 0.6876946687698364, - "learning_rate": 0.00011939694854460696, - "loss": 0.5816070938110351, - "mean_token_accuracy": 0.8145407867431641, - "num_tokens": 2755540.0, - "step": 1150 - }, - { - "entropy": 0.6378493067622185, - "epoch": 2.8926417370325694, - "grad_norm": 0.7532753944396973, - "learning_rate": 0.00011774875634998911, - "loss": 0.5754346084594727, - "mean_token_accuracy": 0.8148861181735992, - "num_tokens": 2874118.0, - "step": 1200 - }, - { - "epoch": 3.0, - "eval_entropy": 0.6912222209941136, - "eval_mean_token_accuracy": 0.7868467750174276, - "eval_not_syn_loss": 0.6782422661781311, - "eval_not_syn_runtime": 53.2627, - "eval_not_syn_samples_per_second": 26.66, - "eval_not_syn_steps_per_second": 3.342, - "eval_num_tokens": 2981418.0, - "step": 1245 - }, - { - "epoch": 3.0, - "eval_entropy": 0.6467212934172555, - "eval_mean_token_accuracy": 0.8215091817834405, - "eval_num_tokens": 2981418.0, - "eval_syn_loss": 0.6301999092102051, - "eval_syn_runtime": 55.0115, - "eval_syn_samples_per_second": 25.813, - "eval_syn_steps_per_second": 3.236, - "step": 1245 - }, - { - "entropy": 0.6316030452949832, - "epoch": 3.012062726176116, - "grad_norm": 0.7761592864990234, - "learning_rate": 0.00011600864037049124, - "loss": 0.5697880935668945, - "mean_token_accuracy": 0.8159356400220081, - "num_tokens": 2994602.0, - "step": 1250 - }, - { - "entropy": 0.5467158386111259, - "epoch": 3.132689987937274, - "grad_norm": 0.8169087767601013, - "learning_rate": 0.00011417967792657078, - "loss": 0.47844127655029295, - "mean_token_accuracy": 0.8378052765130997, - "num_tokens": 3121094.0, - "step": 1300 - }, - { - "entropy": 0.5454566389322281, - "epoch": 3.2533172496984317, - "grad_norm": 0.8219636678695679, - "learning_rate": 0.0001122651034598459, - "loss": 0.4839338302612305, - "mean_token_accuracy": 0.8385071897506714, - "num_tokens": 3248038.0, - "step": 1350 - }, - { - "entropy": 0.5598723018169403, - "epoch": 3.37394451145959, - "grad_norm": 0.9956360459327698, - "learning_rate": 0.00011026830281312423, - "loss": 0.492743034362793, - "mean_token_accuracy": 0.8357370454072952, - "num_tokens": 3363487.0, - "step": 1400 - }, - { - "entropy": 0.5479365903139114, - "epoch": 3.494571773220748, - "grad_norm": 0.8586823344230652, - "learning_rate": 0.00010819280724268524, - "loss": 0.4810820770263672, - "mean_token_accuracy": 0.837928666472435, - "num_tokens": 3480513.0, - "step": 1450 - }, - { - "entropy": 0.5619524419307709, - "epoch": 3.615199034981906, - "grad_norm": 1.0894954204559326, - "learning_rate": 0.0001060422871734048, - "loss": 0.4974747085571289, - "mean_token_accuracy": 0.8349784809350967, - "num_tokens": 3597183.0, - "step": 1500 - }, - { - "entropy": 0.5588027596473694, - "epoch": 3.7358262967430638, - "grad_norm": 0.9905257225036621, - "learning_rate": 0.00010382054570776654, - "loss": 0.49687202453613283, - "mean_token_accuracy": 0.834458841085434, - "num_tokens": 3714173.0, - "step": 1550 - }, - { - "entropy": 0.5654991599917412, - "epoch": 3.856453558504222, - "grad_norm": 1.0116610527038574, - "learning_rate": 0.00010153151190023791, - "loss": 0.5016670989990234, - "mean_token_accuracy": 0.8326652735471726, - "num_tokens": 3832652.0, - "step": 1600 - }, - { - "entropy": 0.5605382058024406, - "epoch": 3.97708082026538, - "grad_norm": 0.8011100888252258, - "learning_rate": 9.917923380890624e-05, - "loss": 0.49551734924316404, - "mean_token_accuracy": 0.8343535929918289, - "num_tokens": 3955167.0, - "step": 1650 - }, - { - "epoch": 4.0, - "eval_entropy": 0.5819476859623127, - "eval_mean_token_accuracy": 0.7963069781158747, - "eval_not_syn_loss": 0.6924306750297546, - "eval_not_syn_runtime": 54.027, - "eval_not_syn_samples_per_second": 26.283, - "eval_not_syn_steps_per_second": 3.295, - "eval_num_tokens": 3975224.0, - "step": 1660 - }, - { - "epoch": 4.0, - "eval_entropy": 0.545470636212424, - "eval_mean_token_accuracy": 0.8127643666910321, - "eval_num_tokens": 3975224.0, - "eval_syn_loss": 0.649576723575592, - "eval_syn_runtime": 55.2296, - "eval_syn_samples_per_second": 25.711, - "eval_syn_steps_per_second": 3.223, - "step": 1660 - }, - { - "entropy": 0.4920780294471317, - "epoch": 4.0965018094089265, - "grad_norm": 0.8362822532653809, - "learning_rate": 9.676787133666148e-05, - "loss": 0.4147996139526367, - "mean_token_accuracy": 0.8581844628459275, - "num_tokens": 4072361.0, - "step": 1700 - }, - { - "entropy": 0.4675139603018761, - "epoch": 4.217129071170085, - "grad_norm": 0.9920316338539124, - "learning_rate": 9.430168887458645e-05, - "loss": 0.3863528823852539, - "mean_token_accuracy": 0.865827117562294, - "num_tokens": 4194978.0, - "step": 1750 - }, - { - "entropy": 0.47163557589054106, - "epoch": 4.337756332931242, - "grad_norm": 1.0789282321929932, - "learning_rate": 9.178504776056418e-05, - "loss": 0.390842170715332, - "mean_token_accuracy": 0.8634983265399933, - "num_tokens": 4315548.0, - "step": 1800 - }, - { - "entropy": 0.47117673605680466, - "epoch": 4.458383594692401, - "grad_norm": 1.0369561910629272, - "learning_rate": 8.922239856643896e-05, - "loss": 0.39387256622314454, - "mean_token_accuracy": 0.8625931572914124, - "num_tokens": 4431469.0, - "step": 1850 - }, - { - "entropy": 0.479859294295311, - "epoch": 4.579010856453559, - "grad_norm": 1.0827305316925049, - "learning_rate": 8.661827322737095e-05, - "loss": 0.40195709228515625, - "mean_token_accuracy": 0.8612629306316376, - "num_tokens": 4551803.0, - "step": 1900 - }, - { - "entropy": 0.4729380664229393, - "epoch": 4.699638118214716, - "grad_norm": 1.0114246606826782, - "learning_rate": 8.397727702730326e-05, - "loss": 0.39648338317871096, - "mean_token_accuracy": 0.8610477095842362, - "num_tokens": 4674468.0, - "step": 1950 - }, - { - "entropy": 0.4903118047118187, - "epoch": 4.820265379975875, - "grad_norm": 1.0115467309951782, - "learning_rate": 8.130408045471496e-05, - "loss": 0.4067733001708984, - "mean_token_accuracy": 0.8597196912765503, - "num_tokens": 4788027.0, - "step": 2000 - }, - { - "entropy": 0.48099748432636263, - "epoch": 4.940892641737032, - "grad_norm": 1.0201005935668945, - "learning_rate": 7.860341094306242e-05, - "loss": 0.4030665588378906, - "mean_token_accuracy": 0.8603792876005173, - "num_tokens": 4907561.0, - "step": 2050 - }, - { - "epoch": 5.0, - "eval_entropy": 0.5492983292997553, - "eval_mean_token_accuracy": 0.796870938847574, - "eval_not_syn_loss": 0.7129728198051453, - "eval_not_syn_runtime": 53.9575, - "eval_not_syn_samples_per_second": 26.317, - "eval_not_syn_steps_per_second": 3.299, - "eval_num_tokens": 4969030.0, - "step": 2075 - }, - { - "epoch": 5.0, - "eval_entropy": 0.513823946037989, - "eval_mean_token_accuracy": 0.8174765464964877, - "eval_num_tokens": 4969030.0, - "eval_syn_loss": 0.6699439883232117, - "eval_syn_runtime": 56.0033, - "eval_syn_samples_per_second": 25.356, - "eval_syn_steps_per_second": 3.178, - "step": 2075 - }, - { - "entropy": 0.43584728511897003, - "epoch": 5.060313630880579, - "grad_norm": 0.9914600253105164, - "learning_rate": 7.588004451051618e-05, - "loss": 0.34999092102050783, - "mean_token_accuracy": 0.8777085295831314, - "num_tokens": 5026625.0, - "step": 2100 - }, - { - "entropy": 0.38114599585533143, - "epoch": 5.180940892641737, - "grad_norm": 1.232283115386963, - "learning_rate": 7.313879731377764e-05, - "loss": 0.2903956604003906, - "mean_token_accuracy": 0.896172685623169, - "num_tokens": 5141010.0, - "step": 2150 - }, - { - "entropy": 0.3792745009064674, - "epoch": 5.301568154402895, - "grad_norm": 1.2843542098999023, - "learning_rate": 7.03845171309126e-05, - "loss": 0.2903476333618164, - "mean_token_accuracy": 0.8951605755090714, - "num_tokens": 5261240.0, - "step": 2200 - }, - { - "entropy": 0.3744692197442055, - "epoch": 5.422195416164053, - "grad_norm": 1.23075270652771, - "learning_rate": 6.76220747882637e-05, - "loss": 0.29011749267578124, - "mean_token_accuracy": 0.8963996982574463, - "num_tokens": 5382160.0, - "step": 2250 - }, - { - "entropy": 0.3795280712842941, - "epoch": 5.542822677925211, - "grad_norm": 1.4011399745941162, - "learning_rate": 6.485635554660295e-05, - "loss": 0.2921779823303223, - "mean_token_accuracy": 0.8954424083232879, - "num_tokens": 5502002.0, - "step": 2300 - }, - { - "entropy": 0.3797056859731674, - "epoch": 5.66344993968637, - "grad_norm": 1.1022812128067017, - "learning_rate": 6.209225046175775e-05, - "loss": 0.29343433380126954, - "mean_token_accuracy": 0.8939031893014908, - "num_tokens": 5629969.0, - "step": 2350 - }, - { - "entropy": 0.3827402114868164, - "epoch": 5.784077201447527, - "grad_norm": 1.22834312915802, - "learning_rate": 5.9334647734988365e-05, - "loss": 0.2954188537597656, - "mean_token_accuracy": 0.8937842601537704, - "num_tokens": 5750596.0, - "step": 2400 - }, - { - "entropy": 0.38422658771276474, - "epoch": 5.904704463208685, - "grad_norm": 1.425066351890564, - "learning_rate": 5.6588424068413775e-05, - "loss": 0.296480770111084, - "mean_token_accuracy": 0.8936110073328019, - "num_tokens": 5868110.0, - "step": 2450 - }, - { - "epoch": 6.0, - "eval_entropy": 0.49826224684045556, - "eval_mean_token_accuracy": 0.795356297760867, - "eval_not_syn_loss": 0.7700047492980957, - "eval_not_syn_runtime": 53.0632, - "eval_not_syn_samples_per_second": 26.761, - "eval_not_syn_steps_per_second": 3.354, - "eval_num_tokens": 5962836.0, - "step": 2490 - }, - { - "epoch": 6.0, - "eval_entropy": 0.4654735526007213, - "eval_mean_token_accuracy": 0.8118122223387944, - "eval_num_tokens": 5962836.0, - "eval_syn_loss": 0.7288840413093567, - "eval_syn_runtime": 54.6517, - "eval_syn_samples_per_second": 25.983, - "eval_syn_steps_per_second": 3.257, - "step": 2490 - } - ], - "logging_steps": 50, - "max_steps": 4150, - "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": 1.229483151455831e+17, - "train_batch_size": 8, - "trial_name": null, - "trial_params": null -} diff --git a/substitutivity_original_Estonian/Qwen3.5-2B-Base_substitutivity_splits_original_features_train_substitutivity_splits_original_features_test2/checkpoint-2490/training_args.bin b/substitutivity_original_Estonian/Qwen3.5-2B-Base_substitutivity_splits_original_features_train_substitutivity_splits_original_features_test2/checkpoint-2490/training_args.bin deleted file mode 100644 index f3d8098f65400413dd700f32ada5649470871017..0000000000000000000000000000000000000000 --- a/substitutivity_original_Estonian/Qwen3.5-2B-Base_substitutivity_splits_original_features_train_substitutivity_splits_original_features_test2/checkpoint-2490/training_args.bin +++ /dev/null @@ -1,3 +0,0 @@ -version https://git-lfs.github.com/spec/v1 -oid sha256:617af0f41414927baaaf158949585bb75eb4b05ff78baa5216c860ef0b431009 -size 6033 diff --git a/substitutivity_original_Estonian/Qwen3.5-2B-Base_substitutivity_splits_original_features_train_substitutivity_splits_original_features_test2/checkpoint-2905/README.md b/substitutivity_original_Estonian/Qwen3.5-2B-Base_substitutivity_splits_original_features_train_substitutivity_splits_original_features_test2/checkpoint-2905/README.md deleted file mode 100644 index ece204d005f7bf161d0e448efdd56fd297b7c8a2..0000000000000000000000000000000000000000 --- a/substitutivity_original_Estonian/Qwen3.5-2B-Base_substitutivity_splits_original_features_train_substitutivity_splits_original_features_test2/checkpoint-2905/README.md +++ /dev/null @@ -1,209 +0,0 @@ ---- -base_model: Qwen/Qwen3.5-2B-Base -library_name: peft -pipeline_tag: text-generation -tags: -- base_model:adapter:Qwen/Qwen3.5-2B-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/substitutivity_original_Estonian/Qwen3.5-2B-Base_substitutivity_splits_original_features_train_substitutivity_splits_original_features_test2/checkpoint-2905/adapter_config.json b/substitutivity_original_Estonian/Qwen3.5-2B-Base_substitutivity_splits_original_features_train_substitutivity_splits_original_features_test2/checkpoint-2905/adapter_config.json deleted file mode 100644 index 6a095bb3bfe249ce8e15df6ae999e88fa1e17162..0000000000000000000000000000000000000000 --- a/substitutivity_original_Estonian/Qwen3.5-2B-Base_substitutivity_splits_original_features_train_substitutivity_splits_original_features_test2/checkpoint-2905/adapter_config.json +++ /dev/null @@ -1,46 +0,0 @@ -{ - "alora_invocation_tokens": null, - "alpha_pattern": {}, - "arrow_config": null, - "auto_mapping": null, - "base_model_name_or_path": "Qwen/Qwen3.5-2B-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": 0.05731002492184622, - "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": [ - "o_proj", - "q_proj", - "k_proj", - "down_proj", - "up_proj", - "v_proj", - "gate_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/substitutivity_original_Estonian/Qwen3.5-2B-Base_substitutivity_splits_original_features_train_substitutivity_splits_original_features_test2/checkpoint-2905/adapter_model.safetensors b/substitutivity_original_Estonian/Qwen3.5-2B-Base_substitutivity_splits_original_features_train_substitutivity_splits_original_features_test2/checkpoint-2905/adapter_model.safetensors deleted file mode 100644 index 69a95dda5ce9ed77e753f31f4d973ed788a4ec85..0000000000000000000000000000000000000000 --- a/substitutivity_original_Estonian/Qwen3.5-2B-Base_substitutivity_splits_original_features_train_substitutivity_splits_original_features_test2/checkpoint-2905/adapter_model.safetensors +++ /dev/null @@ -1,3 +0,0 @@ -version https://git-lfs.github.com/spec/v1 -oid sha256:85db14e8274baa89dc80973c13ac18e8b5d6a62d9e5b2dccb2f3b47af3cd8d5f -size 174613392 diff --git a/substitutivity_original_Estonian/Qwen3.5-2B-Base_substitutivity_splits_original_features_train_substitutivity_splits_original_features_test2/checkpoint-2905/chat_template.jinja b/substitutivity_original_Estonian/Qwen3.5-2B-Base_substitutivity_splits_original_features_train_substitutivity_splits_original_features_test2/checkpoint-2905/chat_template.jinja deleted file mode 100644 index 0ef09f214eaa6d9bca297988afc1454b5827b2c7..0000000000000000000000000000000000000000 --- a/substitutivity_original_Estonian/Qwen3.5-2B-Base_substitutivity_splits_original_features_train_substitutivity_splits_original_features_test2/checkpoint-2905/chat_template.jinja +++ /dev/null @@ -1,154 +0,0 @@ -{%- set image_count = namespace(value=0) %} -{%- set video_count = namespace(value=0) %} -{%- macro render_content(content, do_vision_count, is_system_content=false) %} - {%- if content is string %} - {{- content }} - {%- elif content is iterable and content is not mapping %} - {%- for item in content %} - {%- if 'image' in item or 'image_url' in item or item.type == 'image' %} - {%- if is_system_content %} - {{- raise_exception('System message cannot contain images.') }} - {%- endif %} - {%- if do_vision_count %} - {%- set image_count.value = image_count.value + 1 %} - {%- endif %} - {%- if add_vision_id %} - {{- 'Picture ' ~ image_count.value ~ ': ' }} - {%- endif %} - {{- '<|vision_start|><|image_pad|><|vision_end|>' }} - {%- elif 'video' in item or item.type == 'video' %} - {%- if is_system_content %} - {{- raise_exception('System message cannot contain videos.') }} - {%- endif %} - {%- if do_vision_count %} - {%- set video_count.value = video_count.value + 1 %} - {%- endif %} - {%- if add_vision_id %} - {{- 'Video ' ~ video_count.value ~ ': ' }} - {%- endif %} - {{- '<|vision_start|><|video_pad|><|vision_end|>' }} - {%- elif 'text' in item %} - {{- item.text }} - {%- else %} - {{- raise_exception('Unexpected item type in content.') }} - {%- endif %} - {%- endfor %} - {%- elif content is none or content is undefined %} - {{- '' }} - {%- else %} - {{- raise_exception('Unexpected content type.') }} - {%- endif %} -{%- endmacro %} -{%- if not messages %} - {{- raise_exception('No messages provided.') }} -{%- endif %} -{%- if tools and tools is iterable and tools is not mapping %} - {{- '<|im_start|>system\n' }} - {{- "# Tools\n\nYou have access to the following functions:\n\n" }} - {%- for tool in tools %} - {{- "\n" }} - {{- tool | tojson }} - {%- endfor %} - {{- "\n" }} - {{- '\n\nIf you choose to call a function ONLY reply in the following format with NO suffix:\n\n\n\n\nvalue_1\n\n\nThis is the value for the second parameter\nthat can span\nmultiple lines\n\n\n\n\n\nReminder:\n- Function calls MUST follow the specified format: an inner block must be nested within XML tags\n- Required parameters MUST be specified\n- You may provide optional reasoning for your function call in natural language BEFORE the function call, but NOT after\n- If there is no function call available, answer the question like normal with your current knowledge and do not tell the user about function calls\n' }} - {%- if messages[0].role == 'system' %} - {%- set content = render_content(messages[0].content, false, true)|trim %} - {%- if content %} - {{- '\n\n' + content }} - {%- endif %} - {%- endif %} - {{- '<|im_end|>\n' }} -{%- else %} - {%- if messages[0].role == 'system' %} - {%- set content = render_content(messages[0].content, false, true)|trim %} - {{- '<|im_start|>system\n' + 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" %} - {%- set content = render_content(message.content, false)|trim %} - {%- if not(content.startswith('') and content.endswith('')) %} - {%- set ns.multi_step_tool = false %} - {%- set ns.last_query_index = index %} - {%- endif %} - {%- endif %} -{%- endfor %} -{%- if ns.multi_step_tool %} - {{- raise_exception('No user query found in messages.') }} -{%- endif %} -{%- for message in messages %} - {%- set content = render_content(message.content, true)|trim %} - {%- if message.role == "system" %} - {%- if not loop.first %} - {{- raise_exception('System message must be at the beginning.') }} - {%- endif %} - {%- elif message.role == "user" %} - {{- '<|im_start|>' + message.role + '\n' + content + '<|im_end|>' + '\n' }} - {%- elif message.role == "assistant" %} - {%- set reasoning_content = '' %} - {%- if message.reasoning_content is string %} - {%- set reasoning_content = message.reasoning_content %} - {%- else %} - {%- if '' in content %} - {%- set reasoning_content = content.split('')[0].rstrip('\n').split('')[-1].lstrip('\n') %} - {%- set content = content.split('')[-1].lstrip('\n') %} - {%- endif %} - {%- endif %} - {%- set reasoning_content = reasoning_content|trim %} - {%- if loop.index0 > ns.last_query_index %} - {{- '<|im_start|>' + message.role + '\n\n' + reasoning_content + '\n\n\n' + content }} - {%- else %} - {{- '<|im_start|>' + message.role + '\n' + content }} - {%- endif %} - {%- if message.tool_calls and message.tool_calls is iterable and message.tool_calls is not mapping %} - {%- for tool_call in message.tool_calls %} - {%- if tool_call.function is defined %} - {%- set tool_call = tool_call.function %} - {%- endif %} - {%- if loop.first %} - {%- if content|trim %} - {{- '\n\n\n\n' }} - {%- else %} - {{- '\n\n' }} - {%- endif %} - {%- else %} - {{- '\n\n\n' }} - {%- endif %} - {%- if tool_call.arguments is defined %} - {%- for args_name, args_value in tool_call.arguments|items %} - {{- '\n' }} - {%- set args_value = args_value | tojson | safe if args_value is mapping or (args_value is sequence and args_value is not string) else args_value | string %} - {{- args_value }} - {{- '\n\n' }} - {%- endfor %} - {%- endif %} - {{- '\n' }} - {%- endfor %} - {%- endif %} - {{- '<|im_end|>\n' }} - {%- elif message.role == "tool" %} - {%- if loop.previtem and loop.previtem.role != "tool" %} - {{- '<|im_start|>user' }} - {%- endif %} - {{- '\n\n' }} - {{- content }} - {{- '\n' }} - {%- if not loop.last and loop.nextitem.role != "tool" %} - {{- '<|im_end|>\n' }} - {%- elif loop.last %} - {{- '<|im_end|>\n' }} - {%- endif %} - {%- else %} - {{- raise_exception('Unexpected message role.') }} - {%- endif %} -{%- endfor %} -{%- if add_generation_prompt %} - {{- '<|im_start|>assistant\n' }} - {%- if enable_thinking is defined and enable_thinking is true %} - {{- '\n' }} - {%- else %} - {{- '\n\n\n\n' }} - {%- endif %} -{%- endif %} \ No newline at end of file diff --git a/substitutivity_original_Estonian/Qwen3.5-2B-Base_substitutivity_splits_original_features_train_substitutivity_splits_original_features_test2/checkpoint-2905/tokenizer.json b/substitutivity_original_Estonian/Qwen3.5-2B-Base_substitutivity_splits_original_features_train_substitutivity_splits_original_features_test2/checkpoint-2905/tokenizer.json deleted file mode 100644 index 67741b04f23bfdb46501f748ce27865ec82eccfb..0000000000000000000000000000000000000000 --- a/substitutivity_original_Estonian/Qwen3.5-2B-Base_substitutivity_splits_original_features_train_substitutivity_splits_original_features_test2/checkpoint-2905/tokenizer.json +++ /dev/null @@ -1,3 +0,0 @@ -version https://git-lfs.github.com/spec/v1 -oid sha256:87a7830d63fcf43bf241c3c5242e96e62dd3fdc29224ca26fed8ea333db72de4 -size 19989343 diff --git a/substitutivity_original_Estonian/Qwen3.5-2B-Base_substitutivity_splits_original_features_train_substitutivity_splits_original_features_test2/checkpoint-2905/tokenizer_config.json b/substitutivity_original_Estonian/Qwen3.5-2B-Base_substitutivity_splits_original_features_train_substitutivity_splits_original_features_test2/checkpoint-2905/tokenizer_config.json deleted file mode 100644 index b4a37b2a6fd3ab3317cd7bac72855be1a843b2bb..0000000000000000000000000000000000000000 --- a/substitutivity_original_Estonian/Qwen3.5-2B-Base_substitutivity_splits_original_features_train_substitutivity_splits_original_features_test2/checkpoint-2905/tokenizer_config.json +++ /dev/null @@ -1,31 +0,0 @@ -{ - "add_prefix_space": false, - "audio_bos_token": "<|audio_start|>", - "audio_eos_token": "<|audio_end|>", - "audio_token": "<|audio_pad|>", - "backend": "tokenizers", - "bos_token": null, - "clean_up_tokenization_spaces": false, - "eos_token": "<|endoftext|>", - "errors": "replace", - "image_token": "<|image_pad|>", - "is_local": false, - "model_max_length": 262144, - "model_specific_special_tokens": { - "audio_bos_token": "<|audio_start|>", - "audio_eos_token": "<|audio_end|>", - "audio_token": "<|audio_pad|>", - "image_token": "<|image_pad|>", - "video_token": "<|video_pad|>", - "vision_bos_token": "<|vision_start|>", - "vision_eos_token": "<|vision_end|>" - }, - "pad_token": "<|endoftext|>", - "pretokenize_regex": "(?i:'s|'t|'re|'ve|'m|'ll|'d)|[^\\r\\n\\p{L}\\p{N}]?[\\p{L}\\p{M}]+|\\p{N}| ?[^\\s\\p{L}\\p{M}\\p{N}]+[\\r\\n]*|\\s*[\\r\\n]+|\\s+(?!\\S)|\\s+", - "split_special_tokens": false, - "tokenizer_class": "TokenizersBackend", - "unk_token": null, - "video_token": "<|video_pad|>", - "vision_bos_token": "<|vision_start|>", - "vision_eos_token": "<|vision_end|>" -} diff --git a/substitutivity_original_Estonian/Qwen3.5-2B-Base_substitutivity_splits_original_features_train_substitutivity_splits_original_features_test2/checkpoint-2905/trainer_state.json b/substitutivity_original_Estonian/Qwen3.5-2B-Base_substitutivity_splits_original_features_train_substitutivity_splits_original_features_test2/checkpoint-2905/trainer_state.json deleted file mode 100644 index 3a0faccecaf73890ee5b6a17969b9c15a34a2299..0000000000000000000000000000000000000000 --- a/substitutivity_original_Estonian/Qwen3.5-2B-Base_substitutivity_splits_original_features_train_substitutivity_splits_original_features_test2/checkpoint-2905/trainer_state.json +++ /dev/null @@ -1,768 +0,0 @@ -{ - "best_global_step": null, - "best_metric": null, - "best_model_checkpoint": null, - "epoch": 7.0, - "eval_steps": 500, - "global_step": 2905, - "is_hyper_param_search": false, - "is_local_process_zero": true, - "is_world_process_zero": true, - "log_history": [ - { - "entropy": 2.2933380508422854, - "epoch": 0.12062726176115803, - "grad_norm": 2.8075337409973145, - "learning_rate": 1.5531017374629602e-05, - "loss": 2.1429234313964844, - "mean_token_accuracy": 0.5779684272408485, - "num_tokens": 121428.0, - "step": 50 - }, - { - "entropy": 1.1315062403678895, - "epoch": 0.24125452352231605, - "grad_norm": 1.9415371417999268, - "learning_rate": 3.1378994287516954e-05, - "loss": 1.0461698913574218, - "mean_token_accuracy": 0.7166991305351257, - "num_tokens": 243966.0, - "step": 100 - }, - { - "entropy": 0.9537658470869065, - "epoch": 0.3618817852834741, - "grad_norm": 1.8353157043457031, - "learning_rate": 4.72269712004043e-05, - "loss": 0.8774301147460938, - "mean_token_accuracy": 0.7506789201498032, - "num_tokens": 366649.0, - "step": 150 - }, - { - "entropy": 0.8955666983127594, - "epoch": 0.4825090470446321, - "grad_norm": 1.5419880151748657, - "learning_rate": 6.307494811329165e-05, - "loss": 0.8234302520751953, - "mean_token_accuracy": 0.7616239994764328, - "num_tokens": 488367.0, - "step": 200 - }, - { - "entropy": 0.8689469587802887, - "epoch": 0.6031363088057901, - "grad_norm": 1.5661484003067017, - "learning_rate": 7.8922925026179e-05, - "loss": 0.7961890411376953, - "mean_token_accuracy": 0.767715499997139, - "num_tokens": 603131.0, - "step": 250 - }, - { - "entropy": 0.8425652319192887, - "epoch": 0.7237635705669482, - "grad_norm": 1.5694905519485474, - "learning_rate": 9.477090193906636e-05, - "loss": 0.7718594360351563, - "mean_token_accuracy": 0.7736722755432129, - "num_tokens": 726412.0, - "step": 300 - }, - { - "entropy": 0.843294187784195, - "epoch": 0.8443908323281062, - "grad_norm": 1.1118028163909912, - "learning_rate": 0.00011061887885195371, - "loss": 0.7801297760009765, - "mean_token_accuracy": 0.7714760398864746, - "num_tokens": 846516.0, - "step": 350 - }, - { - "entropy": 0.8213214933872223, - "epoch": 0.9650180940892642, - "grad_norm": 0.9202918410301208, - "learning_rate": 0.00012646685576484106, - "loss": 0.7514453887939453, - "mean_token_accuracy": 0.7774044269323349, - "num_tokens": 958978.0, - "step": 400 - }, - { - "epoch": 1.0, - "eval_entropy": 0.818668900915746, - "eval_mean_token_accuracy": 0.7533496866065464, - "eval_not_syn_loss": 0.7798981070518494, - "eval_not_syn_runtime": 53.8744, - "eval_not_syn_samples_per_second": 26.358, - "eval_not_syn_steps_per_second": 3.304, - "eval_num_tokens": 993806.0, - "step": 415 - }, - { - "epoch": 1.0, - "eval_entropy": 0.7627468125873738, - "eval_mean_token_accuracy": 0.8024111574285486, - "eval_num_tokens": 993806.0, - "eval_syn_loss": 0.7218216061592102, - "eval_syn_runtime": 55.3239, - "eval_syn_samples_per_second": 25.667, - "eval_syn_steps_per_second": 3.217, - "step": 415 - }, - { - "entropy": 0.7802230679627621, - "epoch": 1.0844390832328106, - "grad_norm": 1.0145268440246582, - "learning_rate": 0.00013151131541121804, - "loss": 0.7111608123779297, - "mean_token_accuracy": 0.7845737542768921, - "num_tokens": 1080301.0, - "step": 450 - }, - { - "entropy": 0.7558828103542328, - "epoch": 1.2050663449939687, - "grad_norm": 1.0477591753005981, - "learning_rate": 0.0001313741160170038, - "loss": 0.6933037567138672, - "mean_token_accuracy": 0.7889221370220184, - "num_tokens": 1201215.0, - "step": 500 - }, - { - "entropy": 0.7593123841285706, - "epoch": 1.3256936067551266, - "grad_norm": 0.899161159992218, - "learning_rate": 0.00013112089697387701, - "loss": 0.6962493133544921, - "mean_token_accuracy": 0.7877359414100646, - "num_tokens": 1322895.0, - "step": 550 - }, - { - "entropy": 0.7460986822843552, - "epoch": 1.4463208685162847, - "grad_norm": 1.1061116456985474, - "learning_rate": 0.00013075210608885215, - "loss": 0.6817412567138672, - "mean_token_accuracy": 0.7904874366521836, - "num_tokens": 1436749.0, - "step": 600 - }, - { - "entropy": 0.735516265630722, - "epoch": 1.5669481302774426, - "grad_norm": 0.9925405979156494, - "learning_rate": 0.00013026839555278796, - "loss": 0.6690705871582031, - "mean_token_accuracy": 0.7930250608921051, - "num_tokens": 1555400.0, - "step": 650 - }, - { - "entropy": 0.7302519899606704, - "epoch": 1.6875753920386007, - "grad_norm": 0.800025463104248, - "learning_rate": 0.0001296706207870159, - "loss": 0.6714833831787109, - "mean_token_accuracy": 0.7946307575702667, - "num_tokens": 1676892.0, - "step": 700 - }, - { - "entropy": 0.7321263110637665, - "epoch": 1.8082026537997589, - "grad_norm": 1.0500682592391968, - "learning_rate": 0.00012895983893056423, - "loss": 0.6672868347167968, - "mean_token_accuracy": 0.7947645837068558, - "num_tokens": 1802689.0, - "step": 750 - }, - { - "entropy": 0.7075755137205124, - "epoch": 1.9288299155609168, - "grad_norm": 0.9248411655426025, - "learning_rate": 0.00012813730697065298, - "loss": 0.6466568756103516, - "mean_token_accuracy": 0.798747689127922, - "num_tokens": 1920804.0, - "step": 800 - }, - { - "epoch": 2.0, - "eval_entropy": 0.7426781811740961, - "eval_mean_token_accuracy": 0.7693507239389955, - "eval_not_syn_loss": 0.7139139175415039, - "eval_not_syn_runtime": 54.4286, - "eval_not_syn_samples_per_second": 26.089, - "eval_not_syn_steps_per_second": 3.27, - "eval_num_tokens": 1987612.0, - "step": 830 - }, - { - "epoch": 2.0, - "eval_entropy": 0.6937930352902144, - "eval_mean_token_accuracy": 0.8180048633157537, - "eval_num_tokens": 1987612.0, - "eval_syn_loss": 0.6489660143852234, - "eval_syn_runtime": 56.1525, - "eval_syn_samples_per_second": 25.288, - "eval_syn_steps_per_second": 3.17, - "step": 830 - }, - { - "entropy": 0.6853217440422135, - "epoch": 2.0482509047044632, - "grad_norm": 0.6932648420333862, - "learning_rate": 0.00012720447951976603, - "loss": 0.6212613296508789, - "mean_token_accuracy": 0.8047762579388089, - "num_tokens": 2039660.0, - "step": 850 - }, - { - "entropy": 0.6459032052755356, - "epoch": 2.168878166465621, - "grad_norm": 1.0688549280166626, - "learning_rate": 0.00012616300624323124, - "loss": 0.5740525817871094, - "mean_token_accuracy": 0.8148404061794281, - "num_tokens": 2161772.0, - "step": 900 - }, - { - "entropy": 0.6463843190670013, - "epoch": 2.2895054282267795, - "grad_norm": 0.847271203994751, - "learning_rate": 0.0001250147289418581, - "loss": 0.577717170715332, - "mean_token_accuracy": 0.814143658876419, - "num_tokens": 2280833.0, - "step": 950 - }, - { - "entropy": 0.6461306965351105, - "epoch": 2.4101326899879374, - "grad_norm": 0.9998957514762878, - "learning_rate": 0.00012376167829479217, - "loss": 0.5813011932373047, - "mean_token_accuracy": 0.8144401633739471, - "num_tokens": 2398206.0, - "step": 1000 - }, - { - "entropy": 0.640914659500122, - "epoch": 2.5307599517490953, - "grad_norm": 0.8692949414253235, - "learning_rate": 0.0001224060702683459, - "loss": 0.5755012512207032, - "mean_token_accuracy": 0.8168018108606339, - "num_tokens": 2518888.0, - "step": 1050 - }, - { - "entropy": 0.6421274638175964, - "epoch": 2.651387213510253, - "grad_norm": 0.8920196294784546, - "learning_rate": 0.00012095030219715738, - "loss": 0.5785483932495117, - "mean_token_accuracy": 0.8137972563505173, - "num_tokens": 2639763.0, - "step": 1100 - }, - { - "entropy": 0.6466812592744827, - "epoch": 2.772014475271411, - "grad_norm": 0.6876946687698364, - "learning_rate": 0.00011939694854460696, - "loss": 0.5816070938110351, - "mean_token_accuracy": 0.8145407867431641, - "num_tokens": 2755540.0, - "step": 1150 - }, - { - "entropy": 0.6378493067622185, - "epoch": 2.8926417370325694, - "grad_norm": 0.7532753944396973, - "learning_rate": 0.00011774875634998911, - "loss": 0.5754346084594727, - "mean_token_accuracy": 0.8148861181735992, - "num_tokens": 2874118.0, - "step": 1200 - }, - { - "epoch": 3.0, - "eval_entropy": 0.6912222209941136, - "eval_mean_token_accuracy": 0.7868467750174276, - "eval_not_syn_loss": 0.6782422661781311, - "eval_not_syn_runtime": 53.2627, - "eval_not_syn_samples_per_second": 26.66, - "eval_not_syn_steps_per_second": 3.342, - "eval_num_tokens": 2981418.0, - "step": 1245 - }, - { - "epoch": 3.0, - "eval_entropy": 0.6467212934172555, - "eval_mean_token_accuracy": 0.8215091817834405, - "eval_num_tokens": 2981418.0, - "eval_syn_loss": 0.6301999092102051, - "eval_syn_runtime": 55.0115, - "eval_syn_samples_per_second": 25.813, - "eval_syn_steps_per_second": 3.236, - "step": 1245 - }, - { - "entropy": 0.6316030452949832, - "epoch": 3.012062726176116, - "grad_norm": 0.7761592864990234, - "learning_rate": 0.00011600864037049124, - "loss": 0.5697880935668945, - "mean_token_accuracy": 0.8159356400220081, - "num_tokens": 2994602.0, - "step": 1250 - }, - { - "entropy": 0.5467158386111259, - "epoch": 3.132689987937274, - "grad_norm": 0.8169087767601013, - "learning_rate": 0.00011417967792657078, - "loss": 0.47844127655029295, - "mean_token_accuracy": 0.8378052765130997, - "num_tokens": 3121094.0, - "step": 1300 - }, - { - "entropy": 0.5454566389322281, - "epoch": 3.2533172496984317, - "grad_norm": 0.8219636678695679, - "learning_rate": 0.0001122651034598459, - "loss": 0.4839338302612305, - "mean_token_accuracy": 0.8385071897506714, - "num_tokens": 3248038.0, - "step": 1350 - }, - { - "entropy": 0.5598723018169403, - "epoch": 3.37394451145959, - "grad_norm": 0.9956360459327698, - "learning_rate": 0.00011026830281312423, - "loss": 0.492743034362793, - "mean_token_accuracy": 0.8357370454072952, - "num_tokens": 3363487.0, - "step": 1400 - }, - { - "entropy": 0.5479365903139114, - "epoch": 3.494571773220748, - "grad_norm": 0.8586823344230652, - "learning_rate": 0.00010819280724268524, - "loss": 0.4810820770263672, - "mean_token_accuracy": 0.837928666472435, - "num_tokens": 3480513.0, - "step": 1450 - }, - { - "entropy": 0.5619524419307709, - "epoch": 3.615199034981906, - "grad_norm": 1.0894954204559326, - "learning_rate": 0.0001060422871734048, - "loss": 0.4974747085571289, - "mean_token_accuracy": 0.8349784809350967, - "num_tokens": 3597183.0, - "step": 1500 - }, - { - "entropy": 0.5588027596473694, - "epoch": 3.7358262967430638, - "grad_norm": 0.9905257225036621, - "learning_rate": 0.00010382054570776654, - "loss": 0.49687202453613283, - "mean_token_accuracy": 0.834458841085434, - "num_tokens": 3714173.0, - "step": 1550 - }, - { - "entropy": 0.5654991599917412, - "epoch": 3.856453558504222, - "grad_norm": 1.0116610527038574, - "learning_rate": 0.00010153151190023791, - "loss": 0.5016670989990234, - "mean_token_accuracy": 0.8326652735471726, - "num_tokens": 3832652.0, - "step": 1600 - }, - { - "entropy": 0.5605382058024406, - "epoch": 3.97708082026538, - "grad_norm": 0.8011100888252258, - "learning_rate": 9.917923380890624e-05, - "loss": 0.49551734924316404, - "mean_token_accuracy": 0.8343535929918289, - "num_tokens": 3955167.0, - "step": 1650 - }, - { - "epoch": 4.0, - "eval_entropy": 0.5819476859623127, - "eval_mean_token_accuracy": 0.7963069781158747, - "eval_not_syn_loss": 0.6924306750297546, - "eval_not_syn_runtime": 54.027, - "eval_not_syn_samples_per_second": 26.283, - "eval_not_syn_steps_per_second": 3.295, - "eval_num_tokens": 3975224.0, - "step": 1660 - }, - { - "epoch": 4.0, - "eval_entropy": 0.545470636212424, - "eval_mean_token_accuracy": 0.8127643666910321, - "eval_num_tokens": 3975224.0, - "eval_syn_loss": 0.649576723575592, - "eval_syn_runtime": 55.2296, - "eval_syn_samples_per_second": 25.711, - "eval_syn_steps_per_second": 3.223, - "step": 1660 - }, - { - "entropy": 0.4920780294471317, - "epoch": 4.0965018094089265, - "grad_norm": 0.8362822532653809, - "learning_rate": 9.676787133666148e-05, - "loss": 0.4147996139526367, - "mean_token_accuracy": 0.8581844628459275, - "num_tokens": 4072361.0, - "step": 1700 - }, - { - "entropy": 0.4675139603018761, - "epoch": 4.217129071170085, - "grad_norm": 0.9920316338539124, - "learning_rate": 9.430168887458645e-05, - "loss": 0.3863528823852539, - "mean_token_accuracy": 0.865827117562294, - "num_tokens": 4194978.0, - "step": 1750 - }, - { - "entropy": 0.47163557589054106, - "epoch": 4.337756332931242, - "grad_norm": 1.0789282321929932, - "learning_rate": 9.178504776056418e-05, - "loss": 0.390842170715332, - "mean_token_accuracy": 0.8634983265399933, - "num_tokens": 4315548.0, - "step": 1800 - }, - { - "entropy": 0.47117673605680466, - "epoch": 4.458383594692401, - "grad_norm": 1.0369561910629272, - "learning_rate": 8.922239856643896e-05, - "loss": 0.39387256622314454, - "mean_token_accuracy": 0.8625931572914124, - "num_tokens": 4431469.0, - "step": 1850 - }, - { - "entropy": 0.479859294295311, - "epoch": 4.579010856453559, - "grad_norm": 1.0827305316925049, - "learning_rate": 8.661827322737095e-05, - "loss": 0.40195709228515625, - "mean_token_accuracy": 0.8612629306316376, - "num_tokens": 4551803.0, - "step": 1900 - }, - { - "entropy": 0.4729380664229393, - "epoch": 4.699638118214716, - "grad_norm": 1.0114246606826782, - "learning_rate": 8.397727702730326e-05, - "loss": 0.39648338317871096, - "mean_token_accuracy": 0.8610477095842362, - "num_tokens": 4674468.0, - "step": 1950 - }, - { - "entropy": 0.4903118047118187, - "epoch": 4.820265379975875, - "grad_norm": 1.0115467309951782, - "learning_rate": 8.130408045471496e-05, - "loss": 0.4067733001708984, - "mean_token_accuracy": 0.8597196912765503, - "num_tokens": 4788027.0, - "step": 2000 - }, - { - "entropy": 0.48099748432636263, - "epoch": 4.940892641737032, - "grad_norm": 1.0201005935668945, - "learning_rate": 7.860341094306242e-05, - "loss": 0.4030665588378906, - "mean_token_accuracy": 0.8603792876005173, - "num_tokens": 4907561.0, - "step": 2050 - }, - { - "epoch": 5.0, - "eval_entropy": 0.5492983292997553, - "eval_mean_token_accuracy": 0.796870938847574, - "eval_not_syn_loss": 0.7129728198051453, - "eval_not_syn_runtime": 53.9575, - "eval_not_syn_samples_per_second": 26.317, - "eval_not_syn_steps_per_second": 3.299, - "eval_num_tokens": 4969030.0, - "step": 2075 - }, - { - "epoch": 5.0, - "eval_entropy": 0.513823946037989, - "eval_mean_token_accuracy": 0.8174765464964877, - "eval_num_tokens": 4969030.0, - "eval_syn_loss": 0.6699439883232117, - "eval_syn_runtime": 56.0033, - "eval_syn_samples_per_second": 25.356, - "eval_syn_steps_per_second": 3.178, - "step": 2075 - }, - { - "entropy": 0.43584728511897003, - "epoch": 5.060313630880579, - "grad_norm": 0.9914600253105164, - "learning_rate": 7.588004451051618e-05, - "loss": 0.34999092102050783, - "mean_token_accuracy": 0.8777085295831314, - "num_tokens": 5026625.0, - "step": 2100 - }, - { - "entropy": 0.38114599585533143, - "epoch": 5.180940892641737, - "grad_norm": 1.232283115386963, - "learning_rate": 7.313879731377764e-05, - "loss": 0.2903956604003906, - "mean_token_accuracy": 0.896172685623169, - "num_tokens": 5141010.0, - "step": 2150 - }, - { - "entropy": 0.3792745009064674, - "epoch": 5.301568154402895, - "grad_norm": 1.2843542098999023, - "learning_rate": 7.03845171309126e-05, - "loss": 0.2903476333618164, - "mean_token_accuracy": 0.8951605755090714, - "num_tokens": 5261240.0, - "step": 2200 - }, - { - "entropy": 0.3744692197442055, - "epoch": 5.422195416164053, - "grad_norm": 1.23075270652771, - "learning_rate": 6.76220747882637e-05, - "loss": 0.29011749267578124, - "mean_token_accuracy": 0.8963996982574463, - "num_tokens": 5382160.0, - "step": 2250 - }, - { - "entropy": 0.3795280712842941, - "epoch": 5.542822677925211, - "grad_norm": 1.4011399745941162, - "learning_rate": 6.485635554660295e-05, - "loss": 0.2921779823303223, - "mean_token_accuracy": 0.8954424083232879, - "num_tokens": 5502002.0, - "step": 2300 - }, - { - "entropy": 0.3797056859731674, - "epoch": 5.66344993968637, - "grad_norm": 1.1022812128067017, - "learning_rate": 6.209225046175775e-05, - "loss": 0.29343433380126954, - "mean_token_accuracy": 0.8939031893014908, - "num_tokens": 5629969.0, - "step": 2350 - }, - { - "entropy": 0.3827402114868164, - "epoch": 5.784077201447527, - "grad_norm": 1.22834312915802, - "learning_rate": 5.9334647734988365e-05, - "loss": 0.2954188537597656, - "mean_token_accuracy": 0.8937842601537704, - "num_tokens": 5750596.0, - "step": 2400 - }, - { - "entropy": 0.38422658771276474, - "epoch": 5.904704463208685, - "grad_norm": 1.425066351890564, - "learning_rate": 5.6588424068413775e-05, - "loss": 0.296480770111084, - "mean_token_accuracy": 0.8936110073328019, - "num_tokens": 5868110.0, - "step": 2450 - }, - { - "epoch": 6.0, - "eval_entropy": 0.49826224684045556, - "eval_mean_token_accuracy": 0.795356297760867, - "eval_not_syn_loss": 0.7700047492980957, - "eval_not_syn_runtime": 53.0632, - "eval_not_syn_samples_per_second": 26.761, - "eval_not_syn_steps_per_second": 3.354, - "eval_num_tokens": 5962836.0, - "step": 2490 - }, - { - "epoch": 6.0, - "eval_entropy": 0.4654735526007213, - "eval_mean_token_accuracy": 0.8118122223387944, - "eval_num_tokens": 5962836.0, - "eval_syn_loss": 0.7288840413093567, - "eval_syn_runtime": 54.6517, - "eval_syn_samples_per_second": 25.983, - "eval_syn_steps_per_second": 3.257, - "step": 2490 - }, - { - "entropy": 0.3596555646320786, - "epoch": 6.024125452352232, - "grad_norm": 1.628747820854187, - "learning_rate": 5.385843604077321e-05, - "loss": 0.2706174850463867, - "mean_token_accuracy": 0.9024625179743526, - "num_tokens": 5990107.0, - "step": 2500 - }, - { - "entropy": 0.285839084982872, - "epoch": 6.144752714113389, - "grad_norm": 1.2864570617675781, - "learning_rate": 5.114951151877489e-05, - "loss": 0.19398983001708983, - "mean_token_accuracy": 0.9306288361549377, - "num_tokens": 6104230.0, - "step": 2550 - }, - { - "entropy": 0.2841903820633888, - "epoch": 6.265379975874548, - "grad_norm": 1.3262356519699097, - "learning_rate": 4.846644111922105e-05, - "loss": 0.19344640731811524, - "mean_token_accuracy": 0.9303907889127732, - "num_tokens": 6223548.0, - "step": 2600 - }, - { - "entropy": 0.2794396734237671, - "epoch": 6.386007237635706, - "grad_norm": 1.216142177581787, - "learning_rate": 4.581396973700743e-05, - "loss": 0.19337059020996095, - "mean_token_accuracy": 0.9294343847036362, - "num_tokens": 6349102.0, - "step": 2650 - }, - { - "entropy": 0.28492832884192465, - "epoch": 6.5066344993968634, - "grad_norm": 1.281471610069275, - "learning_rate": 4.3196788153980425e-05, - "loss": 0.19621110916137696, - "mean_token_accuracy": 0.9284999430179596, - "num_tokens": 6468966.0, - "step": 2700 - }, - { - "entropy": 0.28580656126141546, - "epoch": 6.627261761158022, - "grad_norm": 1.581629753112793, - "learning_rate": 4.061952474349045e-05, - "loss": 0.2017466926574707, - "mean_token_accuracy": 0.9269114434719086, - "num_tokens": 6583296.0, - "step": 2750 - }, - { - "entropy": 0.2819211499392986, - "epoch": 6.74788902291918, - "grad_norm": 1.111789584159851, - "learning_rate": 3.8086737285312406e-05, - "loss": 0.19925508499145508, - "mean_token_accuracy": 0.9279456728696823, - "num_tokens": 6698696.0, - "step": 2800 - }, - { - "entropy": 0.27608554884791375, - "epoch": 6.868516284680338, - "grad_norm": 1.337012529373169, - "learning_rate": 3.560290490540746e-05, - "loss": 0.19468421936035157, - "mean_token_accuracy": 0.9297049742937088, - "num_tokens": 6819711.0, - "step": 2850 - }, - { - "entropy": 0.27499831572175026, - "epoch": 6.989143546441496, - "grad_norm": 1.2096495628356934, - "learning_rate": 3.317242015478115e-05, - "loss": 0.19465036392211915, - "mean_token_accuracy": 0.929957520365715, - "num_tokens": 6944941.0, - "step": 2900 - }, - { - "epoch": 7.0, - "eval_entropy": 0.43061937626158253, - "eval_mean_token_accuracy": 0.7864309693320414, - "eval_not_syn_loss": 0.8960050940513611, - "eval_not_syn_runtime": 53.715, - "eval_not_syn_samples_per_second": 26.436, - "eval_not_syn_steps_per_second": 3.314, - "eval_num_tokens": 6956642.0, - "step": 2905 - }, - { - "epoch": 7.0, - "eval_entropy": 0.40260162983047826, - "eval_mean_token_accuracy": 0.81038418077351, - "eval_num_tokens": 6956642.0, - "eval_syn_loss": 0.8312228322029114, - "eval_syn_runtime": 55.7344, - "eval_syn_samples_per_second": 25.478, - "eval_syn_steps_per_second": 3.194, - "step": 2905 - } - ], - "logging_steps": 50, - "max_steps": 4150, - "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": 1.4341938530883034e+17, - "train_batch_size": 8, - "trial_name": null, - "trial_params": null -} diff --git a/substitutivity_original_Estonian/Qwen3.5-2B-Base_substitutivity_splits_original_features_train_substitutivity_splits_original_features_test2/checkpoint-2905/training_args.bin b/substitutivity_original_Estonian/Qwen3.5-2B-Base_substitutivity_splits_original_features_train_substitutivity_splits_original_features_test2/checkpoint-2905/training_args.bin deleted file mode 100644 index f3d8098f65400413dd700f32ada5649470871017..0000000000000000000000000000000000000000 --- a/substitutivity_original_Estonian/Qwen3.5-2B-Base_substitutivity_splits_original_features_train_substitutivity_splits_original_features_test2/checkpoint-2905/training_args.bin +++ /dev/null @@ -1,3 +0,0 @@ -version https://git-lfs.github.com/spec/v1 -oid sha256:617af0f41414927baaaf158949585bb75eb4b05ff78baa5216c860ef0b431009 -size 6033 diff --git a/substitutivity_original_Estonian/Qwen3.5-2B-Base_substitutivity_splits_original_features_train_substitutivity_splits_original_features_test2/checkpoint-3320/README.md b/substitutivity_original_Estonian/Qwen3.5-2B-Base_substitutivity_splits_original_features_train_substitutivity_splits_original_features_test2/checkpoint-3320/README.md deleted file mode 100644 index ece204d005f7bf161d0e448efdd56fd297b7c8a2..0000000000000000000000000000000000000000 --- a/substitutivity_original_Estonian/Qwen3.5-2B-Base_substitutivity_splits_original_features_train_substitutivity_splits_original_features_test2/checkpoint-3320/README.md +++ /dev/null @@ -1,209 +0,0 @@ ---- -base_model: Qwen/Qwen3.5-2B-Base -library_name: peft -pipeline_tag: text-generation -tags: -- base_model:adapter:Qwen/Qwen3.5-2B-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/substitutivity_original_Estonian/Qwen3.5-2B-Base_substitutivity_splits_original_features_train_substitutivity_splits_original_features_test2/checkpoint-3320/adapter_config.json b/substitutivity_original_Estonian/Qwen3.5-2B-Base_substitutivity_splits_original_features_train_substitutivity_splits_original_features_test2/checkpoint-3320/adapter_config.json deleted file mode 100644 index 6a095bb3bfe249ce8e15df6ae999e88fa1e17162..0000000000000000000000000000000000000000 --- a/substitutivity_original_Estonian/Qwen3.5-2B-Base_substitutivity_splits_original_features_train_substitutivity_splits_original_features_test2/checkpoint-3320/adapter_config.json +++ /dev/null @@ -1,46 +0,0 @@ -{ - "alora_invocation_tokens": null, - "alpha_pattern": {}, - "arrow_config": null, - "auto_mapping": null, - "base_model_name_or_path": "Qwen/Qwen3.5-2B-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": 0.05731002492184622, - "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": [ - "o_proj", - "q_proj", - "k_proj", - "down_proj", - "up_proj", - "v_proj", - "gate_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/substitutivity_original_Estonian/Qwen3.5-2B-Base_substitutivity_splits_original_features_train_substitutivity_splits_original_features_test2/checkpoint-3320/adapter_model.safetensors b/substitutivity_original_Estonian/Qwen3.5-2B-Base_substitutivity_splits_original_features_train_substitutivity_splits_original_features_test2/checkpoint-3320/adapter_model.safetensors deleted file mode 100644 index f48ddc0947f382592b044403000c3a7f4d59386f..0000000000000000000000000000000000000000 --- a/substitutivity_original_Estonian/Qwen3.5-2B-Base_substitutivity_splits_original_features_train_substitutivity_splits_original_features_test2/checkpoint-3320/adapter_model.safetensors +++ /dev/null @@ -1,3 +0,0 @@ -version https://git-lfs.github.com/spec/v1 -oid sha256:42842aa833db0d220d1cf385a0ba8349249989f4a4125924d052d2438383ecf2 -size 174613392 diff --git a/substitutivity_original_Estonian/Qwen3.5-2B-Base_substitutivity_splits_original_features_train_substitutivity_splits_original_features_test2/checkpoint-3320/chat_template.jinja b/substitutivity_original_Estonian/Qwen3.5-2B-Base_substitutivity_splits_original_features_train_substitutivity_splits_original_features_test2/checkpoint-3320/chat_template.jinja deleted file mode 100644 index 0ef09f214eaa6d9bca297988afc1454b5827b2c7..0000000000000000000000000000000000000000 --- a/substitutivity_original_Estonian/Qwen3.5-2B-Base_substitutivity_splits_original_features_train_substitutivity_splits_original_features_test2/checkpoint-3320/chat_template.jinja +++ /dev/null @@ -1,154 +0,0 @@ -{%- set image_count = namespace(value=0) %} -{%- set video_count = namespace(value=0) %} -{%- macro render_content(content, do_vision_count, is_system_content=false) %} - {%- if content is string %} - {{- content }} - {%- elif content is iterable and content is not mapping %} - {%- for item in content %} - {%- if 'image' in item or 'image_url' in item or item.type == 'image' %} - {%- if is_system_content %} - {{- raise_exception('System message cannot contain images.') }} - {%- endif %} - {%- if do_vision_count %} - {%- set image_count.value = image_count.value + 1 %} - {%- endif %} - {%- if add_vision_id %} - {{- 'Picture ' ~ image_count.value ~ ': ' }} - {%- endif %} - {{- '<|vision_start|><|image_pad|><|vision_end|>' }} - {%- elif 'video' in item or item.type == 'video' %} - {%- if is_system_content %} - {{- raise_exception('System message cannot contain videos.') }} - {%- endif %} - {%- if do_vision_count %} - {%- set video_count.value = video_count.value + 1 %} - {%- endif %} - {%- if add_vision_id %} - {{- 'Video ' ~ video_count.value ~ ': ' }} - {%- endif %} - {{- '<|vision_start|><|video_pad|><|vision_end|>' }} - {%- elif 'text' in item %} - {{- item.text }} - {%- else %} - {{- raise_exception('Unexpected item type in content.') }} - {%- endif %} - {%- endfor %} - {%- elif content is none or content is undefined %} - {{- '' }} - {%- else %} - {{- raise_exception('Unexpected content type.') }} - {%- endif %} -{%- endmacro %} -{%- if not messages %} - {{- raise_exception('No messages provided.') }} -{%- endif %} -{%- if tools and tools is iterable and tools is not mapping %} - {{- '<|im_start|>system\n' }} - {{- "# Tools\n\nYou have access to the following functions:\n\n" }} - {%- for tool in tools %} - {{- "\n" }} - {{- tool | tojson }} - {%- endfor %} - {{- "\n" }} - {{- '\n\nIf you choose to call a function ONLY reply in the following format with NO suffix:\n\n\n\n\nvalue_1\n\n\nThis is the value for the second parameter\nthat can span\nmultiple lines\n\n\n\n\n\nReminder:\n- Function calls MUST follow the specified format: an inner block must be nested within XML tags\n- Required parameters MUST be specified\n- You may provide optional reasoning for your function call in natural language BEFORE the function call, but NOT after\n- If there is no function call available, answer the question like normal with your current knowledge and do not tell the user about function calls\n' }} - {%- if messages[0].role == 'system' %} - {%- set content = render_content(messages[0].content, false, true)|trim %} - {%- if content %} - {{- '\n\n' + content }} - {%- endif %} - {%- endif %} - {{- '<|im_end|>\n' }} -{%- else %} - {%- if messages[0].role == 'system' %} - {%- set content = render_content(messages[0].content, false, true)|trim %} - {{- '<|im_start|>system\n' + 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" %} - {%- set content = render_content(message.content, false)|trim %} - {%- if not(content.startswith('') and content.endswith('')) %} - {%- set ns.multi_step_tool = false %} - {%- set ns.last_query_index = index %} - {%- endif %} - {%- endif %} -{%- endfor %} -{%- if ns.multi_step_tool %} - {{- raise_exception('No user query found in messages.') }} -{%- endif %} -{%- for message in messages %} - {%- set content = render_content(message.content, true)|trim %} - {%- if message.role == "system" %} - {%- if not loop.first %} - {{- raise_exception('System message must be at the beginning.') }} - {%- endif %} - {%- elif message.role == "user" %} - {{- '<|im_start|>' + message.role + '\n' + content + '<|im_end|>' + '\n' }} - {%- elif message.role == "assistant" %} - {%- set reasoning_content = '' %} - {%- if message.reasoning_content is string %} - {%- set reasoning_content = message.reasoning_content %} - {%- else %} - {%- if '' in content %} - {%- set reasoning_content = content.split('')[0].rstrip('\n').split('')[-1].lstrip('\n') %} - {%- set content = content.split('')[-1].lstrip('\n') %} - {%- endif %} - {%- endif %} - {%- set reasoning_content = reasoning_content|trim %} - {%- if loop.index0 > ns.last_query_index %} - {{- '<|im_start|>' + message.role + '\n\n' + reasoning_content + '\n\n\n' + content }} - {%- else %} - {{- '<|im_start|>' + message.role + '\n' + content }} - {%- endif %} - {%- if message.tool_calls and message.tool_calls is iterable and message.tool_calls is not mapping %} - {%- for tool_call in message.tool_calls %} - {%- if tool_call.function is defined %} - {%- set tool_call = tool_call.function %} - {%- endif %} - {%- if loop.first %} - {%- if content|trim %} - {{- '\n\n\n\n' }} - {%- else %} - {{- '\n\n' }} - {%- endif %} - {%- else %} - {{- '\n\n\n' }} - {%- endif %} - {%- if tool_call.arguments is defined %} - {%- for args_name, args_value in tool_call.arguments|items %} - {{- '\n' }} - {%- set args_value = args_value | tojson | safe if args_value is mapping or (args_value is sequence and args_value is not string) else args_value | string %} - {{- args_value }} - {{- '\n\n' }} - {%- endfor %} - {%- endif %} - {{- '\n' }} - {%- endfor %} - {%- endif %} - {{- '<|im_end|>\n' }} - {%- elif message.role == "tool" %} - {%- if loop.previtem and loop.previtem.role != "tool" %} - {{- '<|im_start|>user' }} - {%- endif %} - {{- '\n\n' }} - {{- content }} - {{- '\n' }} - {%- if not loop.last and loop.nextitem.role != "tool" %} - {{- '<|im_end|>\n' }} - {%- elif loop.last %} - {{- '<|im_end|>\n' }} - {%- endif %} - {%- else %} - {{- raise_exception('Unexpected message role.') }} - {%- endif %} -{%- endfor %} -{%- if add_generation_prompt %} - {{- '<|im_start|>assistant\n' }} - {%- if enable_thinking is defined and enable_thinking is true %} - {{- '\n' }} - {%- else %} - {{- '\n\n\n\n' }} - {%- endif %} -{%- endif %} \ No newline at end of file diff --git a/substitutivity_original_Estonian/Qwen3.5-2B-Base_substitutivity_splits_original_features_train_substitutivity_splits_original_features_test2/checkpoint-3320/tokenizer.json b/substitutivity_original_Estonian/Qwen3.5-2B-Base_substitutivity_splits_original_features_train_substitutivity_splits_original_features_test2/checkpoint-3320/tokenizer.json deleted file mode 100644 index 67741b04f23bfdb46501f748ce27865ec82eccfb..0000000000000000000000000000000000000000 --- a/substitutivity_original_Estonian/Qwen3.5-2B-Base_substitutivity_splits_original_features_train_substitutivity_splits_original_features_test2/checkpoint-3320/tokenizer.json +++ /dev/null @@ -1,3 +0,0 @@ -version https://git-lfs.github.com/spec/v1 -oid sha256:87a7830d63fcf43bf241c3c5242e96e62dd3fdc29224ca26fed8ea333db72de4 -size 19989343 diff --git a/substitutivity_original_Estonian/Qwen3.5-2B-Base_substitutivity_splits_original_features_train_substitutivity_splits_original_features_test2/checkpoint-3320/tokenizer_config.json b/substitutivity_original_Estonian/Qwen3.5-2B-Base_substitutivity_splits_original_features_train_substitutivity_splits_original_features_test2/checkpoint-3320/tokenizer_config.json deleted file mode 100644 index b4a37b2a6fd3ab3317cd7bac72855be1a843b2bb..0000000000000000000000000000000000000000 --- a/substitutivity_original_Estonian/Qwen3.5-2B-Base_substitutivity_splits_original_features_train_substitutivity_splits_original_features_test2/checkpoint-3320/tokenizer_config.json +++ /dev/null @@ -1,31 +0,0 @@ -{ - "add_prefix_space": false, - "audio_bos_token": "<|audio_start|>", - "audio_eos_token": "<|audio_end|>", - "audio_token": "<|audio_pad|>", - "backend": "tokenizers", - "bos_token": null, - "clean_up_tokenization_spaces": false, - "eos_token": "<|endoftext|>", - "errors": "replace", - "image_token": "<|image_pad|>", - "is_local": false, - "model_max_length": 262144, - "model_specific_special_tokens": { - "audio_bos_token": "<|audio_start|>", - "audio_eos_token": "<|audio_end|>", - "audio_token": "<|audio_pad|>", - "image_token": "<|image_pad|>", - "video_token": "<|video_pad|>", - "vision_bos_token": "<|vision_start|>", - "vision_eos_token": "<|vision_end|>" - }, - "pad_token": "<|endoftext|>", - "pretokenize_regex": "(?i:'s|'t|'re|'ve|'m|'ll|'d)|[^\\r\\n\\p{L}\\p{N}]?[\\p{L}\\p{M}]+|\\p{N}| ?[^\\s\\p{L}\\p{M}\\p{N}]+[\\r\\n]*|\\s*[\\r\\n]+|\\s+(?!\\S)|\\s+", - "split_special_tokens": false, - "tokenizer_class": "TokenizersBackend", - "unk_token": null, - "video_token": "<|video_pad|>", - "vision_bos_token": "<|vision_start|>", - "vision_eos_token": "<|vision_end|>" -} diff --git a/substitutivity_original_Estonian/Qwen3.5-2B-Base_substitutivity_splits_original_features_train_substitutivity_splits_original_features_test2/checkpoint-3320/trainer_state.json b/substitutivity_original_Estonian/Qwen3.5-2B-Base_substitutivity_splits_original_features_train_substitutivity_splits_original_features_test2/checkpoint-3320/trainer_state.json deleted file mode 100644 index 1c621608f492330d28dc65dfdad21c8485a4bbac..0000000000000000000000000000000000000000 --- a/substitutivity_original_Estonian/Qwen3.5-2B-Base_substitutivity_splits_original_features_train_substitutivity_splits_original_features_test2/checkpoint-3320/trainer_state.json +++ /dev/null @@ -1,870 +0,0 @@ -{ - "best_global_step": null, - "best_metric": null, - "best_model_checkpoint": null, - "epoch": 8.0, - "eval_steps": 500, - "global_step": 3320, - "is_hyper_param_search": false, - "is_local_process_zero": true, - "is_world_process_zero": true, - "log_history": [ - { - "entropy": 2.2933380508422854, - "epoch": 0.12062726176115803, - "grad_norm": 2.8075337409973145, - "learning_rate": 1.5531017374629602e-05, - "loss": 2.1429234313964844, - "mean_token_accuracy": 0.5779684272408485, - "num_tokens": 121428.0, - "step": 50 - }, - { - "entropy": 1.1315062403678895, - "epoch": 0.24125452352231605, - "grad_norm": 1.9415371417999268, - "learning_rate": 3.1378994287516954e-05, - "loss": 1.0461698913574218, - "mean_token_accuracy": 0.7166991305351257, - "num_tokens": 243966.0, - "step": 100 - }, - { - "entropy": 0.9537658470869065, - "epoch": 0.3618817852834741, - "grad_norm": 1.8353157043457031, - "learning_rate": 4.72269712004043e-05, - "loss": 0.8774301147460938, - "mean_token_accuracy": 0.7506789201498032, - "num_tokens": 366649.0, - "step": 150 - }, - { - "entropy": 0.8955666983127594, - "epoch": 0.4825090470446321, - "grad_norm": 1.5419880151748657, - "learning_rate": 6.307494811329165e-05, - "loss": 0.8234302520751953, - "mean_token_accuracy": 0.7616239994764328, - "num_tokens": 488367.0, - "step": 200 - }, - { - "entropy": 0.8689469587802887, - "epoch": 0.6031363088057901, - "grad_norm": 1.5661484003067017, - "learning_rate": 7.8922925026179e-05, - "loss": 0.7961890411376953, - "mean_token_accuracy": 0.767715499997139, - "num_tokens": 603131.0, - "step": 250 - }, - { - "entropy": 0.8425652319192887, - "epoch": 0.7237635705669482, - "grad_norm": 1.5694905519485474, - "learning_rate": 9.477090193906636e-05, - "loss": 0.7718594360351563, - "mean_token_accuracy": 0.7736722755432129, - "num_tokens": 726412.0, - "step": 300 - }, - { - "entropy": 0.843294187784195, - "epoch": 0.8443908323281062, - "grad_norm": 1.1118028163909912, - "learning_rate": 0.00011061887885195371, - "loss": 0.7801297760009765, - "mean_token_accuracy": 0.7714760398864746, - "num_tokens": 846516.0, - "step": 350 - }, - { - "entropy": 0.8213214933872223, - "epoch": 0.9650180940892642, - "grad_norm": 0.9202918410301208, - "learning_rate": 0.00012646685576484106, - "loss": 0.7514453887939453, - "mean_token_accuracy": 0.7774044269323349, - "num_tokens": 958978.0, - "step": 400 - }, - { - "epoch": 1.0, - "eval_entropy": 0.818668900915746, - "eval_mean_token_accuracy": 0.7533496866065464, - "eval_not_syn_loss": 0.7798981070518494, - "eval_not_syn_runtime": 53.8744, - "eval_not_syn_samples_per_second": 26.358, - "eval_not_syn_steps_per_second": 3.304, - "eval_num_tokens": 993806.0, - "step": 415 - }, - { - "epoch": 1.0, - "eval_entropy": 0.7627468125873738, - "eval_mean_token_accuracy": 0.8024111574285486, - "eval_num_tokens": 993806.0, - "eval_syn_loss": 0.7218216061592102, - "eval_syn_runtime": 55.3239, - "eval_syn_samples_per_second": 25.667, - "eval_syn_steps_per_second": 3.217, - "step": 415 - }, - { - "entropy": 0.7802230679627621, - "epoch": 1.0844390832328106, - "grad_norm": 1.0145268440246582, - "learning_rate": 0.00013151131541121804, - "loss": 0.7111608123779297, - "mean_token_accuracy": 0.7845737542768921, - "num_tokens": 1080301.0, - "step": 450 - }, - { - "entropy": 0.7558828103542328, - "epoch": 1.2050663449939687, - "grad_norm": 1.0477591753005981, - "learning_rate": 0.0001313741160170038, - "loss": 0.6933037567138672, - "mean_token_accuracy": 0.7889221370220184, - "num_tokens": 1201215.0, - "step": 500 - }, - { - "entropy": 0.7593123841285706, - "epoch": 1.3256936067551266, - "grad_norm": 0.899161159992218, - "learning_rate": 0.00013112089697387701, - "loss": 0.6962493133544921, - "mean_token_accuracy": 0.7877359414100646, - "num_tokens": 1322895.0, - "step": 550 - }, - { - "entropy": 0.7460986822843552, - "epoch": 1.4463208685162847, - "grad_norm": 1.1061116456985474, - "learning_rate": 0.00013075210608885215, - "loss": 0.6817412567138672, - "mean_token_accuracy": 0.7904874366521836, - "num_tokens": 1436749.0, - "step": 600 - }, - { - "entropy": 0.735516265630722, - "epoch": 1.5669481302774426, - "grad_norm": 0.9925405979156494, - "learning_rate": 0.00013026839555278796, - "loss": 0.6690705871582031, - "mean_token_accuracy": 0.7930250608921051, - "num_tokens": 1555400.0, - "step": 650 - }, - { - "entropy": 0.7302519899606704, - "epoch": 1.6875753920386007, - "grad_norm": 0.800025463104248, - "learning_rate": 0.0001296706207870159, - "loss": 0.6714833831787109, - "mean_token_accuracy": 0.7946307575702667, - "num_tokens": 1676892.0, - "step": 700 - }, - { - "entropy": 0.7321263110637665, - "epoch": 1.8082026537997589, - "grad_norm": 1.0500682592391968, - "learning_rate": 0.00012895983893056423, - "loss": 0.6672868347167968, - "mean_token_accuracy": 0.7947645837068558, - "num_tokens": 1802689.0, - "step": 750 - }, - { - "entropy": 0.7075755137205124, - "epoch": 1.9288299155609168, - "grad_norm": 0.9248411655426025, - "learning_rate": 0.00012813730697065298, - "loss": 0.6466568756103516, - "mean_token_accuracy": 0.798747689127922, - "num_tokens": 1920804.0, - "step": 800 - }, - { - "epoch": 2.0, - "eval_entropy": 0.7426781811740961, - "eval_mean_token_accuracy": 0.7693507239389955, - "eval_not_syn_loss": 0.7139139175415039, - "eval_not_syn_runtime": 54.4286, - "eval_not_syn_samples_per_second": 26.089, - "eval_not_syn_steps_per_second": 3.27, - "eval_num_tokens": 1987612.0, - "step": 830 - }, - { - "epoch": 2.0, - "eval_entropy": 0.6937930352902144, - "eval_mean_token_accuracy": 0.8180048633157537, - "eval_num_tokens": 1987612.0, - "eval_syn_loss": 0.6489660143852234, - "eval_syn_runtime": 56.1525, - "eval_syn_samples_per_second": 25.288, - "eval_syn_steps_per_second": 3.17, - "step": 830 - }, - { - "entropy": 0.6853217440422135, - "epoch": 2.0482509047044632, - "grad_norm": 0.6932648420333862, - "learning_rate": 0.00012720447951976603, - "loss": 0.6212613296508789, - "mean_token_accuracy": 0.8047762579388089, - "num_tokens": 2039660.0, - "step": 850 - }, - { - "entropy": 0.6459032052755356, - "epoch": 2.168878166465621, - "grad_norm": 1.0688549280166626, - "learning_rate": 0.00012616300624323124, - "loss": 0.5740525817871094, - "mean_token_accuracy": 0.8148404061794281, - "num_tokens": 2161772.0, - "step": 900 - }, - { - "entropy": 0.6463843190670013, - "epoch": 2.2895054282267795, - "grad_norm": 0.847271203994751, - "learning_rate": 0.0001250147289418581, - "loss": 0.577717170715332, - "mean_token_accuracy": 0.814143658876419, - "num_tokens": 2280833.0, - "step": 950 - }, - { - "entropy": 0.6461306965351105, - "epoch": 2.4101326899879374, - "grad_norm": 0.9998957514762878, - "learning_rate": 0.00012376167829479217, - "loss": 0.5813011932373047, - "mean_token_accuracy": 0.8144401633739471, - "num_tokens": 2398206.0, - "step": 1000 - }, - { - "entropy": 0.640914659500122, - "epoch": 2.5307599517490953, - "grad_norm": 0.8692949414253235, - "learning_rate": 0.0001224060702683459, - "loss": 0.5755012512207032, - "mean_token_accuracy": 0.8168018108606339, - "num_tokens": 2518888.0, - "step": 1050 - }, - { - "entropy": 0.6421274638175964, - "epoch": 2.651387213510253, - "grad_norm": 0.8920196294784546, - "learning_rate": 0.00012095030219715738, - "loss": 0.5785483932495117, - "mean_token_accuracy": 0.8137972563505173, - "num_tokens": 2639763.0, - "step": 1100 - }, - { - "entropy": 0.6466812592744827, - "epoch": 2.772014475271411, - "grad_norm": 0.6876946687698364, - "learning_rate": 0.00011939694854460696, - "loss": 0.5816070938110351, - "mean_token_accuracy": 0.8145407867431641, - "num_tokens": 2755540.0, - "step": 1150 - }, - { - "entropy": 0.6378493067622185, - "epoch": 2.8926417370325694, - "grad_norm": 0.7532753944396973, - "learning_rate": 0.00011774875634998911, - "loss": 0.5754346084594727, - "mean_token_accuracy": 0.8148861181735992, - "num_tokens": 2874118.0, - "step": 1200 - }, - { - "epoch": 3.0, - "eval_entropy": 0.6912222209941136, - "eval_mean_token_accuracy": 0.7868467750174276, - "eval_not_syn_loss": 0.6782422661781311, - "eval_not_syn_runtime": 53.2627, - "eval_not_syn_samples_per_second": 26.66, - "eval_not_syn_steps_per_second": 3.342, - "eval_num_tokens": 2981418.0, - "step": 1245 - }, - { - "epoch": 3.0, - "eval_entropy": 0.6467212934172555, - "eval_mean_token_accuracy": 0.8215091817834405, - "eval_num_tokens": 2981418.0, - "eval_syn_loss": 0.6301999092102051, - "eval_syn_runtime": 55.0115, - "eval_syn_samples_per_second": 25.813, - "eval_syn_steps_per_second": 3.236, - "step": 1245 - }, - { - "entropy": 0.6316030452949832, - "epoch": 3.012062726176116, - "grad_norm": 0.7761592864990234, - "learning_rate": 0.00011600864037049124, - "loss": 0.5697880935668945, - "mean_token_accuracy": 0.8159356400220081, - "num_tokens": 2994602.0, - "step": 1250 - }, - { - "entropy": 0.5467158386111259, - "epoch": 3.132689987937274, - "grad_norm": 0.8169087767601013, - "learning_rate": 0.00011417967792657078, - "loss": 0.47844127655029295, - "mean_token_accuracy": 0.8378052765130997, - "num_tokens": 3121094.0, - "step": 1300 - }, - { - "entropy": 0.5454566389322281, - "epoch": 3.2533172496984317, - "grad_norm": 0.8219636678695679, - "learning_rate": 0.0001122651034598459, - "loss": 0.4839338302612305, - "mean_token_accuracy": 0.8385071897506714, - "num_tokens": 3248038.0, - "step": 1350 - }, - { - "entropy": 0.5598723018169403, - "epoch": 3.37394451145959, - "grad_norm": 0.9956360459327698, - "learning_rate": 0.00011026830281312423, - "loss": 0.492743034362793, - "mean_token_accuracy": 0.8357370454072952, - "num_tokens": 3363487.0, - "step": 1400 - }, - { - "entropy": 0.5479365903139114, - "epoch": 3.494571773220748, - "grad_norm": 0.8586823344230652, - "learning_rate": 0.00010819280724268524, - "loss": 0.4810820770263672, - "mean_token_accuracy": 0.837928666472435, - "num_tokens": 3480513.0, - "step": 1450 - }, - { - "entropy": 0.5619524419307709, - "epoch": 3.615199034981906, - "grad_norm": 1.0894954204559326, - "learning_rate": 0.0001060422871734048, - "loss": 0.4974747085571289, - "mean_token_accuracy": 0.8349784809350967, - "num_tokens": 3597183.0, - "step": 1500 - }, - { - "entropy": 0.5588027596473694, - "epoch": 3.7358262967430638, - "grad_norm": 0.9905257225036621, - "learning_rate": 0.00010382054570776654, - "loss": 0.49687202453613283, - "mean_token_accuracy": 0.834458841085434, - "num_tokens": 3714173.0, - "step": 1550 - }, - { - "entropy": 0.5654991599917412, - "epoch": 3.856453558504222, - "grad_norm": 1.0116610527038574, - "learning_rate": 0.00010153151190023791, - "loss": 0.5016670989990234, - "mean_token_accuracy": 0.8326652735471726, - "num_tokens": 3832652.0, - "step": 1600 - }, - { - "entropy": 0.5605382058024406, - "epoch": 3.97708082026538, - "grad_norm": 0.8011100888252258, - "learning_rate": 9.917923380890624e-05, - "loss": 0.49551734924316404, - "mean_token_accuracy": 0.8343535929918289, - "num_tokens": 3955167.0, - "step": 1650 - }, - { - "epoch": 4.0, - "eval_entropy": 0.5819476859623127, - "eval_mean_token_accuracy": 0.7963069781158747, - "eval_not_syn_loss": 0.6924306750297546, - "eval_not_syn_runtime": 54.027, - "eval_not_syn_samples_per_second": 26.283, - "eval_not_syn_steps_per_second": 3.295, - "eval_num_tokens": 3975224.0, - "step": 1660 - }, - { - "epoch": 4.0, - "eval_entropy": 0.545470636212424, - "eval_mean_token_accuracy": 0.8127643666910321, - "eval_num_tokens": 3975224.0, - "eval_syn_loss": 0.649576723575592, - "eval_syn_runtime": 55.2296, - "eval_syn_samples_per_second": 25.711, - "eval_syn_steps_per_second": 3.223, - "step": 1660 - }, - { - "entropy": 0.4920780294471317, - "epoch": 4.0965018094089265, - "grad_norm": 0.8362822532653809, - "learning_rate": 9.676787133666148e-05, - "loss": 0.4147996139526367, - "mean_token_accuracy": 0.8581844628459275, - "num_tokens": 4072361.0, - "step": 1700 - }, - { - "entropy": 0.4675139603018761, - "epoch": 4.217129071170085, - "grad_norm": 0.9920316338539124, - "learning_rate": 9.430168887458645e-05, - "loss": 0.3863528823852539, - "mean_token_accuracy": 0.865827117562294, - "num_tokens": 4194978.0, - "step": 1750 - }, - { - "entropy": 0.47163557589054106, - "epoch": 4.337756332931242, - "grad_norm": 1.0789282321929932, - "learning_rate": 9.178504776056418e-05, - "loss": 0.390842170715332, - "mean_token_accuracy": 0.8634983265399933, - "num_tokens": 4315548.0, - "step": 1800 - }, - { - "entropy": 0.47117673605680466, - "epoch": 4.458383594692401, - "grad_norm": 1.0369561910629272, - "learning_rate": 8.922239856643896e-05, - "loss": 0.39387256622314454, - "mean_token_accuracy": 0.8625931572914124, - "num_tokens": 4431469.0, - "step": 1850 - }, - { - "entropy": 0.479859294295311, - "epoch": 4.579010856453559, - "grad_norm": 1.0827305316925049, - "learning_rate": 8.661827322737095e-05, - "loss": 0.40195709228515625, - "mean_token_accuracy": 0.8612629306316376, - "num_tokens": 4551803.0, - "step": 1900 - }, - { - "entropy": 0.4729380664229393, - "epoch": 4.699638118214716, - "grad_norm": 1.0114246606826782, - "learning_rate": 8.397727702730326e-05, - "loss": 0.39648338317871096, - "mean_token_accuracy": 0.8610477095842362, - "num_tokens": 4674468.0, - "step": 1950 - }, - { - "entropy": 0.4903118047118187, - "epoch": 4.820265379975875, - "grad_norm": 1.0115467309951782, - "learning_rate": 8.130408045471496e-05, - "loss": 0.4067733001708984, - "mean_token_accuracy": 0.8597196912765503, - "num_tokens": 4788027.0, - "step": 2000 - }, - { - "entropy": 0.48099748432636263, - "epoch": 4.940892641737032, - "grad_norm": 1.0201005935668945, - "learning_rate": 7.860341094306242e-05, - "loss": 0.4030665588378906, - "mean_token_accuracy": 0.8603792876005173, - "num_tokens": 4907561.0, - "step": 2050 - }, - { - "epoch": 5.0, - "eval_entropy": 0.5492983292997553, - "eval_mean_token_accuracy": 0.796870938847574, - "eval_not_syn_loss": 0.7129728198051453, - "eval_not_syn_runtime": 53.9575, - "eval_not_syn_samples_per_second": 26.317, - "eval_not_syn_steps_per_second": 3.299, - "eval_num_tokens": 4969030.0, - "step": 2075 - }, - { - "epoch": 5.0, - "eval_entropy": 0.513823946037989, - "eval_mean_token_accuracy": 0.8174765464964877, - "eval_num_tokens": 4969030.0, - "eval_syn_loss": 0.6699439883232117, - "eval_syn_runtime": 56.0033, - "eval_syn_samples_per_second": 25.356, - "eval_syn_steps_per_second": 3.178, - "step": 2075 - }, - { - "entropy": 0.43584728511897003, - "epoch": 5.060313630880579, - "grad_norm": 0.9914600253105164, - "learning_rate": 7.588004451051618e-05, - "loss": 0.34999092102050783, - "mean_token_accuracy": 0.8777085295831314, - "num_tokens": 5026625.0, - "step": 2100 - }, - { - "entropy": 0.38114599585533143, - "epoch": 5.180940892641737, - "grad_norm": 1.232283115386963, - "learning_rate": 7.313879731377764e-05, - "loss": 0.2903956604003906, - "mean_token_accuracy": 0.896172685623169, - "num_tokens": 5141010.0, - "step": 2150 - }, - { - "entropy": 0.3792745009064674, - "epoch": 5.301568154402895, - "grad_norm": 1.2843542098999023, - "learning_rate": 7.03845171309126e-05, - "loss": 0.2903476333618164, - "mean_token_accuracy": 0.8951605755090714, - "num_tokens": 5261240.0, - "step": 2200 - }, - { - "entropy": 0.3744692197442055, - "epoch": 5.422195416164053, - "grad_norm": 1.23075270652771, - "learning_rate": 6.76220747882637e-05, - "loss": 0.29011749267578124, - "mean_token_accuracy": 0.8963996982574463, - "num_tokens": 5382160.0, - "step": 2250 - }, - { - "entropy": 0.3795280712842941, - "epoch": 5.542822677925211, - "grad_norm": 1.4011399745941162, - "learning_rate": 6.485635554660295e-05, - "loss": 0.2921779823303223, - "mean_token_accuracy": 0.8954424083232879, - "num_tokens": 5502002.0, - "step": 2300 - }, - { - "entropy": 0.3797056859731674, - "epoch": 5.66344993968637, - "grad_norm": 1.1022812128067017, - "learning_rate": 6.209225046175775e-05, - "loss": 0.29343433380126954, - "mean_token_accuracy": 0.8939031893014908, - "num_tokens": 5629969.0, - "step": 2350 - }, - { - "entropy": 0.3827402114868164, - "epoch": 5.784077201447527, - "grad_norm": 1.22834312915802, - "learning_rate": 5.9334647734988365e-05, - "loss": 0.2954188537597656, - "mean_token_accuracy": 0.8937842601537704, - "num_tokens": 5750596.0, - "step": 2400 - }, - { - "entropy": 0.38422658771276474, - "epoch": 5.904704463208685, - "grad_norm": 1.425066351890564, - "learning_rate": 5.6588424068413775e-05, - "loss": 0.296480770111084, - "mean_token_accuracy": 0.8936110073328019, - "num_tokens": 5868110.0, - "step": 2450 - }, - { - "epoch": 6.0, - "eval_entropy": 0.49826224684045556, - "eval_mean_token_accuracy": 0.795356297760867, - "eval_not_syn_loss": 0.7700047492980957, - "eval_not_syn_runtime": 53.0632, - "eval_not_syn_samples_per_second": 26.761, - "eval_not_syn_steps_per_second": 3.354, - "eval_num_tokens": 5962836.0, - "step": 2490 - }, - { - "epoch": 6.0, - "eval_entropy": 0.4654735526007213, - "eval_mean_token_accuracy": 0.8118122223387944, - "eval_num_tokens": 5962836.0, - "eval_syn_loss": 0.7288840413093567, - "eval_syn_runtime": 54.6517, - "eval_syn_samples_per_second": 25.983, - "eval_syn_steps_per_second": 3.257, - "step": 2490 - }, - { - "entropy": 0.3596555646320786, - "epoch": 6.024125452352232, - "grad_norm": 1.628747820854187, - "learning_rate": 5.385843604077321e-05, - "loss": 0.2706174850463867, - "mean_token_accuracy": 0.9024625179743526, - "num_tokens": 5990107.0, - "step": 2500 - }, - { - "entropy": 0.285839084982872, - "epoch": 6.144752714113389, - "grad_norm": 1.2864570617675781, - "learning_rate": 5.114951151877489e-05, - "loss": 0.19398983001708983, - "mean_token_accuracy": 0.9306288361549377, - "num_tokens": 6104230.0, - "step": 2550 - }, - { - "entropy": 0.2841903820633888, - "epoch": 6.265379975874548, - "grad_norm": 1.3262356519699097, - "learning_rate": 4.846644111922105e-05, - "loss": 0.19344640731811524, - "mean_token_accuracy": 0.9303907889127732, - "num_tokens": 6223548.0, - "step": 2600 - }, - { - "entropy": 0.2794396734237671, - "epoch": 6.386007237635706, - "grad_norm": 1.216142177581787, - "learning_rate": 4.581396973700743e-05, - "loss": 0.19337059020996095, - "mean_token_accuracy": 0.9294343847036362, - "num_tokens": 6349102.0, - "step": 2650 - }, - { - "entropy": 0.28492832884192465, - "epoch": 6.5066344993968634, - "grad_norm": 1.281471610069275, - "learning_rate": 4.3196788153980425e-05, - "loss": 0.19621110916137696, - "mean_token_accuracy": 0.9284999430179596, - "num_tokens": 6468966.0, - "step": 2700 - }, - { - "entropy": 0.28580656126141546, - "epoch": 6.627261761158022, - "grad_norm": 1.581629753112793, - "learning_rate": 4.061952474349045e-05, - "loss": 0.2017466926574707, - "mean_token_accuracy": 0.9269114434719086, - "num_tokens": 6583296.0, - "step": 2750 - }, - { - "entropy": 0.2819211499392986, - "epoch": 6.74788902291918, - "grad_norm": 1.111789584159851, - "learning_rate": 3.8086737285312406e-05, - "loss": 0.19925508499145508, - "mean_token_accuracy": 0.9279456728696823, - "num_tokens": 6698696.0, - "step": 2800 - }, - { - "entropy": 0.27608554884791375, - "epoch": 6.868516284680338, - "grad_norm": 1.337012529373169, - "learning_rate": 3.560290490540746e-05, - "loss": 0.19468421936035157, - "mean_token_accuracy": 0.9297049742937088, - "num_tokens": 6819711.0, - "step": 2850 - }, - { - "entropy": 0.27499831572175026, - "epoch": 6.989143546441496, - "grad_norm": 1.2096495628356934, - "learning_rate": 3.317242015478115e-05, - "loss": 0.19465036392211915, - "mean_token_accuracy": 0.929957520365715, - "num_tokens": 6944941.0, - "step": 2900 - }, - { - "epoch": 7.0, - "eval_entropy": 0.43061937626158253, - "eval_mean_token_accuracy": 0.7864309693320414, - "eval_not_syn_loss": 0.8960050940513611, - "eval_not_syn_runtime": 53.715, - "eval_not_syn_samples_per_second": 26.436, - "eval_not_syn_steps_per_second": 3.314, - "eval_num_tokens": 6956642.0, - "step": 2905 - }, - { - "epoch": 7.0, - "eval_entropy": 0.40260162983047826, - "eval_mean_token_accuracy": 0.81038418077351, - "eval_num_tokens": 6956642.0, - "eval_syn_loss": 0.8312228322029114, - "eval_syn_runtime": 55.7344, - "eval_syn_samples_per_second": 25.478, - "eval_syn_steps_per_second": 3.194, - "step": 2905 - }, - { - "entropy": 0.22430229894440584, - "epoch": 7.108564535585042, - "grad_norm": 1.1575374603271484, - "learning_rate": 3.0799581241445255e-05, - "loss": 0.13364919662475586, - "mean_token_accuracy": 0.9534954422652119, - "num_tokens": 7064818.0, - "step": 2950 - }, - { - "entropy": 0.21469997629523277, - "epoch": 7.2291917973462, - "grad_norm": 1.0705780982971191, - "learning_rate": 2.8488584429221433e-05, - "loss": 0.13137072563171387, - "mean_token_accuracy": 0.9546506702899933, - "num_tokens": 7183624.0, - "step": 3000 - }, - { - "entropy": 0.21114618048071862, - "epoch": 7.349819059107358, - "grad_norm": 1.1908551454544067, - "learning_rate": 2.6243516616828438e-05, - "loss": 0.1303229808807373, - "mean_token_accuracy": 0.9540972465276718, - "num_tokens": 7304122.0, - "step": 3050 - }, - { - "entropy": 0.2161379286646843, - "epoch": 7.470446320868517, - "grad_norm": 1.198324203491211, - "learning_rate": 2.4068348110377418e-05, - "loss": 0.1322408390045166, - "mean_token_accuracy": 0.9531843411922455, - "num_tokens": 7422217.0, - "step": 3100 - }, - { - "entropy": 0.20395339593291284, - "epoch": 7.591073582629674, - "grad_norm": 1.3267316818237305, - "learning_rate": 2.1966925602055525e-05, - "loss": 0.12691154479980468, - "mean_token_accuracy": 0.954550067782402, - "num_tokens": 7547071.0, - "step": 3150 - }, - { - "entropy": 0.2153193709254265, - "epoch": 7.711700844390832, - "grad_norm": 1.2034279108047485, - "learning_rate": 1.9942965367416116e-05, - "loss": 0.13238741874694823, - "mean_token_accuracy": 0.9535881149768829, - "num_tokens": 7663184.0, - "step": 3200 - }, - { - "entropy": 0.2055867248773575, - "epoch": 7.832328106151991, - "grad_norm": 1.3387084007263184, - "learning_rate": 1.8000046693304666e-05, - "loss": 0.12809998512268067, - "mean_token_accuracy": 0.9540317648649216, - "num_tokens": 7785293.0, - "step": 3250 - }, - { - "entropy": 0.21260966420173644, - "epoch": 7.952955367913148, - "grad_norm": 1.201429009437561, - "learning_rate": 1.6141605548043607e-05, - "loss": 0.13191481590270995, - "mean_token_accuracy": 0.952368374466896, - "num_tokens": 7903616.0, - "step": 3300 - }, - { - "epoch": 8.0, - "eval_entropy": 0.3767001710915833, - "eval_mean_token_accuracy": 0.7909231467193432, - "eval_not_syn_loss": 1.0326374769210815, - "eval_not_syn_runtime": 54.1394, - "eval_not_syn_samples_per_second": 26.229, - "eval_not_syn_steps_per_second": 3.288, - "eval_num_tokens": 7950448.0, - "step": 3320 - }, - { - "epoch": 8.0, - "eval_entropy": 0.35253987854786134, - "eval_mean_token_accuracy": 0.803997924488582, - "eval_num_tokens": 7950448.0, - "eval_syn_loss": 0.9743913412094116, - "eval_syn_runtime": 55.701, - "eval_syn_samples_per_second": 25.493, - "eval_syn_steps_per_second": 3.196, - "step": 3320 - } - ], - "logging_steps": 50, - "max_steps": 4150, - "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": 1.639931278004348e+17, - "train_batch_size": 8, - "trial_name": null, - "trial_params": null -} diff --git a/substitutivity_original_Estonian/Qwen3.5-2B-Base_substitutivity_splits_original_features_train_substitutivity_splits_original_features_test2/checkpoint-3320/training_args.bin b/substitutivity_original_Estonian/Qwen3.5-2B-Base_substitutivity_splits_original_features_train_substitutivity_splits_original_features_test2/checkpoint-3320/training_args.bin deleted file mode 100644 index f3d8098f65400413dd700f32ada5649470871017..0000000000000000000000000000000000000000 --- a/substitutivity_original_Estonian/Qwen3.5-2B-Base_substitutivity_splits_original_features_train_substitutivity_splits_original_features_test2/checkpoint-3320/training_args.bin +++ /dev/null @@ -1,3 +0,0 @@ -version https://git-lfs.github.com/spec/v1 -oid sha256:617af0f41414927baaaf158949585bb75eb4b05ff78baa5216c860ef0b431009 -size 6033 diff --git a/substitutivity_original_Estonian/Qwen3.5-2B-Base_substitutivity_splits_original_features_train_substitutivity_splits_original_features_test2/checkpoint-3735/README.md b/substitutivity_original_Estonian/Qwen3.5-2B-Base_substitutivity_splits_original_features_train_substitutivity_splits_original_features_test2/checkpoint-3735/README.md deleted file mode 100644 index ece204d005f7bf161d0e448efdd56fd297b7c8a2..0000000000000000000000000000000000000000 --- a/substitutivity_original_Estonian/Qwen3.5-2B-Base_substitutivity_splits_original_features_train_substitutivity_splits_original_features_test2/checkpoint-3735/README.md +++ /dev/null @@ -1,209 +0,0 @@ ---- -base_model: Qwen/Qwen3.5-2B-Base -library_name: peft -pipeline_tag: text-generation -tags: -- base_model:adapter:Qwen/Qwen3.5-2B-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/substitutivity_original_Estonian/Qwen3.5-2B-Base_substitutivity_splits_original_features_train_substitutivity_splits_original_features_test2/checkpoint-3735/adapter_config.json b/substitutivity_original_Estonian/Qwen3.5-2B-Base_substitutivity_splits_original_features_train_substitutivity_splits_original_features_test2/checkpoint-3735/adapter_config.json deleted file mode 100644 index 6a095bb3bfe249ce8e15df6ae999e88fa1e17162..0000000000000000000000000000000000000000 --- a/substitutivity_original_Estonian/Qwen3.5-2B-Base_substitutivity_splits_original_features_train_substitutivity_splits_original_features_test2/checkpoint-3735/adapter_config.json +++ /dev/null @@ -1,46 +0,0 @@ -{ - "alora_invocation_tokens": null, - "alpha_pattern": {}, - "arrow_config": null, - "auto_mapping": null, - "base_model_name_or_path": "Qwen/Qwen3.5-2B-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": 0.05731002492184622, - "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": [ - "o_proj", - "q_proj", - "k_proj", - "down_proj", - "up_proj", - "v_proj", - "gate_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/substitutivity_original_Estonian/Qwen3.5-2B-Base_substitutivity_splits_original_features_train_substitutivity_splits_original_features_test2/checkpoint-3735/adapter_model.safetensors b/substitutivity_original_Estonian/Qwen3.5-2B-Base_substitutivity_splits_original_features_train_substitutivity_splits_original_features_test2/checkpoint-3735/adapter_model.safetensors deleted file mode 100644 index 73788597ce48d57e3bcd832f4ce20438eda9f4b1..0000000000000000000000000000000000000000 --- a/substitutivity_original_Estonian/Qwen3.5-2B-Base_substitutivity_splits_original_features_train_substitutivity_splits_original_features_test2/checkpoint-3735/adapter_model.safetensors +++ /dev/null @@ -1,3 +0,0 @@ -version https://git-lfs.github.com/spec/v1 -oid sha256:c5c55e096c9cdd4a9405ded559d4a3129bcec7320f65e0e5214101329c8b2c6d -size 174613392 diff --git a/substitutivity_original_Estonian/Qwen3.5-2B-Base_substitutivity_splits_original_features_train_substitutivity_splits_original_features_test2/checkpoint-3735/chat_template.jinja b/substitutivity_original_Estonian/Qwen3.5-2B-Base_substitutivity_splits_original_features_train_substitutivity_splits_original_features_test2/checkpoint-3735/chat_template.jinja deleted file mode 100644 index 0ef09f214eaa6d9bca297988afc1454b5827b2c7..0000000000000000000000000000000000000000 --- a/substitutivity_original_Estonian/Qwen3.5-2B-Base_substitutivity_splits_original_features_train_substitutivity_splits_original_features_test2/checkpoint-3735/chat_template.jinja +++ /dev/null @@ -1,154 +0,0 @@ -{%- set image_count = namespace(value=0) %} -{%- set video_count = namespace(value=0) %} -{%- macro render_content(content, do_vision_count, is_system_content=false) %} - {%- if content is string %} - {{- content }} - {%- elif content is iterable and content is not mapping %} - {%- for item in content %} - {%- if 'image' in item or 'image_url' in item or item.type == 'image' %} - {%- if is_system_content %} - {{- raise_exception('System message cannot contain images.') }} - {%- endif %} - {%- if do_vision_count %} - {%- set image_count.value = image_count.value + 1 %} - {%- endif %} - {%- if add_vision_id %} - {{- 'Picture ' ~ image_count.value ~ ': ' }} - {%- endif %} - {{- '<|vision_start|><|image_pad|><|vision_end|>' }} - {%- elif 'video' in item or item.type == 'video' %} - {%- if is_system_content %} - {{- raise_exception('System message cannot contain videos.') }} - {%- endif %} - {%- if do_vision_count %} - {%- set video_count.value = video_count.value + 1 %} - {%- endif %} - {%- if add_vision_id %} - {{- 'Video ' ~ video_count.value ~ ': ' }} - {%- endif %} - {{- '<|vision_start|><|video_pad|><|vision_end|>' }} - {%- elif 'text' in item %} - {{- item.text }} - {%- else %} - {{- raise_exception('Unexpected item type in content.') }} - {%- endif %} - {%- endfor %} - {%- elif content is none or content is undefined %} - {{- '' }} - {%- else %} - {{- raise_exception('Unexpected content type.') }} - {%- endif %} -{%- endmacro %} -{%- if not messages %} - {{- raise_exception('No messages provided.') }} -{%- endif %} -{%- if tools and tools is iterable and tools is not mapping %} - {{- '<|im_start|>system\n' }} - {{- "# Tools\n\nYou have access to the following functions:\n\n" }} - {%- for tool in tools %} - {{- "\n" }} - {{- tool | tojson }} - {%- endfor %} - {{- "\n" }} - {{- '\n\nIf you choose to call a function ONLY reply in the following format with NO suffix:\n\n\n\n\nvalue_1\n\n\nThis is the value for the second parameter\nthat can span\nmultiple lines\n\n\n\n\n\nReminder:\n- Function calls MUST follow the specified format: an inner block must be nested within XML tags\n- Required parameters MUST be specified\n- You may provide optional reasoning for your function call in natural language BEFORE the function call, but NOT after\n- If there is no function call available, answer the question like normal with your current knowledge and do not tell the user about function calls\n' }} - {%- if messages[0].role == 'system' %} - {%- set content = render_content(messages[0].content, false, true)|trim %} - {%- if content %} - {{- '\n\n' + content }} - {%- endif %} - {%- endif %} - {{- '<|im_end|>\n' }} -{%- else %} - {%- if messages[0].role == 'system' %} - {%- set content = render_content(messages[0].content, false, true)|trim %} - {{- '<|im_start|>system\n' + 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" %} - {%- set content = render_content(message.content, false)|trim %} - {%- if not(content.startswith('') and content.endswith('')) %} - {%- set ns.multi_step_tool = false %} - {%- set ns.last_query_index = index %} - {%- endif %} - {%- endif %} -{%- endfor %} -{%- if ns.multi_step_tool %} - {{- raise_exception('No user query found in messages.') }} -{%- endif %} -{%- for message in messages %} - {%- set content = render_content(message.content, true)|trim %} - {%- if message.role == "system" %} - {%- if not loop.first %} - {{- raise_exception('System message must be at the beginning.') }} - {%- endif %} - {%- elif message.role == "user" %} - {{- '<|im_start|>' + message.role + '\n' + content + '<|im_end|>' + '\n' }} - {%- elif message.role == "assistant" %} - {%- set reasoning_content = '' %} - {%- if message.reasoning_content is string %} - {%- set reasoning_content = message.reasoning_content %} - {%- else %} - {%- if '' in content %} - {%- set reasoning_content = content.split('')[0].rstrip('\n').split('')[-1].lstrip('\n') %} - {%- set content = content.split('')[-1].lstrip('\n') %} - {%- endif %} - {%- endif %} - {%- set reasoning_content = reasoning_content|trim %} - {%- if loop.index0 > ns.last_query_index %} - {{- '<|im_start|>' + message.role + '\n\n' + reasoning_content + '\n\n\n' + content }} - {%- else %} - {{- '<|im_start|>' + message.role + '\n' + content }} - {%- endif %} - {%- if message.tool_calls and message.tool_calls is iterable and message.tool_calls is not mapping %} - {%- for tool_call in message.tool_calls %} - {%- if tool_call.function is defined %} - {%- set tool_call = tool_call.function %} - {%- endif %} - {%- if loop.first %} - {%- if content|trim %} - {{- '\n\n\n\n' }} - {%- else %} - {{- '\n\n' }} - {%- endif %} - {%- else %} - {{- '\n\n\n' }} - {%- endif %} - {%- if tool_call.arguments is defined %} - {%- for args_name, args_value in tool_call.arguments|items %} - {{- '\n' }} - {%- set args_value = args_value | tojson | safe if args_value is mapping or (args_value is sequence and args_value is not string) else args_value | string %} - {{- args_value }} - {{- '\n\n' }} - {%- endfor %} - {%- endif %} - {{- '\n' }} - {%- endfor %} - {%- endif %} - {{- '<|im_end|>\n' }} - {%- elif message.role == "tool" %} - {%- if loop.previtem and loop.previtem.role != "tool" %} - {{- '<|im_start|>user' }} - {%- endif %} - {{- '\n\n' }} - {{- content }} - {{- '\n' }} - {%- if not loop.last and loop.nextitem.role != "tool" %} - {{- '<|im_end|>\n' }} - {%- elif loop.last %} - {{- '<|im_end|>\n' }} - {%- endif %} - {%- else %} - {{- raise_exception('Unexpected message role.') }} - {%- endif %} -{%- endfor %} -{%- if add_generation_prompt %} - {{- '<|im_start|>assistant\n' }} - {%- if enable_thinking is defined and enable_thinking is true %} - {{- '\n' }} - {%- else %} - {{- '\n\n\n\n' }} - {%- endif %} -{%- endif %} \ No newline at end of file diff --git a/substitutivity_original_Estonian/Qwen3.5-2B-Base_substitutivity_splits_original_features_train_substitutivity_splits_original_features_test2/checkpoint-3735/tokenizer.json b/substitutivity_original_Estonian/Qwen3.5-2B-Base_substitutivity_splits_original_features_train_substitutivity_splits_original_features_test2/checkpoint-3735/tokenizer.json deleted file mode 100644 index 67741b04f23bfdb46501f748ce27865ec82eccfb..0000000000000000000000000000000000000000 --- a/substitutivity_original_Estonian/Qwen3.5-2B-Base_substitutivity_splits_original_features_train_substitutivity_splits_original_features_test2/checkpoint-3735/tokenizer.json +++ /dev/null @@ -1,3 +0,0 @@ -version https://git-lfs.github.com/spec/v1 -oid sha256:87a7830d63fcf43bf241c3c5242e96e62dd3fdc29224ca26fed8ea333db72de4 -size 19989343 diff --git a/substitutivity_original_Estonian/Qwen3.5-2B-Base_substitutivity_splits_original_features_train_substitutivity_splits_original_features_test2/checkpoint-3735/tokenizer_config.json b/substitutivity_original_Estonian/Qwen3.5-2B-Base_substitutivity_splits_original_features_train_substitutivity_splits_original_features_test2/checkpoint-3735/tokenizer_config.json deleted file mode 100644 index b4a37b2a6fd3ab3317cd7bac72855be1a843b2bb..0000000000000000000000000000000000000000 --- a/substitutivity_original_Estonian/Qwen3.5-2B-Base_substitutivity_splits_original_features_train_substitutivity_splits_original_features_test2/checkpoint-3735/tokenizer_config.json +++ /dev/null @@ -1,31 +0,0 @@ -{ - "add_prefix_space": false, - "audio_bos_token": "<|audio_start|>", - "audio_eos_token": "<|audio_end|>", - "audio_token": "<|audio_pad|>", - "backend": "tokenizers", - "bos_token": null, - "clean_up_tokenization_spaces": false, - "eos_token": "<|endoftext|>", - "errors": "replace", - "image_token": "<|image_pad|>", - "is_local": false, - "model_max_length": 262144, - "model_specific_special_tokens": { - "audio_bos_token": "<|audio_start|>", - "audio_eos_token": "<|audio_end|>", - "audio_token": "<|audio_pad|>", - "image_token": "<|image_pad|>", - "video_token": "<|video_pad|>", - "vision_bos_token": "<|vision_start|>", - "vision_eos_token": "<|vision_end|>" - }, - "pad_token": "<|endoftext|>", - "pretokenize_regex": "(?i:'s|'t|'re|'ve|'m|'ll|'d)|[^\\r\\n\\p{L}\\p{N}]?[\\p{L}\\p{M}]+|\\p{N}| ?[^\\s\\p{L}\\p{M}\\p{N}]+[\\r\\n]*|\\s*[\\r\\n]+|\\s+(?!\\S)|\\s+", - "split_special_tokens": false, - "tokenizer_class": "TokenizersBackend", - "unk_token": null, - "video_token": "<|video_pad|>", - "vision_bos_token": "<|vision_start|>", - "vision_eos_token": "<|vision_end|>" -} diff --git a/substitutivity_original_Estonian/Qwen3.5-2B-Base_substitutivity_splits_original_features_train_substitutivity_splits_original_features_test2/checkpoint-3735/trainer_state.json b/substitutivity_original_Estonian/Qwen3.5-2B-Base_substitutivity_splits_original_features_train_substitutivity_splits_original_features_test2/checkpoint-3735/trainer_state.json deleted file mode 100644 index e9896cd5f70fa3aff67a2b98f5482e82105cf02d..0000000000000000000000000000000000000000 --- a/substitutivity_original_Estonian/Qwen3.5-2B-Base_substitutivity_splits_original_features_train_substitutivity_splits_original_features_test2/checkpoint-3735/trainer_state.json +++ /dev/null @@ -1,972 +0,0 @@ -{ - "best_global_step": null, - "best_metric": null, - "best_model_checkpoint": null, - "epoch": 9.0, - "eval_steps": 500, - "global_step": 3735, - "is_hyper_param_search": false, - "is_local_process_zero": true, - "is_world_process_zero": true, - "log_history": [ - { - "entropy": 2.2933380508422854, - "epoch": 0.12062726176115803, - "grad_norm": 2.8075337409973145, - "learning_rate": 1.5531017374629602e-05, - "loss": 2.1429234313964844, - "mean_token_accuracy": 0.5779684272408485, - "num_tokens": 121428.0, - "step": 50 - }, - { - "entropy": 1.1315062403678895, - "epoch": 0.24125452352231605, - "grad_norm": 1.9415371417999268, - "learning_rate": 3.1378994287516954e-05, - "loss": 1.0461698913574218, - "mean_token_accuracy": 0.7166991305351257, - "num_tokens": 243966.0, - "step": 100 - }, - { - "entropy": 0.9537658470869065, - "epoch": 0.3618817852834741, - "grad_norm": 1.8353157043457031, - "learning_rate": 4.72269712004043e-05, - "loss": 0.8774301147460938, - "mean_token_accuracy": 0.7506789201498032, - "num_tokens": 366649.0, - "step": 150 - }, - { - "entropy": 0.8955666983127594, - "epoch": 0.4825090470446321, - "grad_norm": 1.5419880151748657, - "learning_rate": 6.307494811329165e-05, - "loss": 0.8234302520751953, - "mean_token_accuracy": 0.7616239994764328, - "num_tokens": 488367.0, - "step": 200 - }, - { - "entropy": 0.8689469587802887, - "epoch": 0.6031363088057901, - "grad_norm": 1.5661484003067017, - "learning_rate": 7.8922925026179e-05, - "loss": 0.7961890411376953, - "mean_token_accuracy": 0.767715499997139, - "num_tokens": 603131.0, - "step": 250 - }, - { - "entropy": 0.8425652319192887, - "epoch": 0.7237635705669482, - "grad_norm": 1.5694905519485474, - "learning_rate": 9.477090193906636e-05, - "loss": 0.7718594360351563, - "mean_token_accuracy": 0.7736722755432129, - "num_tokens": 726412.0, - "step": 300 - }, - { - "entropy": 0.843294187784195, - "epoch": 0.8443908323281062, - "grad_norm": 1.1118028163909912, - "learning_rate": 0.00011061887885195371, - "loss": 0.7801297760009765, - "mean_token_accuracy": 0.7714760398864746, - "num_tokens": 846516.0, - "step": 350 - }, - { - "entropy": 0.8213214933872223, - "epoch": 0.9650180940892642, - "grad_norm": 0.9202918410301208, - "learning_rate": 0.00012646685576484106, - "loss": 0.7514453887939453, - "mean_token_accuracy": 0.7774044269323349, - "num_tokens": 958978.0, - "step": 400 - }, - { - "epoch": 1.0, - "eval_entropy": 0.818668900915746, - "eval_mean_token_accuracy": 0.7533496866065464, - "eval_not_syn_loss": 0.7798981070518494, - "eval_not_syn_runtime": 53.8744, - "eval_not_syn_samples_per_second": 26.358, - "eval_not_syn_steps_per_second": 3.304, - "eval_num_tokens": 993806.0, - "step": 415 - }, - { - "epoch": 1.0, - "eval_entropy": 0.7627468125873738, - "eval_mean_token_accuracy": 0.8024111574285486, - "eval_num_tokens": 993806.0, - "eval_syn_loss": 0.7218216061592102, - "eval_syn_runtime": 55.3239, - "eval_syn_samples_per_second": 25.667, - "eval_syn_steps_per_second": 3.217, - "step": 415 - }, - { - "entropy": 0.7802230679627621, - "epoch": 1.0844390832328106, - "grad_norm": 1.0145268440246582, - "learning_rate": 0.00013151131541121804, - "loss": 0.7111608123779297, - "mean_token_accuracy": 0.7845737542768921, - "num_tokens": 1080301.0, - "step": 450 - }, - { - "entropy": 0.7558828103542328, - "epoch": 1.2050663449939687, - "grad_norm": 1.0477591753005981, - "learning_rate": 0.0001313741160170038, - "loss": 0.6933037567138672, - "mean_token_accuracy": 0.7889221370220184, - "num_tokens": 1201215.0, - "step": 500 - }, - { - "entropy": 0.7593123841285706, - "epoch": 1.3256936067551266, - "grad_norm": 0.899161159992218, - "learning_rate": 0.00013112089697387701, - "loss": 0.6962493133544921, - "mean_token_accuracy": 0.7877359414100646, - "num_tokens": 1322895.0, - "step": 550 - }, - { - "entropy": 0.7460986822843552, - "epoch": 1.4463208685162847, - "grad_norm": 1.1061116456985474, - "learning_rate": 0.00013075210608885215, - "loss": 0.6817412567138672, - "mean_token_accuracy": 0.7904874366521836, - "num_tokens": 1436749.0, - "step": 600 - }, - { - "entropy": 0.735516265630722, - "epoch": 1.5669481302774426, - "grad_norm": 0.9925405979156494, - "learning_rate": 0.00013026839555278796, - "loss": 0.6690705871582031, - "mean_token_accuracy": 0.7930250608921051, - "num_tokens": 1555400.0, - "step": 650 - }, - { - "entropy": 0.7302519899606704, - "epoch": 1.6875753920386007, - "grad_norm": 0.800025463104248, - "learning_rate": 0.0001296706207870159, - "loss": 0.6714833831787109, - "mean_token_accuracy": 0.7946307575702667, - "num_tokens": 1676892.0, - "step": 700 - }, - { - "entropy": 0.7321263110637665, - "epoch": 1.8082026537997589, - "grad_norm": 1.0500682592391968, - "learning_rate": 0.00012895983893056423, - "loss": 0.6672868347167968, - "mean_token_accuracy": 0.7947645837068558, - "num_tokens": 1802689.0, - "step": 750 - }, - { - "entropy": 0.7075755137205124, - "epoch": 1.9288299155609168, - "grad_norm": 0.9248411655426025, - "learning_rate": 0.00012813730697065298, - "loss": 0.6466568756103516, - "mean_token_accuracy": 0.798747689127922, - "num_tokens": 1920804.0, - "step": 800 - }, - { - "epoch": 2.0, - "eval_entropy": 0.7426781811740961, - "eval_mean_token_accuracy": 0.7693507239389955, - "eval_not_syn_loss": 0.7139139175415039, - "eval_not_syn_runtime": 54.4286, - "eval_not_syn_samples_per_second": 26.089, - "eval_not_syn_steps_per_second": 3.27, - "eval_num_tokens": 1987612.0, - "step": 830 - }, - { - "epoch": 2.0, - "eval_entropy": 0.6937930352902144, - "eval_mean_token_accuracy": 0.8180048633157537, - "eval_num_tokens": 1987612.0, - "eval_syn_loss": 0.6489660143852234, - "eval_syn_runtime": 56.1525, - "eval_syn_samples_per_second": 25.288, - "eval_syn_steps_per_second": 3.17, - "step": 830 - }, - { - "entropy": 0.6853217440422135, - "epoch": 2.0482509047044632, - "grad_norm": 0.6932648420333862, - "learning_rate": 0.00012720447951976603, - "loss": 0.6212613296508789, - "mean_token_accuracy": 0.8047762579388089, - "num_tokens": 2039660.0, - "step": 850 - }, - { - "entropy": 0.6459032052755356, - "epoch": 2.168878166465621, - "grad_norm": 1.0688549280166626, - "learning_rate": 0.00012616300624323124, - "loss": 0.5740525817871094, - "mean_token_accuracy": 0.8148404061794281, - "num_tokens": 2161772.0, - "step": 900 - }, - { - "entropy": 0.6463843190670013, - "epoch": 2.2895054282267795, - "grad_norm": 0.847271203994751, - "learning_rate": 0.0001250147289418581, - "loss": 0.577717170715332, - "mean_token_accuracy": 0.814143658876419, - "num_tokens": 2280833.0, - "step": 950 - }, - { - "entropy": 0.6461306965351105, - "epoch": 2.4101326899879374, - "grad_norm": 0.9998957514762878, - "learning_rate": 0.00012376167829479217, - "loss": 0.5813011932373047, - "mean_token_accuracy": 0.8144401633739471, - "num_tokens": 2398206.0, - "step": 1000 - }, - { - "entropy": 0.640914659500122, - "epoch": 2.5307599517490953, - "grad_norm": 0.8692949414253235, - "learning_rate": 0.0001224060702683459, - "loss": 0.5755012512207032, - "mean_token_accuracy": 0.8168018108606339, - "num_tokens": 2518888.0, - "step": 1050 - }, - { - "entropy": 0.6421274638175964, - "epoch": 2.651387213510253, - "grad_norm": 0.8920196294784546, - "learning_rate": 0.00012095030219715738, - "loss": 0.5785483932495117, - "mean_token_accuracy": 0.8137972563505173, - "num_tokens": 2639763.0, - "step": 1100 - }, - { - "entropy": 0.6466812592744827, - "epoch": 2.772014475271411, - "grad_norm": 0.6876946687698364, - "learning_rate": 0.00011939694854460696, - "loss": 0.5816070938110351, - "mean_token_accuracy": 0.8145407867431641, - "num_tokens": 2755540.0, - "step": 1150 - }, - { - "entropy": 0.6378493067622185, - "epoch": 2.8926417370325694, - "grad_norm": 0.7532753944396973, - "learning_rate": 0.00011774875634998911, - "loss": 0.5754346084594727, - "mean_token_accuracy": 0.8148861181735992, - "num_tokens": 2874118.0, - "step": 1200 - }, - { - "epoch": 3.0, - "eval_entropy": 0.6912222209941136, - "eval_mean_token_accuracy": 0.7868467750174276, - "eval_not_syn_loss": 0.6782422661781311, - "eval_not_syn_runtime": 53.2627, - "eval_not_syn_samples_per_second": 26.66, - "eval_not_syn_steps_per_second": 3.342, - "eval_num_tokens": 2981418.0, - "step": 1245 - }, - { - "epoch": 3.0, - "eval_entropy": 0.6467212934172555, - "eval_mean_token_accuracy": 0.8215091817834405, - "eval_num_tokens": 2981418.0, - "eval_syn_loss": 0.6301999092102051, - "eval_syn_runtime": 55.0115, - "eval_syn_samples_per_second": 25.813, - "eval_syn_steps_per_second": 3.236, - "step": 1245 - }, - { - "entropy": 0.6316030452949832, - "epoch": 3.012062726176116, - "grad_norm": 0.7761592864990234, - "learning_rate": 0.00011600864037049124, - "loss": 0.5697880935668945, - "mean_token_accuracy": 0.8159356400220081, - "num_tokens": 2994602.0, - "step": 1250 - }, - { - "entropy": 0.5467158386111259, - "epoch": 3.132689987937274, - "grad_norm": 0.8169087767601013, - "learning_rate": 0.00011417967792657078, - "loss": 0.47844127655029295, - "mean_token_accuracy": 0.8378052765130997, - "num_tokens": 3121094.0, - "step": 1300 - }, - { - "entropy": 0.5454566389322281, - "epoch": 3.2533172496984317, - "grad_norm": 0.8219636678695679, - "learning_rate": 0.0001122651034598459, - "loss": 0.4839338302612305, - "mean_token_accuracy": 0.8385071897506714, - "num_tokens": 3248038.0, - "step": 1350 - }, - { - "entropy": 0.5598723018169403, - "epoch": 3.37394451145959, - "grad_norm": 0.9956360459327698, - "learning_rate": 0.00011026830281312423, - "loss": 0.492743034362793, - "mean_token_accuracy": 0.8357370454072952, - "num_tokens": 3363487.0, - "step": 1400 - }, - { - "entropy": 0.5479365903139114, - "epoch": 3.494571773220748, - "grad_norm": 0.8586823344230652, - "learning_rate": 0.00010819280724268524, - "loss": 0.4810820770263672, - "mean_token_accuracy": 0.837928666472435, - "num_tokens": 3480513.0, - "step": 1450 - }, - { - "entropy": 0.5619524419307709, - "epoch": 3.615199034981906, - "grad_norm": 1.0894954204559326, - "learning_rate": 0.0001060422871734048, - "loss": 0.4974747085571289, - "mean_token_accuracy": 0.8349784809350967, - "num_tokens": 3597183.0, - "step": 1500 - }, - { - "entropy": 0.5588027596473694, - "epoch": 3.7358262967430638, - "grad_norm": 0.9905257225036621, - "learning_rate": 0.00010382054570776654, - "loss": 0.49687202453613283, - "mean_token_accuracy": 0.834458841085434, - "num_tokens": 3714173.0, - "step": 1550 - }, - { - "entropy": 0.5654991599917412, - "epoch": 3.856453558504222, - "grad_norm": 1.0116610527038574, - "learning_rate": 0.00010153151190023791, - "loss": 0.5016670989990234, - "mean_token_accuracy": 0.8326652735471726, - "num_tokens": 3832652.0, - "step": 1600 - }, - { - "entropy": 0.5605382058024406, - "epoch": 3.97708082026538, - "grad_norm": 0.8011100888252258, - "learning_rate": 9.917923380890624e-05, - "loss": 0.49551734924316404, - "mean_token_accuracy": 0.8343535929918289, - "num_tokens": 3955167.0, - "step": 1650 - }, - { - "epoch": 4.0, - "eval_entropy": 0.5819476859623127, - "eval_mean_token_accuracy": 0.7963069781158747, - "eval_not_syn_loss": 0.6924306750297546, - "eval_not_syn_runtime": 54.027, - "eval_not_syn_samples_per_second": 26.283, - "eval_not_syn_steps_per_second": 3.295, - "eval_num_tokens": 3975224.0, - "step": 1660 - }, - { - "epoch": 4.0, - "eval_entropy": 0.545470636212424, - "eval_mean_token_accuracy": 0.8127643666910321, - "eval_num_tokens": 3975224.0, - "eval_syn_loss": 0.649576723575592, - "eval_syn_runtime": 55.2296, - "eval_syn_samples_per_second": 25.711, - "eval_syn_steps_per_second": 3.223, - "step": 1660 - }, - { - "entropy": 0.4920780294471317, - "epoch": 4.0965018094089265, - "grad_norm": 0.8362822532653809, - "learning_rate": 9.676787133666148e-05, - "loss": 0.4147996139526367, - "mean_token_accuracy": 0.8581844628459275, - "num_tokens": 4072361.0, - "step": 1700 - }, - { - "entropy": 0.4675139603018761, - "epoch": 4.217129071170085, - "grad_norm": 0.9920316338539124, - "learning_rate": 9.430168887458645e-05, - "loss": 0.3863528823852539, - "mean_token_accuracy": 0.865827117562294, - "num_tokens": 4194978.0, - "step": 1750 - }, - { - "entropy": 0.47163557589054106, - "epoch": 4.337756332931242, - "grad_norm": 1.0789282321929932, - "learning_rate": 9.178504776056418e-05, - "loss": 0.390842170715332, - "mean_token_accuracy": 0.8634983265399933, - "num_tokens": 4315548.0, - "step": 1800 - }, - { - "entropy": 0.47117673605680466, - "epoch": 4.458383594692401, - "grad_norm": 1.0369561910629272, - "learning_rate": 8.922239856643896e-05, - "loss": 0.39387256622314454, - "mean_token_accuracy": 0.8625931572914124, - "num_tokens": 4431469.0, - "step": 1850 - }, - { - "entropy": 0.479859294295311, - "epoch": 4.579010856453559, - "grad_norm": 1.0827305316925049, - "learning_rate": 8.661827322737095e-05, - "loss": 0.40195709228515625, - "mean_token_accuracy": 0.8612629306316376, - "num_tokens": 4551803.0, - "step": 1900 - }, - { - "entropy": 0.4729380664229393, - "epoch": 4.699638118214716, - "grad_norm": 1.0114246606826782, - "learning_rate": 8.397727702730326e-05, - "loss": 0.39648338317871096, - "mean_token_accuracy": 0.8610477095842362, - "num_tokens": 4674468.0, - "step": 1950 - }, - { - "entropy": 0.4903118047118187, - "epoch": 4.820265379975875, - "grad_norm": 1.0115467309951782, - "learning_rate": 8.130408045471496e-05, - "loss": 0.4067733001708984, - "mean_token_accuracy": 0.8597196912765503, - "num_tokens": 4788027.0, - "step": 2000 - }, - { - "entropy": 0.48099748432636263, - "epoch": 4.940892641737032, - "grad_norm": 1.0201005935668945, - "learning_rate": 7.860341094306242e-05, - "loss": 0.4030665588378906, - "mean_token_accuracy": 0.8603792876005173, - "num_tokens": 4907561.0, - "step": 2050 - }, - { - "epoch": 5.0, - "eval_entropy": 0.5492983292997553, - "eval_mean_token_accuracy": 0.796870938847574, - "eval_not_syn_loss": 0.7129728198051453, - "eval_not_syn_runtime": 53.9575, - "eval_not_syn_samples_per_second": 26.317, - "eval_not_syn_steps_per_second": 3.299, - "eval_num_tokens": 4969030.0, - "step": 2075 - }, - { - "epoch": 5.0, - "eval_entropy": 0.513823946037989, - "eval_mean_token_accuracy": 0.8174765464964877, - "eval_num_tokens": 4969030.0, - "eval_syn_loss": 0.6699439883232117, - "eval_syn_runtime": 56.0033, - "eval_syn_samples_per_second": 25.356, - "eval_syn_steps_per_second": 3.178, - "step": 2075 - }, - { - "entropy": 0.43584728511897003, - "epoch": 5.060313630880579, - "grad_norm": 0.9914600253105164, - "learning_rate": 7.588004451051618e-05, - "loss": 0.34999092102050783, - "mean_token_accuracy": 0.8777085295831314, - "num_tokens": 5026625.0, - "step": 2100 - }, - { - "entropy": 0.38114599585533143, - "epoch": 5.180940892641737, - "grad_norm": 1.232283115386963, - "learning_rate": 7.313879731377764e-05, - "loss": 0.2903956604003906, - "mean_token_accuracy": 0.896172685623169, - "num_tokens": 5141010.0, - "step": 2150 - }, - { - "entropy": 0.3792745009064674, - "epoch": 5.301568154402895, - "grad_norm": 1.2843542098999023, - "learning_rate": 7.03845171309126e-05, - "loss": 0.2903476333618164, - "mean_token_accuracy": 0.8951605755090714, - "num_tokens": 5261240.0, - "step": 2200 - }, - { - "entropy": 0.3744692197442055, - "epoch": 5.422195416164053, - "grad_norm": 1.23075270652771, - "learning_rate": 6.76220747882637e-05, - "loss": 0.29011749267578124, - "mean_token_accuracy": 0.8963996982574463, - "num_tokens": 5382160.0, - "step": 2250 - }, - { - "entropy": 0.3795280712842941, - "epoch": 5.542822677925211, - "grad_norm": 1.4011399745941162, - "learning_rate": 6.485635554660295e-05, - "loss": 0.2921779823303223, - "mean_token_accuracy": 0.8954424083232879, - "num_tokens": 5502002.0, - "step": 2300 - }, - { - "entropy": 0.3797056859731674, - "epoch": 5.66344993968637, - "grad_norm": 1.1022812128067017, - "learning_rate": 6.209225046175775e-05, - "loss": 0.29343433380126954, - "mean_token_accuracy": 0.8939031893014908, - "num_tokens": 5629969.0, - "step": 2350 - }, - { - "entropy": 0.3827402114868164, - "epoch": 5.784077201447527, - "grad_norm": 1.22834312915802, - "learning_rate": 5.9334647734988365e-05, - "loss": 0.2954188537597656, - "mean_token_accuracy": 0.8937842601537704, - "num_tokens": 5750596.0, - "step": 2400 - }, - { - "entropy": 0.38422658771276474, - "epoch": 5.904704463208685, - "grad_norm": 1.425066351890564, - "learning_rate": 5.6588424068413775e-05, - "loss": 0.296480770111084, - "mean_token_accuracy": 0.8936110073328019, - "num_tokens": 5868110.0, - "step": 2450 - }, - { - "epoch": 6.0, - "eval_entropy": 0.49826224684045556, - "eval_mean_token_accuracy": 0.795356297760867, - "eval_not_syn_loss": 0.7700047492980957, - "eval_not_syn_runtime": 53.0632, - "eval_not_syn_samples_per_second": 26.761, - "eval_not_syn_steps_per_second": 3.354, - "eval_num_tokens": 5962836.0, - "step": 2490 - }, - { - "epoch": 6.0, - "eval_entropy": 0.4654735526007213, - "eval_mean_token_accuracy": 0.8118122223387944, - "eval_num_tokens": 5962836.0, - "eval_syn_loss": 0.7288840413093567, - "eval_syn_runtime": 54.6517, - "eval_syn_samples_per_second": 25.983, - "eval_syn_steps_per_second": 3.257, - "step": 2490 - }, - { - "entropy": 0.3596555646320786, - "epoch": 6.024125452352232, - "grad_norm": 1.628747820854187, - "learning_rate": 5.385843604077321e-05, - "loss": 0.2706174850463867, - "mean_token_accuracy": 0.9024625179743526, - "num_tokens": 5990107.0, - "step": 2500 - }, - { - "entropy": 0.285839084982872, - "epoch": 6.144752714113389, - "grad_norm": 1.2864570617675781, - "learning_rate": 5.114951151877489e-05, - "loss": 0.19398983001708983, - "mean_token_accuracy": 0.9306288361549377, - "num_tokens": 6104230.0, - "step": 2550 - }, - { - "entropy": 0.2841903820633888, - "epoch": 6.265379975874548, - "grad_norm": 1.3262356519699097, - "learning_rate": 4.846644111922105e-05, - "loss": 0.19344640731811524, - "mean_token_accuracy": 0.9303907889127732, - "num_tokens": 6223548.0, - "step": 2600 - }, - { - "entropy": 0.2794396734237671, - "epoch": 6.386007237635706, - "grad_norm": 1.216142177581787, - "learning_rate": 4.581396973700743e-05, - "loss": 0.19337059020996095, - "mean_token_accuracy": 0.9294343847036362, - "num_tokens": 6349102.0, - "step": 2650 - }, - { - "entropy": 0.28492832884192465, - "epoch": 6.5066344993968634, - "grad_norm": 1.281471610069275, - "learning_rate": 4.3196788153980425e-05, - "loss": 0.19621110916137696, - "mean_token_accuracy": 0.9284999430179596, - "num_tokens": 6468966.0, - "step": 2700 - }, - { - "entropy": 0.28580656126141546, - "epoch": 6.627261761158022, - "grad_norm": 1.581629753112793, - "learning_rate": 4.061952474349045e-05, - "loss": 0.2017466926574707, - "mean_token_accuracy": 0.9269114434719086, - "num_tokens": 6583296.0, - "step": 2750 - }, - { - "entropy": 0.2819211499392986, - "epoch": 6.74788902291918, - "grad_norm": 1.111789584159851, - "learning_rate": 3.8086737285312406e-05, - "loss": 0.19925508499145508, - "mean_token_accuracy": 0.9279456728696823, - "num_tokens": 6698696.0, - "step": 2800 - }, - { - "entropy": 0.27608554884791375, - "epoch": 6.868516284680338, - "grad_norm": 1.337012529373169, - "learning_rate": 3.560290490540746e-05, - "loss": 0.19468421936035157, - "mean_token_accuracy": 0.9297049742937088, - "num_tokens": 6819711.0, - "step": 2850 - }, - { - "entropy": 0.27499831572175026, - "epoch": 6.989143546441496, - "grad_norm": 1.2096495628356934, - "learning_rate": 3.317242015478115e-05, - "loss": 0.19465036392211915, - "mean_token_accuracy": 0.929957520365715, - "num_tokens": 6944941.0, - "step": 2900 - }, - { - "epoch": 7.0, - "eval_entropy": 0.43061937626158253, - "eval_mean_token_accuracy": 0.7864309693320414, - "eval_not_syn_loss": 0.8960050940513611, - "eval_not_syn_runtime": 53.715, - "eval_not_syn_samples_per_second": 26.436, - "eval_not_syn_steps_per_second": 3.314, - "eval_num_tokens": 6956642.0, - "step": 2905 - }, - { - "epoch": 7.0, - "eval_entropy": 0.40260162983047826, - "eval_mean_token_accuracy": 0.81038418077351, - "eval_num_tokens": 6956642.0, - "eval_syn_loss": 0.8312228322029114, - "eval_syn_runtime": 55.7344, - "eval_syn_samples_per_second": 25.478, - "eval_syn_steps_per_second": 3.194, - "step": 2905 - }, - { - "entropy": 0.22430229894440584, - "epoch": 7.108564535585042, - "grad_norm": 1.1575374603271484, - "learning_rate": 3.0799581241445255e-05, - "loss": 0.13364919662475586, - "mean_token_accuracy": 0.9534954422652119, - "num_tokens": 7064818.0, - "step": 2950 - }, - { - "entropy": 0.21469997629523277, - "epoch": 7.2291917973462, - "grad_norm": 1.0705780982971191, - "learning_rate": 2.8488584429221433e-05, - "loss": 0.13137072563171387, - "mean_token_accuracy": 0.9546506702899933, - "num_tokens": 7183624.0, - "step": 3000 - }, - { - "entropy": 0.21114618048071862, - "epoch": 7.349819059107358, - "grad_norm": 1.1908551454544067, - "learning_rate": 2.6243516616828438e-05, - "loss": 0.1303229808807373, - "mean_token_accuracy": 0.9540972465276718, - "num_tokens": 7304122.0, - "step": 3050 - }, - { - "entropy": 0.2161379286646843, - "epoch": 7.470446320868517, - "grad_norm": 1.198324203491211, - "learning_rate": 2.4068348110377418e-05, - "loss": 0.1322408390045166, - "mean_token_accuracy": 0.9531843411922455, - "num_tokens": 7422217.0, - "step": 3100 - }, - { - "entropy": 0.20395339593291284, - "epoch": 7.591073582629674, - "grad_norm": 1.3267316818237305, - "learning_rate": 2.1966925602055525e-05, - "loss": 0.12691154479980468, - "mean_token_accuracy": 0.954550067782402, - "num_tokens": 7547071.0, - "step": 3150 - }, - { - "entropy": 0.2153193709254265, - "epoch": 7.711700844390832, - "grad_norm": 1.2034279108047485, - "learning_rate": 1.9942965367416116e-05, - "loss": 0.13238741874694823, - "mean_token_accuracy": 0.9535881149768829, - "num_tokens": 7663184.0, - "step": 3200 - }, - { - "entropy": 0.2055867248773575, - "epoch": 7.832328106151991, - "grad_norm": 1.3387084007263184, - "learning_rate": 1.8000046693304666e-05, - "loss": 0.12809998512268067, - "mean_token_accuracy": 0.9540317648649216, - "num_tokens": 7785293.0, - "step": 3250 - }, - { - "entropy": 0.21260966420173644, - "epoch": 7.952955367913148, - "grad_norm": 1.201429009437561, - "learning_rate": 1.6141605548043607e-05, - "loss": 0.13191481590270995, - "mean_token_accuracy": 0.952368374466896, - "num_tokens": 7903616.0, - "step": 3300 - }, - { - "epoch": 8.0, - "eval_entropy": 0.3767001710915833, - "eval_mean_token_accuracy": 0.7909231467193432, - "eval_not_syn_loss": 1.0326374769210815, - "eval_not_syn_runtime": 54.1394, - "eval_not_syn_samples_per_second": 26.229, - "eval_not_syn_steps_per_second": 3.288, - "eval_num_tokens": 7950448.0, - "step": 3320 - }, - { - "epoch": 8.0, - "eval_entropy": 0.35253987854786134, - "eval_mean_token_accuracy": 0.803997924488582, - "eval_num_tokens": 7950448.0, - "eval_syn_loss": 0.9743913412094116, - "eval_syn_runtime": 55.701, - "eval_syn_samples_per_second": 25.493, - "eval_syn_steps_per_second": 3.196, - "step": 3320 - }, - { - "entropy": 0.1920483737732425, - "epoch": 8.072376357056696, - "grad_norm": 1.1402173042297363, - "learning_rate": 1.4370928505069763e-05, - "loss": 0.10957123756408692, - "mean_token_accuracy": 0.9624346013020988, - "num_tokens": 8020867.0, - "step": 3350 - }, - { - "entropy": 0.17265828780829906, - "epoch": 8.193003618817853, - "grad_norm": 1.285122036933899, - "learning_rate": 1.2691146930770208e-05, - "loss": 0.09433144569396973, - "mean_token_accuracy": 0.9681703245639801, - "num_tokens": 8143032.0, - "step": 3400 - }, - { - "entropy": 0.1748460330814123, - "epoch": 8.31363088057901, - "grad_norm": 0.9159325957298279, - "learning_rate": 1.110523144679515e-05, - "loss": 0.09543456077575684, - "mean_token_accuracy": 0.9679662960767746, - "num_tokens": 8262875.0, - "step": 3450 - }, - { - "entropy": 0.16876790300011635, - "epoch": 8.43425814234017, - "grad_norm": 1.0886222124099731, - "learning_rate": 9.615986676640981e-06, - "loss": 0.09403505325317382, - "mean_token_accuracy": 0.9679694646596908, - "num_tokens": 8383147.0, - "step": 3500 - }, - { - "entropy": 0.1674938587099314, - "epoch": 8.554885404101327, - "grad_norm": 0.8731210231781006, - "learning_rate": 8.226046285793963e-06, - "loss": 0.09352604866027832, - "mean_token_accuracy": 0.9685411840677262, - "num_tokens": 8502743.0, - "step": 3550 - }, - { - "entropy": 0.1702917966991663, - "epoch": 8.675512665862485, - "grad_norm": 1.2127553224563599, - "learning_rate": 6.937868324205779e-06, - "loss": 0.09577181816101074, - "mean_token_accuracy": 0.9679328501224518, - "num_tokens": 8621550.0, - "step": 3600 - }, - { - "entropy": 0.16657552666962147, - "epoch": 8.796139927623642, - "grad_norm": 0.7716957330703735, - "learning_rate": 5.7537308793376514e-06, - "loss": 0.09437231063842773, - "mean_token_accuracy": 0.9675484406948089, - "num_tokens": 8742590.0, - "step": 3650 - }, - { - "entropy": 0.16926519222557546, - "epoch": 8.916767189384801, - "grad_norm": 0.9887721538543701, - "learning_rate": 4.675728047460445e-06, - "loss": 0.09573341369628906, - "mean_token_accuracy": 0.9675081825256348, - "num_tokens": 8862250.0, - "step": 3700 - }, - { - "epoch": 9.0, - "eval_entropy": 0.348999947309494, - "eval_mean_token_accuracy": 0.7875944831398096, - "eval_not_syn_loss": 1.15312659740448, - "eval_not_syn_runtime": 54.232, - "eval_not_syn_samples_per_second": 26.184, - "eval_not_syn_steps_per_second": 3.282, - "eval_num_tokens": 8944254.0, - "step": 3735 - }, - { - "epoch": 9.0, - "eval_entropy": 0.32651834206634694, - "eval_mean_token_accuracy": 0.8028985322191474, - "eval_num_tokens": 8944254.0, - "eval_syn_loss": 1.0929256677627563, - "eval_syn_runtime": 56.5213, - "eval_syn_samples_per_second": 25.123, - "eval_syn_steps_per_second": 3.149, - "step": 3735 - } - ], - "logging_steps": 50, - "max_steps": 4150, - "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": 1.8444938839209984e+17, - "train_batch_size": 8, - "trial_name": null, - "trial_params": null -} diff --git a/substitutivity_original_Estonian/Qwen3.5-2B-Base_substitutivity_splits_original_features_train_substitutivity_splits_original_features_test2/checkpoint-3735/training_args.bin b/substitutivity_original_Estonian/Qwen3.5-2B-Base_substitutivity_splits_original_features_train_substitutivity_splits_original_features_test2/checkpoint-3735/training_args.bin deleted file mode 100644 index f3d8098f65400413dd700f32ada5649470871017..0000000000000000000000000000000000000000 --- a/substitutivity_original_Estonian/Qwen3.5-2B-Base_substitutivity_splits_original_features_train_substitutivity_splits_original_features_test2/checkpoint-3735/training_args.bin +++ /dev/null @@ -1,3 +0,0 @@ -version https://git-lfs.github.com/spec/v1 -oid sha256:617af0f41414927baaaf158949585bb75eb4b05ff78baa5216c860ef0b431009 -size 6033 diff --git a/substitutivity_original_Estonian/Qwen3.5-2B-Base_substitutivity_splits_original_features_train_substitutivity_splits_original_features_test2/checkpoint-415/README.md b/substitutivity_original_Estonian/Qwen3.5-2B-Base_substitutivity_splits_original_features_train_substitutivity_splits_original_features_test2/checkpoint-415/README.md deleted file mode 100644 index ece204d005f7bf161d0e448efdd56fd297b7c8a2..0000000000000000000000000000000000000000 --- a/substitutivity_original_Estonian/Qwen3.5-2B-Base_substitutivity_splits_original_features_train_substitutivity_splits_original_features_test2/checkpoint-415/README.md +++ /dev/null @@ -1,209 +0,0 @@ ---- -base_model: Qwen/Qwen3.5-2B-Base -library_name: peft -pipeline_tag: text-generation -tags: -- base_model:adapter:Qwen/Qwen3.5-2B-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/substitutivity_original_Estonian/Qwen3.5-2B-Base_substitutivity_splits_original_features_train_substitutivity_splits_original_features_test2/checkpoint-415/adapter_config.json b/substitutivity_original_Estonian/Qwen3.5-2B-Base_substitutivity_splits_original_features_train_substitutivity_splits_original_features_test2/checkpoint-415/adapter_config.json deleted file mode 100644 index 6a095bb3bfe249ce8e15df6ae999e88fa1e17162..0000000000000000000000000000000000000000 --- a/substitutivity_original_Estonian/Qwen3.5-2B-Base_substitutivity_splits_original_features_train_substitutivity_splits_original_features_test2/checkpoint-415/adapter_config.json +++ /dev/null @@ -1,46 +0,0 @@ -{ - "alora_invocation_tokens": null, - "alpha_pattern": {}, - "arrow_config": null, - "auto_mapping": null, - "base_model_name_or_path": "Qwen/Qwen3.5-2B-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": 0.05731002492184622, - "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": [ - "o_proj", - "q_proj", - "k_proj", - "down_proj", - "up_proj", - "v_proj", - "gate_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/substitutivity_original_Estonian/Qwen3.5-2B-Base_substitutivity_splits_original_features_train_substitutivity_splits_original_features_test2/checkpoint-415/adapter_model.safetensors b/substitutivity_original_Estonian/Qwen3.5-2B-Base_substitutivity_splits_original_features_train_substitutivity_splits_original_features_test2/checkpoint-415/adapter_model.safetensors deleted file mode 100644 index d94f813a077874df832d2baf64e27f5f2af298ee..0000000000000000000000000000000000000000 --- a/substitutivity_original_Estonian/Qwen3.5-2B-Base_substitutivity_splits_original_features_train_substitutivity_splits_original_features_test2/checkpoint-415/adapter_model.safetensors +++ /dev/null @@ -1,3 +0,0 @@ -version https://git-lfs.github.com/spec/v1 -oid sha256:a8d13296e58d06d9b79f7f40b727719e14353424e8a2f574cf02b65ca09de9d7 -size 174613392 diff --git a/substitutivity_original_Estonian/Qwen3.5-2B-Base_substitutivity_splits_original_features_train_substitutivity_splits_original_features_test2/checkpoint-415/chat_template.jinja b/substitutivity_original_Estonian/Qwen3.5-2B-Base_substitutivity_splits_original_features_train_substitutivity_splits_original_features_test2/checkpoint-415/chat_template.jinja deleted file mode 100644 index 0ef09f214eaa6d9bca297988afc1454b5827b2c7..0000000000000000000000000000000000000000 --- a/substitutivity_original_Estonian/Qwen3.5-2B-Base_substitutivity_splits_original_features_train_substitutivity_splits_original_features_test2/checkpoint-415/chat_template.jinja +++ /dev/null @@ -1,154 +0,0 @@ -{%- set image_count = namespace(value=0) %} -{%- set video_count = namespace(value=0) %} -{%- macro render_content(content, do_vision_count, is_system_content=false) %} - {%- if content is string %} - {{- content }} - {%- elif content is iterable and content is not mapping %} - {%- for item in content %} - {%- if 'image' in item or 'image_url' in item or item.type == 'image' %} - {%- if is_system_content %} - {{- raise_exception('System message cannot contain images.') }} - {%- endif %} - {%- if do_vision_count %} - {%- set image_count.value = image_count.value + 1 %} - {%- endif %} - {%- if add_vision_id %} - {{- 'Picture ' ~ image_count.value ~ ': ' }} - {%- endif %} - {{- '<|vision_start|><|image_pad|><|vision_end|>' }} - {%- elif 'video' in item or item.type == 'video' %} - {%- if is_system_content %} - {{- raise_exception('System message cannot contain videos.') }} - {%- endif %} - {%- if do_vision_count %} - {%- set video_count.value = video_count.value + 1 %} - {%- endif %} - {%- if add_vision_id %} - {{- 'Video ' ~ video_count.value ~ ': ' }} - {%- endif %} - {{- '<|vision_start|><|video_pad|><|vision_end|>' }} - {%- elif 'text' in item %} - {{- item.text }} - {%- else %} - {{- raise_exception('Unexpected item type in content.') }} - {%- endif %} - {%- endfor %} - {%- elif content is none or content is undefined %} - {{- '' }} - {%- else %} - {{- raise_exception('Unexpected content type.') }} - {%- endif %} -{%- endmacro %} -{%- if not messages %} - {{- raise_exception('No messages provided.') }} -{%- endif %} -{%- if tools and tools is iterable and tools is not mapping %} - {{- '<|im_start|>system\n' }} - {{- "# Tools\n\nYou have access to the following functions:\n\n" }} - {%- for tool in tools %} - {{- "\n" }} - {{- tool | tojson }} - {%- endfor %} - {{- "\n" }} - {{- '\n\nIf you choose to call a function ONLY reply in the following format with NO suffix:\n\n\n\n\nvalue_1\n\n\nThis is the value for the second parameter\nthat can span\nmultiple lines\n\n\n\n\n\nReminder:\n- Function calls MUST follow the specified format: an inner block must be nested within XML tags\n- Required parameters MUST be specified\n- You may provide optional reasoning for your function call in natural language BEFORE the function call, but NOT after\n- If there is no function call available, answer the question like normal with your current knowledge and do not tell the user about function calls\n' }} - {%- if messages[0].role == 'system' %} - {%- set content = render_content(messages[0].content, false, true)|trim %} - {%- if content %} - {{- '\n\n' + content }} - {%- endif %} - {%- endif %} - {{- '<|im_end|>\n' }} -{%- else %} - {%- if messages[0].role == 'system' %} - {%- set content = render_content(messages[0].content, false, true)|trim %} - {{- '<|im_start|>system\n' + 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" %} - {%- set content = render_content(message.content, false)|trim %} - {%- if not(content.startswith('') and content.endswith('')) %} - {%- set ns.multi_step_tool = false %} - {%- set ns.last_query_index = index %} - {%- endif %} - {%- endif %} -{%- endfor %} -{%- if ns.multi_step_tool %} - {{- raise_exception('No user query found in messages.') }} -{%- endif %} -{%- for message in messages %} - {%- set content = render_content(message.content, true)|trim %} - {%- if message.role == "system" %} - {%- if not loop.first %} - {{- raise_exception('System message must be at the beginning.') }} - {%- endif %} - {%- elif message.role == "user" %} - {{- '<|im_start|>' + message.role + '\n' + content + '<|im_end|>' + '\n' }} - {%- elif message.role == "assistant" %} - {%- set reasoning_content = '' %} - {%- if message.reasoning_content is string %} - {%- set reasoning_content = message.reasoning_content %} - {%- else %} - {%- if '' in content %} - {%- set reasoning_content = content.split('')[0].rstrip('\n').split('')[-1].lstrip('\n') %} - {%- set content = content.split('')[-1].lstrip('\n') %} - {%- endif %} - {%- endif %} - {%- set reasoning_content = reasoning_content|trim %} - {%- if loop.index0 > ns.last_query_index %} - {{- '<|im_start|>' + message.role + '\n\n' + reasoning_content + '\n\n\n' + content }} - {%- else %} - {{- '<|im_start|>' + message.role + '\n' + content }} - {%- endif %} - {%- if message.tool_calls and message.tool_calls is iterable and message.tool_calls is not mapping %} - {%- for tool_call in message.tool_calls %} - {%- if tool_call.function is defined %} - {%- set tool_call = tool_call.function %} - {%- endif %} - {%- if loop.first %} - {%- if content|trim %} - {{- '\n\n\n\n' }} - {%- else %} - {{- '\n\n' }} - {%- endif %} - {%- else %} - {{- '\n\n\n' }} - {%- endif %} - {%- if tool_call.arguments is defined %} - {%- for args_name, args_value in tool_call.arguments|items %} - {{- '\n' }} - {%- set args_value = args_value | tojson | safe if args_value is mapping or (args_value is sequence and args_value is not string) else args_value | string %} - {{- args_value }} - {{- '\n\n' }} - {%- endfor %} - {%- endif %} - {{- '\n' }} - {%- endfor %} - {%- endif %} - {{- '<|im_end|>\n' }} - {%- elif message.role == "tool" %} - {%- if loop.previtem and loop.previtem.role != "tool" %} - {{- '<|im_start|>user' }} - {%- endif %} - {{- '\n\n' }} - {{- content }} - {{- '\n' }} - {%- if not loop.last and loop.nextitem.role != "tool" %} - {{- '<|im_end|>\n' }} - {%- elif loop.last %} - {{- '<|im_end|>\n' }} - {%- endif %} - {%- else %} - {{- raise_exception('Unexpected message role.') }} - {%- endif %} -{%- endfor %} -{%- if add_generation_prompt %} - {{- '<|im_start|>assistant\n' }} - {%- if enable_thinking is defined and enable_thinking is true %} - {{- '\n' }} - {%- else %} - {{- '\n\n\n\n' }} - {%- endif %} -{%- endif %} \ No newline at end of file diff --git a/substitutivity_original_Estonian/Qwen3.5-2B-Base_substitutivity_splits_original_features_train_substitutivity_splits_original_features_test2/checkpoint-415/tokenizer.json b/substitutivity_original_Estonian/Qwen3.5-2B-Base_substitutivity_splits_original_features_train_substitutivity_splits_original_features_test2/checkpoint-415/tokenizer.json deleted file mode 100644 index 67741b04f23bfdb46501f748ce27865ec82eccfb..0000000000000000000000000000000000000000 --- a/substitutivity_original_Estonian/Qwen3.5-2B-Base_substitutivity_splits_original_features_train_substitutivity_splits_original_features_test2/checkpoint-415/tokenizer.json +++ /dev/null @@ -1,3 +0,0 @@ -version https://git-lfs.github.com/spec/v1 -oid sha256:87a7830d63fcf43bf241c3c5242e96e62dd3fdc29224ca26fed8ea333db72de4 -size 19989343 diff --git a/substitutivity_original_Estonian/Qwen3.5-2B-Base_substitutivity_splits_original_features_train_substitutivity_splits_original_features_test2/checkpoint-415/tokenizer_config.json b/substitutivity_original_Estonian/Qwen3.5-2B-Base_substitutivity_splits_original_features_train_substitutivity_splits_original_features_test2/checkpoint-415/tokenizer_config.json deleted file mode 100644 index b4a37b2a6fd3ab3317cd7bac72855be1a843b2bb..0000000000000000000000000000000000000000 --- a/substitutivity_original_Estonian/Qwen3.5-2B-Base_substitutivity_splits_original_features_train_substitutivity_splits_original_features_test2/checkpoint-415/tokenizer_config.json +++ /dev/null @@ -1,31 +0,0 @@ -{ - "add_prefix_space": false, - "audio_bos_token": "<|audio_start|>", - "audio_eos_token": "<|audio_end|>", - "audio_token": "<|audio_pad|>", - "backend": "tokenizers", - "bos_token": null, - "clean_up_tokenization_spaces": false, - "eos_token": "<|endoftext|>", - "errors": "replace", - "image_token": "<|image_pad|>", - "is_local": false, - "model_max_length": 262144, - "model_specific_special_tokens": { - "audio_bos_token": "<|audio_start|>", - "audio_eos_token": "<|audio_end|>", - "audio_token": "<|audio_pad|>", - "image_token": "<|image_pad|>", - "video_token": "<|video_pad|>", - "vision_bos_token": "<|vision_start|>", - "vision_eos_token": "<|vision_end|>" - }, - "pad_token": "<|endoftext|>", - "pretokenize_regex": "(?i:'s|'t|'re|'ve|'m|'ll|'d)|[^\\r\\n\\p{L}\\p{N}]?[\\p{L}\\p{M}]+|\\p{N}| ?[^\\s\\p{L}\\p{M}\\p{N}]+[\\r\\n]*|\\s*[\\r\\n]+|\\s+(?!\\S)|\\s+", - "split_special_tokens": false, - "tokenizer_class": "TokenizersBackend", - "unk_token": null, - "video_token": "<|video_pad|>", - "vision_bos_token": "<|vision_start|>", - "vision_eos_token": "<|vision_end|>" -} diff --git a/substitutivity_original_Estonian/Qwen3.5-2B-Base_substitutivity_splits_original_features_train_substitutivity_splits_original_features_test2/checkpoint-415/trainer_state.json b/substitutivity_original_Estonian/Qwen3.5-2B-Base_substitutivity_splits_original_features_train_substitutivity_splits_original_features_test2/checkpoint-415/trainer_state.json deleted file mode 100644 index bbb11865d8ba147470a539562418fa23df975bc9..0000000000000000000000000000000000000000 --- a/substitutivity_original_Estonian/Qwen3.5-2B-Base_substitutivity_splits_original_features_train_substitutivity_splits_original_features_test2/checkpoint-415/trainer_state.json +++ /dev/null @@ -1,136 +0,0 @@ -{ - "best_global_step": null, - "best_metric": null, - "best_model_checkpoint": null, - "epoch": 1.0, - "eval_steps": 500, - "global_step": 415, - "is_hyper_param_search": false, - "is_local_process_zero": true, - "is_world_process_zero": true, - "log_history": [ - { - "entropy": 2.2933380508422854, - "epoch": 0.12062726176115803, - "grad_norm": 2.8075337409973145, - "learning_rate": 1.5531017374629602e-05, - "loss": 2.1429234313964844, - "mean_token_accuracy": 0.5779684272408485, - "num_tokens": 121428.0, - "step": 50 - }, - { - "entropy": 1.1315062403678895, - "epoch": 0.24125452352231605, - "grad_norm": 1.9415371417999268, - "learning_rate": 3.1378994287516954e-05, - "loss": 1.0461698913574218, - "mean_token_accuracy": 0.7166991305351257, - "num_tokens": 243966.0, - "step": 100 - }, - { - "entropy": 0.9537658470869065, - "epoch": 0.3618817852834741, - "grad_norm": 1.8353157043457031, - "learning_rate": 4.72269712004043e-05, - "loss": 0.8774301147460938, - "mean_token_accuracy": 0.7506789201498032, - "num_tokens": 366649.0, - "step": 150 - }, - { - "entropy": 0.8955666983127594, - "epoch": 0.4825090470446321, - "grad_norm": 1.5419880151748657, - "learning_rate": 6.307494811329165e-05, - "loss": 0.8234302520751953, - "mean_token_accuracy": 0.7616239994764328, - "num_tokens": 488367.0, - "step": 200 - }, - { - "entropy": 0.8689469587802887, - "epoch": 0.6031363088057901, - "grad_norm": 1.5661484003067017, - "learning_rate": 7.8922925026179e-05, - "loss": 0.7961890411376953, - "mean_token_accuracy": 0.767715499997139, - "num_tokens": 603131.0, - "step": 250 - }, - { - "entropy": 0.8425652319192887, - "epoch": 0.7237635705669482, - "grad_norm": 1.5694905519485474, - "learning_rate": 9.477090193906636e-05, - "loss": 0.7718594360351563, - "mean_token_accuracy": 0.7736722755432129, - "num_tokens": 726412.0, - "step": 300 - }, - { - "entropy": 0.843294187784195, - "epoch": 0.8443908323281062, - "grad_norm": 1.1118028163909912, - "learning_rate": 0.00011061887885195371, - "loss": 0.7801297760009765, - "mean_token_accuracy": 0.7714760398864746, - "num_tokens": 846516.0, - "step": 350 - }, - { - "entropy": 0.8213214933872223, - "epoch": 0.9650180940892642, - "grad_norm": 0.9202918410301208, - "learning_rate": 0.00012646685576484106, - "loss": 0.7514453887939453, - "mean_token_accuracy": 0.7774044269323349, - "num_tokens": 958978.0, - "step": 400 - }, - { - "epoch": 1.0, - "eval_entropy": 0.818668900915746, - "eval_mean_token_accuracy": 0.7533496866065464, - "eval_not_syn_loss": 0.7798981070518494, - "eval_not_syn_runtime": 53.8744, - "eval_not_syn_samples_per_second": 26.358, - "eval_not_syn_steps_per_second": 3.304, - "eval_num_tokens": 993806.0, - "step": 415 - }, - { - "epoch": 1.0, - "eval_entropy": 0.7627468125873738, - "eval_mean_token_accuracy": 0.8024111574285486, - "eval_num_tokens": 993806.0, - "eval_syn_loss": 0.7218216061592102, - "eval_syn_runtime": 55.3239, - "eval_syn_samples_per_second": 25.667, - "eval_syn_steps_per_second": 3.217, - "step": 415 - } - ], - "logging_steps": 50, - "max_steps": 4150, - "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": 2.0596135379978496e+16, - "train_batch_size": 8, - "trial_name": null, - "trial_params": null -} diff --git a/substitutivity_original_Estonian/Qwen3.5-2B-Base_substitutivity_splits_original_features_train_substitutivity_splits_original_features_test2/checkpoint-415/training_args.bin b/substitutivity_original_Estonian/Qwen3.5-2B-Base_substitutivity_splits_original_features_train_substitutivity_splits_original_features_test2/checkpoint-415/training_args.bin deleted file mode 100644 index f3d8098f65400413dd700f32ada5649470871017..0000000000000000000000000000000000000000 --- a/substitutivity_original_Estonian/Qwen3.5-2B-Base_substitutivity_splits_original_features_train_substitutivity_splits_original_features_test2/checkpoint-415/training_args.bin +++ /dev/null @@ -1,3 +0,0 @@ -version https://git-lfs.github.com/spec/v1 -oid sha256:617af0f41414927baaaf158949585bb75eb4b05ff78baa5216c860ef0b431009 -size 6033 diff --git a/substitutivity_original_Estonian/Qwen3.5-2B-Base_substitutivity_splits_original_features_train_substitutivity_splits_original_features_test2/checkpoint-4150/README.md b/substitutivity_original_Estonian/Qwen3.5-2B-Base_substitutivity_splits_original_features_train_substitutivity_splits_original_features_test2/checkpoint-4150/README.md deleted file mode 100644 index ece204d005f7bf161d0e448efdd56fd297b7c8a2..0000000000000000000000000000000000000000 --- a/substitutivity_original_Estonian/Qwen3.5-2B-Base_substitutivity_splits_original_features_train_substitutivity_splits_original_features_test2/checkpoint-4150/README.md +++ /dev/null @@ -1,209 +0,0 @@ ---- -base_model: Qwen/Qwen3.5-2B-Base -library_name: peft -pipeline_tag: text-generation -tags: -- base_model:adapter:Qwen/Qwen3.5-2B-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/substitutivity_original_Estonian/Qwen3.5-2B-Base_substitutivity_splits_original_features_train_substitutivity_splits_original_features_test2/checkpoint-4150/adapter_config.json b/substitutivity_original_Estonian/Qwen3.5-2B-Base_substitutivity_splits_original_features_train_substitutivity_splits_original_features_test2/checkpoint-4150/adapter_config.json deleted file mode 100644 index 6a095bb3bfe249ce8e15df6ae999e88fa1e17162..0000000000000000000000000000000000000000 --- a/substitutivity_original_Estonian/Qwen3.5-2B-Base_substitutivity_splits_original_features_train_substitutivity_splits_original_features_test2/checkpoint-4150/adapter_config.json +++ /dev/null @@ -1,46 +0,0 @@ -{ - "alora_invocation_tokens": null, - "alpha_pattern": {}, - "arrow_config": null, - "auto_mapping": null, - "base_model_name_or_path": "Qwen/Qwen3.5-2B-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": 0.05731002492184622, - "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": [ - "o_proj", - "q_proj", - "k_proj", - "down_proj", - "up_proj", - "v_proj", - "gate_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/substitutivity_original_Estonian/Qwen3.5-2B-Base_substitutivity_splits_original_features_train_substitutivity_splits_original_features_test2/checkpoint-4150/adapter_model.safetensors b/substitutivity_original_Estonian/Qwen3.5-2B-Base_substitutivity_splits_original_features_train_substitutivity_splits_original_features_test2/checkpoint-4150/adapter_model.safetensors deleted file mode 100644 index e3e5231513a61d1c9a4507b18120656a2fbaee66..0000000000000000000000000000000000000000 --- a/substitutivity_original_Estonian/Qwen3.5-2B-Base_substitutivity_splits_original_features_train_substitutivity_splits_original_features_test2/checkpoint-4150/adapter_model.safetensors +++ /dev/null @@ -1,3 +0,0 @@ -version https://git-lfs.github.com/spec/v1 -oid sha256:f96ea5edb05fd6a156e34b71e3250c108cdf74b5760924b0ec133e3527874a6d -size 174613392 diff --git a/substitutivity_original_Estonian/Qwen3.5-2B-Base_substitutivity_splits_original_features_train_substitutivity_splits_original_features_test2/checkpoint-4150/chat_template.jinja b/substitutivity_original_Estonian/Qwen3.5-2B-Base_substitutivity_splits_original_features_train_substitutivity_splits_original_features_test2/checkpoint-4150/chat_template.jinja deleted file mode 100644 index 0ef09f214eaa6d9bca297988afc1454b5827b2c7..0000000000000000000000000000000000000000 --- a/substitutivity_original_Estonian/Qwen3.5-2B-Base_substitutivity_splits_original_features_train_substitutivity_splits_original_features_test2/checkpoint-4150/chat_template.jinja +++ /dev/null @@ -1,154 +0,0 @@ -{%- set image_count = namespace(value=0) %} -{%- set video_count = namespace(value=0) %} -{%- macro render_content(content, do_vision_count, is_system_content=false) %} - {%- if content is string %} - {{- content }} - {%- elif content is iterable and content is not mapping %} - {%- for item in content %} - {%- if 'image' in item or 'image_url' in item or item.type == 'image' %} - {%- if is_system_content %} - {{- raise_exception('System message cannot contain images.') }} - {%- endif %} - {%- if do_vision_count %} - {%- set image_count.value = image_count.value + 1 %} - {%- endif %} - {%- if add_vision_id %} - {{- 'Picture ' ~ image_count.value ~ ': ' }} - {%- endif %} - {{- '<|vision_start|><|image_pad|><|vision_end|>' }} - {%- elif 'video' in item or item.type == 'video' %} - {%- if is_system_content %} - {{- raise_exception('System message cannot contain videos.') }} - {%- endif %} - {%- if do_vision_count %} - {%- set video_count.value = video_count.value + 1 %} - {%- endif %} - {%- if add_vision_id %} - {{- 'Video ' ~ video_count.value ~ ': ' }} - {%- endif %} - {{- '<|vision_start|><|video_pad|><|vision_end|>' }} - {%- elif 'text' in item %} - {{- item.text }} - {%- else %} - {{- raise_exception('Unexpected item type in content.') }} - {%- endif %} - {%- endfor %} - {%- elif content is none or content is undefined %} - {{- '' }} - {%- else %} - {{- raise_exception('Unexpected content type.') }} - {%- endif %} -{%- endmacro %} -{%- if not messages %} - {{- raise_exception('No messages provided.') }} -{%- endif %} -{%- if tools and tools is iterable and tools is not mapping %} - {{- '<|im_start|>system\n' }} - {{- "# Tools\n\nYou have access to the following functions:\n\n" }} - {%- for tool in tools %} - {{- "\n" }} - {{- tool | tojson }} - {%- endfor %} - {{- "\n" }} - {{- '\n\nIf you choose to call a function ONLY reply in the following format with NO suffix:\n\n\n\n\nvalue_1\n\n\nThis is the value for the second parameter\nthat can span\nmultiple lines\n\n\n\n\n\nReminder:\n- Function calls MUST follow the specified format: an inner block must be nested within XML tags\n- Required parameters MUST be specified\n- You may provide optional reasoning for your function call in natural language BEFORE the function call, but NOT after\n- If there is no function call available, answer the question like normal with your current knowledge and do not tell the user about function calls\n' }} - {%- if messages[0].role == 'system' %} - {%- set content = render_content(messages[0].content, false, true)|trim %} - {%- if content %} - {{- '\n\n' + content }} - {%- endif %} - {%- endif %} - {{- '<|im_end|>\n' }} -{%- else %} - {%- if messages[0].role == 'system' %} - {%- set content = render_content(messages[0].content, false, true)|trim %} - {{- '<|im_start|>system\n' + 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" %} - {%- set content = render_content(message.content, false)|trim %} - {%- if not(content.startswith('') and content.endswith('')) %} - {%- set ns.multi_step_tool = false %} - {%- set ns.last_query_index = index %} - {%- endif %} - {%- endif %} -{%- endfor %} -{%- if ns.multi_step_tool %} - {{- raise_exception('No user query found in messages.') }} -{%- endif %} -{%- for message in messages %} - {%- set content = render_content(message.content, true)|trim %} - {%- if message.role == "system" %} - {%- if not loop.first %} - {{- raise_exception('System message must be at the beginning.') }} - {%- endif %} - {%- elif message.role == "user" %} - {{- '<|im_start|>' + message.role + '\n' + content + '<|im_end|>' + '\n' }} - {%- elif message.role == "assistant" %} - {%- set reasoning_content = '' %} - {%- if message.reasoning_content is string %} - {%- set reasoning_content = message.reasoning_content %} - {%- else %} - {%- if '' in content %} - {%- set reasoning_content = content.split('')[0].rstrip('\n').split('')[-1].lstrip('\n') %} - {%- set content = content.split('')[-1].lstrip('\n') %} - {%- endif %} - {%- endif %} - {%- set reasoning_content = reasoning_content|trim %} - {%- if loop.index0 > ns.last_query_index %} - {{- '<|im_start|>' + message.role + '\n\n' + reasoning_content + '\n\n\n' + content }} - {%- else %} - {{- '<|im_start|>' + message.role + '\n' + content }} - {%- endif %} - {%- if message.tool_calls and message.tool_calls is iterable and message.tool_calls is not mapping %} - {%- for tool_call in message.tool_calls %} - {%- if tool_call.function is defined %} - {%- set tool_call = tool_call.function %} - {%- endif %} - {%- if loop.first %} - {%- if content|trim %} - {{- '\n\n\n\n' }} - {%- else %} - {{- '\n\n' }} - {%- endif %} - {%- else %} - {{- '\n\n\n' }} - {%- endif %} - {%- if tool_call.arguments is defined %} - {%- for args_name, args_value in tool_call.arguments|items %} - {{- '\n' }} - {%- set args_value = args_value | tojson | safe if args_value is mapping or (args_value is sequence and args_value is not string) else args_value | string %} - {{- args_value }} - {{- '\n\n' }} - {%- endfor %} - {%- endif %} - {{- '\n' }} - {%- endfor %} - {%- endif %} - {{- '<|im_end|>\n' }} - {%- elif message.role == "tool" %} - {%- if loop.previtem and loop.previtem.role != "tool" %} - {{- '<|im_start|>user' }} - {%- endif %} - {{- '\n\n' }} - {{- content }} - {{- '\n' }} - {%- if not loop.last and loop.nextitem.role != "tool" %} - {{- '<|im_end|>\n' }} - {%- elif loop.last %} - {{- '<|im_end|>\n' }} - {%- endif %} - {%- else %} - {{- raise_exception('Unexpected message role.') }} - {%- endif %} -{%- endfor %} -{%- if add_generation_prompt %} - {{- '<|im_start|>assistant\n' }} - {%- if enable_thinking is defined and enable_thinking is true %} - {{- '\n' }} - {%- else %} - {{- '\n\n\n\n' }} - {%- endif %} -{%- endif %} \ No newline at end of file diff --git a/substitutivity_original_Estonian/Qwen3.5-2B-Base_substitutivity_splits_original_features_train_substitutivity_splits_original_features_test2/checkpoint-4150/tokenizer.json b/substitutivity_original_Estonian/Qwen3.5-2B-Base_substitutivity_splits_original_features_train_substitutivity_splits_original_features_test2/checkpoint-4150/tokenizer.json deleted file mode 100644 index 67741b04f23bfdb46501f748ce27865ec82eccfb..0000000000000000000000000000000000000000 --- a/substitutivity_original_Estonian/Qwen3.5-2B-Base_substitutivity_splits_original_features_train_substitutivity_splits_original_features_test2/checkpoint-4150/tokenizer.json +++ /dev/null @@ -1,3 +0,0 @@ -version https://git-lfs.github.com/spec/v1 -oid sha256:87a7830d63fcf43bf241c3c5242e96e62dd3fdc29224ca26fed8ea333db72de4 -size 19989343 diff --git a/substitutivity_original_Estonian/Qwen3.5-2B-Base_substitutivity_splits_original_features_train_substitutivity_splits_original_features_test2/checkpoint-4150/tokenizer_config.json b/substitutivity_original_Estonian/Qwen3.5-2B-Base_substitutivity_splits_original_features_train_substitutivity_splits_original_features_test2/checkpoint-4150/tokenizer_config.json deleted file mode 100644 index b4a37b2a6fd3ab3317cd7bac72855be1a843b2bb..0000000000000000000000000000000000000000 --- a/substitutivity_original_Estonian/Qwen3.5-2B-Base_substitutivity_splits_original_features_train_substitutivity_splits_original_features_test2/checkpoint-4150/tokenizer_config.json +++ /dev/null @@ -1,31 +0,0 @@ -{ - "add_prefix_space": false, - "audio_bos_token": "<|audio_start|>", - "audio_eos_token": "<|audio_end|>", - "audio_token": "<|audio_pad|>", - "backend": "tokenizers", - "bos_token": null, - "clean_up_tokenization_spaces": false, - "eos_token": "<|endoftext|>", - "errors": "replace", - "image_token": "<|image_pad|>", - "is_local": false, - "model_max_length": 262144, - "model_specific_special_tokens": { - "audio_bos_token": "<|audio_start|>", - "audio_eos_token": "<|audio_end|>", - "audio_token": "<|audio_pad|>", - "image_token": "<|image_pad|>", - "video_token": "<|video_pad|>", - "vision_bos_token": "<|vision_start|>", - "vision_eos_token": "<|vision_end|>" - }, - "pad_token": "<|endoftext|>", - "pretokenize_regex": "(?i:'s|'t|'re|'ve|'m|'ll|'d)|[^\\r\\n\\p{L}\\p{N}]?[\\p{L}\\p{M}]+|\\p{N}| ?[^\\s\\p{L}\\p{M}\\p{N}]+[\\r\\n]*|\\s*[\\r\\n]+|\\s+(?!\\S)|\\s+", - "split_special_tokens": false, - "tokenizer_class": "TokenizersBackend", - "unk_token": null, - "video_token": "<|video_pad|>", - "vision_bos_token": "<|vision_start|>", - "vision_eos_token": "<|vision_end|>" -} diff --git a/substitutivity_original_Estonian/Qwen3.5-2B-Base_substitutivity_splits_original_features_train_substitutivity_splits_original_features_test2/checkpoint-4150/trainer_state.json b/substitutivity_original_Estonian/Qwen3.5-2B-Base_substitutivity_splits_original_features_train_substitutivity_splits_original_features_test2/checkpoint-4150/trainer_state.json deleted file mode 100644 index abc63c7060170e47ad8a8913983d40d8b61d1948..0000000000000000000000000000000000000000 --- a/substitutivity_original_Estonian/Qwen3.5-2B-Base_substitutivity_splits_original_features_train_substitutivity_splits_original_features_test2/checkpoint-4150/trainer_state.json +++ /dev/null @@ -1,1084 +0,0 @@ -{ - "best_global_step": null, - "best_metric": null, - "best_model_checkpoint": null, - "epoch": 10.0, - "eval_steps": 500, - "global_step": 4150, - "is_hyper_param_search": false, - "is_local_process_zero": true, - "is_world_process_zero": true, - "log_history": [ - { - "entropy": 2.2933380508422854, - "epoch": 0.12062726176115803, - "grad_norm": 2.8075337409973145, - "learning_rate": 1.5531017374629602e-05, - "loss": 2.1429234313964844, - "mean_token_accuracy": 0.5779684272408485, - "num_tokens": 121428.0, - "step": 50 - }, - { - "entropy": 1.1315062403678895, - "epoch": 0.24125452352231605, - "grad_norm": 1.9415371417999268, - "learning_rate": 3.1378994287516954e-05, - "loss": 1.0461698913574218, - "mean_token_accuracy": 0.7166991305351257, - "num_tokens": 243966.0, - "step": 100 - }, - { - "entropy": 0.9537658470869065, - "epoch": 0.3618817852834741, - "grad_norm": 1.8353157043457031, - "learning_rate": 4.72269712004043e-05, - "loss": 0.8774301147460938, - "mean_token_accuracy": 0.7506789201498032, - "num_tokens": 366649.0, - "step": 150 - }, - { - "entropy": 0.8955666983127594, - "epoch": 0.4825090470446321, - "grad_norm": 1.5419880151748657, - "learning_rate": 6.307494811329165e-05, - "loss": 0.8234302520751953, - "mean_token_accuracy": 0.7616239994764328, - "num_tokens": 488367.0, - "step": 200 - }, - { - "entropy": 0.8689469587802887, - "epoch": 0.6031363088057901, - "grad_norm": 1.5661484003067017, - "learning_rate": 7.8922925026179e-05, - "loss": 0.7961890411376953, - "mean_token_accuracy": 0.767715499997139, - "num_tokens": 603131.0, - "step": 250 - }, - { - "entropy": 0.8425652319192887, - "epoch": 0.7237635705669482, - "grad_norm": 1.5694905519485474, - "learning_rate": 9.477090193906636e-05, - "loss": 0.7718594360351563, - "mean_token_accuracy": 0.7736722755432129, - "num_tokens": 726412.0, - "step": 300 - }, - { - "entropy": 0.843294187784195, - "epoch": 0.8443908323281062, - "grad_norm": 1.1118028163909912, - "learning_rate": 0.00011061887885195371, - "loss": 0.7801297760009765, - "mean_token_accuracy": 0.7714760398864746, - "num_tokens": 846516.0, - "step": 350 - }, - { - "entropy": 0.8213214933872223, - "epoch": 0.9650180940892642, - "grad_norm": 0.9202918410301208, - "learning_rate": 0.00012646685576484106, - "loss": 0.7514453887939453, - "mean_token_accuracy": 0.7774044269323349, - "num_tokens": 958978.0, - "step": 400 - }, - { - "epoch": 1.0, - "eval_entropy": 0.818668900915746, - "eval_mean_token_accuracy": 0.7533496866065464, - "eval_not_syn_loss": 0.7798981070518494, - "eval_not_syn_runtime": 53.8744, - "eval_not_syn_samples_per_second": 26.358, - "eval_not_syn_steps_per_second": 3.304, - "eval_num_tokens": 993806.0, - "step": 415 - }, - { - "epoch": 1.0, - "eval_entropy": 0.7627468125873738, - "eval_mean_token_accuracy": 0.8024111574285486, - "eval_num_tokens": 993806.0, - "eval_syn_loss": 0.7218216061592102, - "eval_syn_runtime": 55.3239, - "eval_syn_samples_per_second": 25.667, - "eval_syn_steps_per_second": 3.217, - "step": 415 - }, - { - "entropy": 0.7802230679627621, - "epoch": 1.0844390832328106, - "grad_norm": 1.0145268440246582, - "learning_rate": 0.00013151131541121804, - "loss": 0.7111608123779297, - "mean_token_accuracy": 0.7845737542768921, - "num_tokens": 1080301.0, - "step": 450 - }, - { - "entropy": 0.7558828103542328, - "epoch": 1.2050663449939687, - "grad_norm": 1.0477591753005981, - "learning_rate": 0.0001313741160170038, - "loss": 0.6933037567138672, - "mean_token_accuracy": 0.7889221370220184, - "num_tokens": 1201215.0, - "step": 500 - }, - { - "entropy": 0.7593123841285706, - "epoch": 1.3256936067551266, - "grad_norm": 0.899161159992218, - "learning_rate": 0.00013112089697387701, - "loss": 0.6962493133544921, - "mean_token_accuracy": 0.7877359414100646, - "num_tokens": 1322895.0, - "step": 550 - }, - { - "entropy": 0.7460986822843552, - "epoch": 1.4463208685162847, - "grad_norm": 1.1061116456985474, - "learning_rate": 0.00013075210608885215, - "loss": 0.6817412567138672, - "mean_token_accuracy": 0.7904874366521836, - "num_tokens": 1436749.0, - "step": 600 - }, - { - "entropy": 0.735516265630722, - "epoch": 1.5669481302774426, - "grad_norm": 0.9925405979156494, - "learning_rate": 0.00013026839555278796, - "loss": 0.6690705871582031, - "mean_token_accuracy": 0.7930250608921051, - "num_tokens": 1555400.0, - "step": 650 - }, - { - "entropy": 0.7302519899606704, - "epoch": 1.6875753920386007, - "grad_norm": 0.800025463104248, - "learning_rate": 0.0001296706207870159, - "loss": 0.6714833831787109, - "mean_token_accuracy": 0.7946307575702667, - "num_tokens": 1676892.0, - "step": 700 - }, - { - "entropy": 0.7321263110637665, - "epoch": 1.8082026537997589, - "grad_norm": 1.0500682592391968, - "learning_rate": 0.00012895983893056423, - "loss": 0.6672868347167968, - "mean_token_accuracy": 0.7947645837068558, - "num_tokens": 1802689.0, - "step": 750 - }, - { - "entropy": 0.7075755137205124, - "epoch": 1.9288299155609168, - "grad_norm": 0.9248411655426025, - "learning_rate": 0.00012813730697065298, - "loss": 0.6466568756103516, - "mean_token_accuracy": 0.798747689127922, - "num_tokens": 1920804.0, - "step": 800 - }, - { - "epoch": 2.0, - "eval_entropy": 0.7426781811740961, - "eval_mean_token_accuracy": 0.7693507239389955, - "eval_not_syn_loss": 0.7139139175415039, - "eval_not_syn_runtime": 54.4286, - "eval_not_syn_samples_per_second": 26.089, - "eval_not_syn_steps_per_second": 3.27, - "eval_num_tokens": 1987612.0, - "step": 830 - }, - { - "epoch": 2.0, - "eval_entropy": 0.6937930352902144, - "eval_mean_token_accuracy": 0.8180048633157537, - "eval_num_tokens": 1987612.0, - "eval_syn_loss": 0.6489660143852234, - "eval_syn_runtime": 56.1525, - "eval_syn_samples_per_second": 25.288, - "eval_syn_steps_per_second": 3.17, - "step": 830 - }, - { - "entropy": 0.6853217440422135, - "epoch": 2.0482509047044632, - "grad_norm": 0.6932648420333862, - "learning_rate": 0.00012720447951976603, - "loss": 0.6212613296508789, - "mean_token_accuracy": 0.8047762579388089, - "num_tokens": 2039660.0, - "step": 850 - }, - { - "entropy": 0.6459032052755356, - "epoch": 2.168878166465621, - "grad_norm": 1.0688549280166626, - "learning_rate": 0.00012616300624323124, - "loss": 0.5740525817871094, - "mean_token_accuracy": 0.8148404061794281, - "num_tokens": 2161772.0, - "step": 900 - }, - { - "entropy": 0.6463843190670013, - "epoch": 2.2895054282267795, - "grad_norm": 0.847271203994751, - "learning_rate": 0.0001250147289418581, - "loss": 0.577717170715332, - "mean_token_accuracy": 0.814143658876419, - "num_tokens": 2280833.0, - "step": 950 - }, - { - "entropy": 0.6461306965351105, - "epoch": 2.4101326899879374, - "grad_norm": 0.9998957514762878, - "learning_rate": 0.00012376167829479217, - "loss": 0.5813011932373047, - "mean_token_accuracy": 0.8144401633739471, - "num_tokens": 2398206.0, - "step": 1000 - }, - { - "entropy": 0.640914659500122, - "epoch": 2.5307599517490953, - "grad_norm": 0.8692949414253235, - "learning_rate": 0.0001224060702683459, - "loss": 0.5755012512207032, - "mean_token_accuracy": 0.8168018108606339, - "num_tokens": 2518888.0, - "step": 1050 - }, - { - "entropy": 0.6421274638175964, - "epoch": 2.651387213510253, - "grad_norm": 0.8920196294784546, - "learning_rate": 0.00012095030219715738, - "loss": 0.5785483932495117, - "mean_token_accuracy": 0.8137972563505173, - "num_tokens": 2639763.0, - "step": 1100 - }, - { - "entropy": 0.6466812592744827, - "epoch": 2.772014475271411, - "grad_norm": 0.6876946687698364, - "learning_rate": 0.00011939694854460696, - "loss": 0.5816070938110351, - "mean_token_accuracy": 0.8145407867431641, - "num_tokens": 2755540.0, - "step": 1150 - }, - { - "entropy": 0.6378493067622185, - "epoch": 2.8926417370325694, - "grad_norm": 0.7532753944396973, - "learning_rate": 0.00011774875634998911, - "loss": 0.5754346084594727, - "mean_token_accuracy": 0.8148861181735992, - "num_tokens": 2874118.0, - "step": 1200 - }, - { - "epoch": 3.0, - "eval_entropy": 0.6912222209941136, - "eval_mean_token_accuracy": 0.7868467750174276, - "eval_not_syn_loss": 0.6782422661781311, - "eval_not_syn_runtime": 53.2627, - "eval_not_syn_samples_per_second": 26.66, - "eval_not_syn_steps_per_second": 3.342, - "eval_num_tokens": 2981418.0, - "step": 1245 - }, - { - "epoch": 3.0, - "eval_entropy": 0.6467212934172555, - "eval_mean_token_accuracy": 0.8215091817834405, - "eval_num_tokens": 2981418.0, - "eval_syn_loss": 0.6301999092102051, - "eval_syn_runtime": 55.0115, - "eval_syn_samples_per_second": 25.813, - "eval_syn_steps_per_second": 3.236, - "step": 1245 - }, - { - "entropy": 0.6316030452949832, - "epoch": 3.012062726176116, - "grad_norm": 0.7761592864990234, - "learning_rate": 0.00011600864037049124, - "loss": 0.5697880935668945, - "mean_token_accuracy": 0.8159356400220081, - "num_tokens": 2994602.0, - "step": 1250 - }, - { - "entropy": 0.5467158386111259, - "epoch": 3.132689987937274, - "grad_norm": 0.8169087767601013, - "learning_rate": 0.00011417967792657078, - "loss": 0.47844127655029295, - "mean_token_accuracy": 0.8378052765130997, - "num_tokens": 3121094.0, - "step": 1300 - }, - { - "entropy": 0.5454566389322281, - "epoch": 3.2533172496984317, - "grad_norm": 0.8219636678695679, - "learning_rate": 0.0001122651034598459, - "loss": 0.4839338302612305, - "mean_token_accuracy": 0.8385071897506714, - "num_tokens": 3248038.0, - "step": 1350 - }, - { - "entropy": 0.5598723018169403, - "epoch": 3.37394451145959, - "grad_norm": 0.9956360459327698, - "learning_rate": 0.00011026830281312423, - "loss": 0.492743034362793, - "mean_token_accuracy": 0.8357370454072952, - "num_tokens": 3363487.0, - "step": 1400 - }, - { - "entropy": 0.5479365903139114, - "epoch": 3.494571773220748, - "grad_norm": 0.8586823344230652, - "learning_rate": 0.00010819280724268524, - "loss": 0.4810820770263672, - "mean_token_accuracy": 0.837928666472435, - "num_tokens": 3480513.0, - "step": 1450 - }, - { - "entropy": 0.5619524419307709, - "epoch": 3.615199034981906, - "grad_norm": 1.0894954204559326, - "learning_rate": 0.0001060422871734048, - "loss": 0.4974747085571289, - "mean_token_accuracy": 0.8349784809350967, - "num_tokens": 3597183.0, - "step": 1500 - }, - { - "entropy": 0.5588027596473694, - "epoch": 3.7358262967430638, - "grad_norm": 0.9905257225036621, - "learning_rate": 0.00010382054570776654, - "loss": 0.49687202453613283, - "mean_token_accuracy": 0.834458841085434, - "num_tokens": 3714173.0, - "step": 1550 - }, - { - "entropy": 0.5654991599917412, - "epoch": 3.856453558504222, - "grad_norm": 1.0116610527038574, - "learning_rate": 0.00010153151190023791, - "loss": 0.5016670989990234, - "mean_token_accuracy": 0.8326652735471726, - "num_tokens": 3832652.0, - "step": 1600 - }, - { - "entropy": 0.5605382058024406, - "epoch": 3.97708082026538, - "grad_norm": 0.8011100888252258, - "learning_rate": 9.917923380890624e-05, - "loss": 0.49551734924316404, - "mean_token_accuracy": 0.8343535929918289, - "num_tokens": 3955167.0, - "step": 1650 - }, - { - "epoch": 4.0, - "eval_entropy": 0.5819476859623127, - "eval_mean_token_accuracy": 0.7963069781158747, - "eval_not_syn_loss": 0.6924306750297546, - "eval_not_syn_runtime": 54.027, - "eval_not_syn_samples_per_second": 26.283, - "eval_not_syn_steps_per_second": 3.295, - "eval_num_tokens": 3975224.0, - "step": 1660 - }, - { - "epoch": 4.0, - "eval_entropy": 0.545470636212424, - "eval_mean_token_accuracy": 0.8127643666910321, - "eval_num_tokens": 3975224.0, - "eval_syn_loss": 0.649576723575592, - "eval_syn_runtime": 55.2296, - "eval_syn_samples_per_second": 25.711, - "eval_syn_steps_per_second": 3.223, - "step": 1660 - }, - { - "entropy": 0.4920780294471317, - "epoch": 4.0965018094089265, - "grad_norm": 0.8362822532653809, - "learning_rate": 9.676787133666148e-05, - "loss": 0.4147996139526367, - "mean_token_accuracy": 0.8581844628459275, - "num_tokens": 4072361.0, - "step": 1700 - }, - { - "entropy": 0.4675139603018761, - "epoch": 4.217129071170085, - "grad_norm": 0.9920316338539124, - "learning_rate": 9.430168887458645e-05, - "loss": 0.3863528823852539, - "mean_token_accuracy": 0.865827117562294, - "num_tokens": 4194978.0, - "step": 1750 - }, - { - "entropy": 0.47163557589054106, - "epoch": 4.337756332931242, - "grad_norm": 1.0789282321929932, - "learning_rate": 9.178504776056418e-05, - "loss": 0.390842170715332, - "mean_token_accuracy": 0.8634983265399933, - "num_tokens": 4315548.0, - "step": 1800 - }, - { - "entropy": 0.47117673605680466, - "epoch": 4.458383594692401, - "grad_norm": 1.0369561910629272, - "learning_rate": 8.922239856643896e-05, - "loss": 0.39387256622314454, - "mean_token_accuracy": 0.8625931572914124, - "num_tokens": 4431469.0, - "step": 1850 - }, - { - "entropy": 0.479859294295311, - "epoch": 4.579010856453559, - "grad_norm": 1.0827305316925049, - "learning_rate": 8.661827322737095e-05, - "loss": 0.40195709228515625, - "mean_token_accuracy": 0.8612629306316376, - "num_tokens": 4551803.0, - "step": 1900 - }, - { - "entropy": 0.4729380664229393, - "epoch": 4.699638118214716, - "grad_norm": 1.0114246606826782, - "learning_rate": 8.397727702730326e-05, - "loss": 0.39648338317871096, - "mean_token_accuracy": 0.8610477095842362, - "num_tokens": 4674468.0, - "step": 1950 - }, - { - "entropy": 0.4903118047118187, - "epoch": 4.820265379975875, - "grad_norm": 1.0115467309951782, - "learning_rate": 8.130408045471496e-05, - "loss": 0.4067733001708984, - "mean_token_accuracy": 0.8597196912765503, - "num_tokens": 4788027.0, - "step": 2000 - }, - { - "entropy": 0.48099748432636263, - "epoch": 4.940892641737032, - "grad_norm": 1.0201005935668945, - "learning_rate": 7.860341094306242e-05, - "loss": 0.4030665588378906, - "mean_token_accuracy": 0.8603792876005173, - "num_tokens": 4907561.0, - "step": 2050 - }, - { - "epoch": 5.0, - "eval_entropy": 0.5492983292997553, - "eval_mean_token_accuracy": 0.796870938847574, - "eval_not_syn_loss": 0.7129728198051453, - "eval_not_syn_runtime": 53.9575, - "eval_not_syn_samples_per_second": 26.317, - "eval_not_syn_steps_per_second": 3.299, - "eval_num_tokens": 4969030.0, - "step": 2075 - }, - { - "epoch": 5.0, - "eval_entropy": 0.513823946037989, - "eval_mean_token_accuracy": 0.8174765464964877, - "eval_num_tokens": 4969030.0, - "eval_syn_loss": 0.6699439883232117, - "eval_syn_runtime": 56.0033, - "eval_syn_samples_per_second": 25.356, - "eval_syn_steps_per_second": 3.178, - "step": 2075 - }, - { - "entropy": 0.43584728511897003, - "epoch": 5.060313630880579, - "grad_norm": 0.9914600253105164, - "learning_rate": 7.588004451051618e-05, - "loss": 0.34999092102050783, - "mean_token_accuracy": 0.8777085295831314, - "num_tokens": 5026625.0, - "step": 2100 - }, - { - "entropy": 0.38114599585533143, - "epoch": 5.180940892641737, - "grad_norm": 1.232283115386963, - "learning_rate": 7.313879731377764e-05, - "loss": 0.2903956604003906, - "mean_token_accuracy": 0.896172685623169, - "num_tokens": 5141010.0, - "step": 2150 - }, - { - "entropy": 0.3792745009064674, - "epoch": 5.301568154402895, - "grad_norm": 1.2843542098999023, - "learning_rate": 7.03845171309126e-05, - "loss": 0.2903476333618164, - "mean_token_accuracy": 0.8951605755090714, - "num_tokens": 5261240.0, - "step": 2200 - }, - { - "entropy": 0.3744692197442055, - "epoch": 5.422195416164053, - "grad_norm": 1.23075270652771, - "learning_rate": 6.76220747882637e-05, - "loss": 0.29011749267578124, - "mean_token_accuracy": 0.8963996982574463, - "num_tokens": 5382160.0, - "step": 2250 - }, - { - "entropy": 0.3795280712842941, - "epoch": 5.542822677925211, - "grad_norm": 1.4011399745941162, - "learning_rate": 6.485635554660295e-05, - "loss": 0.2921779823303223, - "mean_token_accuracy": 0.8954424083232879, - "num_tokens": 5502002.0, - "step": 2300 - }, - { - "entropy": 0.3797056859731674, - "epoch": 5.66344993968637, - "grad_norm": 1.1022812128067017, - "learning_rate": 6.209225046175775e-05, - "loss": 0.29343433380126954, - "mean_token_accuracy": 0.8939031893014908, - "num_tokens": 5629969.0, - "step": 2350 - }, - { - "entropy": 0.3827402114868164, - "epoch": 5.784077201447527, - "grad_norm": 1.22834312915802, - "learning_rate": 5.9334647734988365e-05, - "loss": 0.2954188537597656, - "mean_token_accuracy": 0.8937842601537704, - "num_tokens": 5750596.0, - "step": 2400 - }, - { - "entropy": 0.38422658771276474, - "epoch": 5.904704463208685, - "grad_norm": 1.425066351890564, - "learning_rate": 5.6588424068413775e-05, - "loss": 0.296480770111084, - "mean_token_accuracy": 0.8936110073328019, - "num_tokens": 5868110.0, - "step": 2450 - }, - { - "epoch": 6.0, - "eval_entropy": 0.49826224684045556, - "eval_mean_token_accuracy": 0.795356297760867, - "eval_not_syn_loss": 0.7700047492980957, - "eval_not_syn_runtime": 53.0632, - "eval_not_syn_samples_per_second": 26.761, - "eval_not_syn_steps_per_second": 3.354, - "eval_num_tokens": 5962836.0, - "step": 2490 - }, - { - "epoch": 6.0, - "eval_entropy": 0.4654735526007213, - "eval_mean_token_accuracy": 0.8118122223387944, - "eval_num_tokens": 5962836.0, - "eval_syn_loss": 0.7288840413093567, - "eval_syn_runtime": 54.6517, - "eval_syn_samples_per_second": 25.983, - "eval_syn_steps_per_second": 3.257, - "step": 2490 - }, - { - "entropy": 0.3596555646320786, - "epoch": 6.024125452352232, - "grad_norm": 1.628747820854187, - "learning_rate": 5.385843604077321e-05, - "loss": 0.2706174850463867, - "mean_token_accuracy": 0.9024625179743526, - "num_tokens": 5990107.0, - "step": 2500 - }, - { - "entropy": 0.285839084982872, - "epoch": 6.144752714113389, - "grad_norm": 1.2864570617675781, - "learning_rate": 5.114951151877489e-05, - "loss": 0.19398983001708983, - "mean_token_accuracy": 0.9306288361549377, - "num_tokens": 6104230.0, - "step": 2550 - }, - { - "entropy": 0.2841903820633888, - "epoch": 6.265379975874548, - "grad_norm": 1.3262356519699097, - "learning_rate": 4.846644111922105e-05, - "loss": 0.19344640731811524, - "mean_token_accuracy": 0.9303907889127732, - "num_tokens": 6223548.0, - "step": 2600 - }, - { - "entropy": 0.2794396734237671, - "epoch": 6.386007237635706, - "grad_norm": 1.216142177581787, - "learning_rate": 4.581396973700743e-05, - "loss": 0.19337059020996095, - "mean_token_accuracy": 0.9294343847036362, - "num_tokens": 6349102.0, - "step": 2650 - }, - { - "entropy": 0.28492832884192465, - "epoch": 6.5066344993968634, - "grad_norm": 1.281471610069275, - "learning_rate": 4.3196788153980425e-05, - "loss": 0.19621110916137696, - "mean_token_accuracy": 0.9284999430179596, - "num_tokens": 6468966.0, - "step": 2700 - }, - { - "entropy": 0.28580656126141546, - "epoch": 6.627261761158022, - "grad_norm": 1.581629753112793, - "learning_rate": 4.061952474349045e-05, - "loss": 0.2017466926574707, - "mean_token_accuracy": 0.9269114434719086, - "num_tokens": 6583296.0, - "step": 2750 - }, - { - "entropy": 0.2819211499392986, - "epoch": 6.74788902291918, - "grad_norm": 1.111789584159851, - "learning_rate": 3.8086737285312406e-05, - "loss": 0.19925508499145508, - "mean_token_accuracy": 0.9279456728696823, - "num_tokens": 6698696.0, - "step": 2800 - }, - { - "entropy": 0.27608554884791375, - "epoch": 6.868516284680338, - "grad_norm": 1.337012529373169, - "learning_rate": 3.560290490540746e-05, - "loss": 0.19468421936035157, - "mean_token_accuracy": 0.9297049742937088, - "num_tokens": 6819711.0, - "step": 2850 - }, - { - "entropy": 0.27499831572175026, - "epoch": 6.989143546441496, - "grad_norm": 1.2096495628356934, - "learning_rate": 3.317242015478115e-05, - "loss": 0.19465036392211915, - "mean_token_accuracy": 0.929957520365715, - "num_tokens": 6944941.0, - "step": 2900 - }, - { - "epoch": 7.0, - "eval_entropy": 0.43061937626158253, - "eval_mean_token_accuracy": 0.7864309693320414, - "eval_not_syn_loss": 0.8960050940513611, - "eval_not_syn_runtime": 53.715, - "eval_not_syn_samples_per_second": 26.436, - "eval_not_syn_steps_per_second": 3.314, - "eval_num_tokens": 6956642.0, - "step": 2905 - }, - { - "epoch": 7.0, - "eval_entropy": 0.40260162983047826, - "eval_mean_token_accuracy": 0.81038418077351, - "eval_num_tokens": 6956642.0, - "eval_syn_loss": 0.8312228322029114, - "eval_syn_runtime": 55.7344, - "eval_syn_samples_per_second": 25.478, - "eval_syn_steps_per_second": 3.194, - "step": 2905 - }, - { - "entropy": 0.22430229894440584, - "epoch": 7.108564535585042, - "grad_norm": 1.1575374603271484, - "learning_rate": 3.0799581241445255e-05, - "loss": 0.13364919662475586, - "mean_token_accuracy": 0.9534954422652119, - "num_tokens": 7064818.0, - "step": 2950 - }, - { - "entropy": 0.21469997629523277, - "epoch": 7.2291917973462, - "grad_norm": 1.0705780982971191, - "learning_rate": 2.8488584429221433e-05, - "loss": 0.13137072563171387, - "mean_token_accuracy": 0.9546506702899933, - "num_tokens": 7183624.0, - "step": 3000 - }, - { - "entropy": 0.21114618048071862, - "epoch": 7.349819059107358, - "grad_norm": 1.1908551454544067, - "learning_rate": 2.6243516616828438e-05, - "loss": 0.1303229808807373, - "mean_token_accuracy": 0.9540972465276718, - "num_tokens": 7304122.0, - "step": 3050 - }, - { - "entropy": 0.2161379286646843, - "epoch": 7.470446320868517, - "grad_norm": 1.198324203491211, - "learning_rate": 2.4068348110377418e-05, - "loss": 0.1322408390045166, - "mean_token_accuracy": 0.9531843411922455, - "num_tokens": 7422217.0, - "step": 3100 - }, - { - "entropy": 0.20395339593291284, - "epoch": 7.591073582629674, - "grad_norm": 1.3267316818237305, - "learning_rate": 2.1966925602055525e-05, - "loss": 0.12691154479980468, - "mean_token_accuracy": 0.954550067782402, - "num_tokens": 7547071.0, - "step": 3150 - }, - { - "entropy": 0.2153193709254265, - "epoch": 7.711700844390832, - "grad_norm": 1.2034279108047485, - "learning_rate": 1.9942965367416116e-05, - "loss": 0.13238741874694823, - "mean_token_accuracy": 0.9535881149768829, - "num_tokens": 7663184.0, - "step": 3200 - }, - { - "entropy": 0.2055867248773575, - "epoch": 7.832328106151991, - "grad_norm": 1.3387084007263184, - "learning_rate": 1.8000046693304666e-05, - "loss": 0.12809998512268067, - "mean_token_accuracy": 0.9540317648649216, - "num_tokens": 7785293.0, - "step": 3250 - }, - { - "entropy": 0.21260966420173644, - "epoch": 7.952955367913148, - "grad_norm": 1.201429009437561, - "learning_rate": 1.6141605548043607e-05, - "loss": 0.13191481590270995, - "mean_token_accuracy": 0.952368374466896, - "num_tokens": 7903616.0, - "step": 3300 - }, - { - "epoch": 8.0, - "eval_entropy": 0.3767001710915833, - "eval_mean_token_accuracy": 0.7909231467193432, - "eval_not_syn_loss": 1.0326374769210815, - "eval_not_syn_runtime": 54.1394, - "eval_not_syn_samples_per_second": 26.229, - "eval_not_syn_steps_per_second": 3.288, - "eval_num_tokens": 7950448.0, - "step": 3320 - }, - { - "epoch": 8.0, - "eval_entropy": 0.35253987854786134, - "eval_mean_token_accuracy": 0.803997924488582, - "eval_num_tokens": 7950448.0, - "eval_syn_loss": 0.9743913412094116, - "eval_syn_runtime": 55.701, - "eval_syn_samples_per_second": 25.493, - "eval_syn_steps_per_second": 3.196, - "step": 3320 - }, - { - "entropy": 0.1920483737732425, - "epoch": 8.072376357056696, - "grad_norm": 1.1402173042297363, - "learning_rate": 1.4370928505069763e-05, - "loss": 0.10957123756408692, - "mean_token_accuracy": 0.9624346013020988, - "num_tokens": 8020867.0, - "step": 3350 - }, - { - "entropy": 0.17265828780829906, - "epoch": 8.193003618817853, - "grad_norm": 1.285122036933899, - "learning_rate": 1.2691146930770208e-05, - "loss": 0.09433144569396973, - "mean_token_accuracy": 0.9681703245639801, - "num_tokens": 8143032.0, - "step": 3400 - }, - { - "entropy": 0.1748460330814123, - "epoch": 8.31363088057901, - "grad_norm": 0.9159325957298279, - "learning_rate": 1.110523144679515e-05, - "loss": 0.09543456077575684, - "mean_token_accuracy": 0.9679662960767746, - "num_tokens": 8262875.0, - "step": 3450 - }, - { - "entropy": 0.16876790300011635, - "epoch": 8.43425814234017, - "grad_norm": 1.0886222124099731, - "learning_rate": 9.615986676640981e-06, - "loss": 0.09403505325317382, - "mean_token_accuracy": 0.9679694646596908, - "num_tokens": 8383147.0, - "step": 3500 - }, - { - "entropy": 0.1674938587099314, - "epoch": 8.554885404101327, - "grad_norm": 0.8731210231781006, - "learning_rate": 8.226046285793963e-06, - "loss": 0.09352604866027832, - "mean_token_accuracy": 0.9685411840677262, - "num_tokens": 8502743.0, - "step": 3550 - }, - { - "entropy": 0.1702917966991663, - "epoch": 8.675512665862485, - "grad_norm": 1.2127553224563599, - "learning_rate": 6.937868324205779e-06, - "loss": 0.09577181816101074, - "mean_token_accuracy": 0.9679328501224518, - "num_tokens": 8621550.0, - "step": 3600 - }, - { - "entropy": 0.16657552666962147, - "epoch": 8.796139927623642, - "grad_norm": 0.7716957330703735, - "learning_rate": 5.7537308793376514e-06, - "loss": 0.09437231063842773, - "mean_token_accuracy": 0.9675484406948089, - "num_tokens": 8742590.0, - "step": 3650 - }, - { - "entropy": 0.16926519222557546, - "epoch": 8.916767189384801, - "grad_norm": 0.9887721538543701, - "learning_rate": 4.675728047460445e-06, - "loss": 0.09573341369628906, - "mean_token_accuracy": 0.9675081825256348, - "num_tokens": 8862250.0, - "step": 3700 - }, - { - "epoch": 9.0, - "eval_entropy": 0.348999947309494, - "eval_mean_token_accuracy": 0.7875944831398096, - "eval_not_syn_loss": 1.15312659740448, - "eval_not_syn_runtime": 54.232, - "eval_not_syn_samples_per_second": 26.184, - "eval_not_syn_steps_per_second": 3.282, - "eval_num_tokens": 8944254.0, - "step": 3735 - }, - { - "epoch": 9.0, - "eval_entropy": 0.32651834206634694, - "eval_mean_token_accuracy": 0.8028985322191474, - "eval_num_tokens": 8944254.0, - "eval_syn_loss": 1.0929256677627563, - "eval_syn_runtime": 56.5213, - "eval_syn_samples_per_second": 25.123, - "eval_syn_steps_per_second": 3.149, - "step": 3735 - }, - { - "entropy": 0.16505921361121265, - "epoch": 9.036188178528347, - "grad_norm": 0.5117172002792358, - "learning_rate": 3.7057662303351195e-06, - "loss": 0.0903382968902588, - "mean_token_accuracy": 0.9698322872922878, - "num_tokens": 8981377.0, - "step": 3750 - }, - { - "entropy": 0.16023802682757377, - "epoch": 9.156815440289506, - "grad_norm": 0.7043508887290955, - "learning_rate": 2.8455607638231843e-06, - "loss": 0.08237065315246582, - "mean_token_accuracy": 0.973184723854065, - "num_tokens": 9099674.0, - "step": 3800 - }, - { - "entropy": 0.15360147535800933, - "epoch": 9.277442702050664, - "grad_norm": 0.8259885311126709, - "learning_rate": 2.096632884388653e-06, - "loss": 0.0790113353729248, - "mean_token_accuracy": 0.9740552854537964, - "num_tokens": 9222172.0, - "step": 3850 - }, - { - "entropy": 0.1572684432566166, - "epoch": 9.398069963811821, - "grad_norm": 0.7598025798797607, - "learning_rate": 1.4603070388567325e-06, - "loss": 0.08192154884338379, - "mean_token_accuracy": 0.9732418739795685, - "num_tokens": 9341003.0, - "step": 3900 - }, - { - "entropy": 0.1555389292538166, - "epoch": 9.51869722557298, - "grad_norm": 1.0184978246688843, - "learning_rate": 9.377085421863613e-07, - "loss": 0.08013585090637207, - "mean_token_accuracy": 0.9729253125190734, - "num_tokens": 9461190.0, - "step": 3950 - }, - { - "entropy": 0.15520566545426845, - "epoch": 9.639324487334138, - "grad_norm": 0.825910210609436, - "learning_rate": 5.297615873990316e-07, - "loss": 0.0808549690246582, - "mean_token_accuracy": 0.9731829220056534, - "num_tokens": 9577775.0, - "step": 4000 - }, - { - "entropy": 0.15546547561883928, - "epoch": 9.759951749095295, - "grad_norm": 0.9650986790657043, - "learning_rate": 2.3718761118302318e-07, - "loss": 0.08213469505310059, - "mean_token_accuracy": 0.9729492688179016, - "num_tokens": 9696918.0, - "step": 4050 - }, - { - "entropy": 0.15284155048429965, - "epoch": 9.880579010856454, - "grad_norm": 0.5635292530059814, - "learning_rate": 6.05040180636361e-08, - "loss": 0.07865472316741944, - "mean_token_accuracy": 0.9736917620897293, - "num_tokens": 9821164.0, - "step": 4100 - }, - { - "entropy": 0.15636218971375263, - "epoch": 10.0, - "grad_norm": 0.9871779084205627, - "learning_rate": 2.326539542532457e-11, - "loss": 0.08128765106201172, - "mean_token_accuracy": 0.9726401260404876, - "num_tokens": 9938060.0, - "step": 4150 - }, - { - "epoch": 10.0, - "eval_entropy": 0.3390489054864712, - "eval_mean_token_accuracy": 0.7870067391502723, - "eval_not_syn_loss": 1.2185481786727905, - "eval_not_syn_runtime": 53.9521, - "eval_not_syn_samples_per_second": 26.32, - "eval_not_syn_steps_per_second": 3.299, - "eval_num_tokens": 9938060.0, - "step": 4150 - }, - { - "epoch": 10.0, - "eval_entropy": 0.31715648001834246, - "eval_mean_token_accuracy": 0.802757600050294, - "eval_num_tokens": 9938060.0, - "eval_syn_loss": 1.150896668434143, - "eval_syn_runtime": 55.6165, - "eval_syn_samples_per_second": 25.532, - "eval_syn_steps_per_second": 3.2, - "step": 4150 - } - ], - "logging_steps": 50, - "max_steps": 4150, - "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": true - }, - "attributes": {} - } - }, - "total_flos": 2.0486433180931622e+17, - "train_batch_size": 8, - "trial_name": null, - "trial_params": null -} diff --git a/substitutivity_original_Estonian/Qwen3.5-2B-Base_substitutivity_splits_original_features_train_substitutivity_splits_original_features_test2/checkpoint-4150/training_args.bin b/substitutivity_original_Estonian/Qwen3.5-2B-Base_substitutivity_splits_original_features_train_substitutivity_splits_original_features_test2/checkpoint-4150/training_args.bin deleted file mode 100644 index f3d8098f65400413dd700f32ada5649470871017..0000000000000000000000000000000000000000 --- a/substitutivity_original_Estonian/Qwen3.5-2B-Base_substitutivity_splits_original_features_train_substitutivity_splits_original_features_test2/checkpoint-4150/training_args.bin +++ /dev/null @@ -1,3 +0,0 @@ -version https://git-lfs.github.com/spec/v1 -oid sha256:617af0f41414927baaaf158949585bb75eb4b05ff78baa5216c860ef0b431009 -size 6033 diff --git a/substitutivity_original_Estonian/Qwen3.5-2B-Base_substitutivity_splits_original_features_train_substitutivity_splits_original_features_test2/checkpoint-830/README.md b/substitutivity_original_Estonian/Qwen3.5-2B-Base_substitutivity_splits_original_features_train_substitutivity_splits_original_features_test2/checkpoint-830/README.md deleted file mode 100644 index ece204d005f7bf161d0e448efdd56fd297b7c8a2..0000000000000000000000000000000000000000 --- a/substitutivity_original_Estonian/Qwen3.5-2B-Base_substitutivity_splits_original_features_train_substitutivity_splits_original_features_test2/checkpoint-830/README.md +++ /dev/null @@ -1,209 +0,0 @@ ---- -base_model: Qwen/Qwen3.5-2B-Base -library_name: peft -pipeline_tag: text-generation -tags: -- base_model:adapter:Qwen/Qwen3.5-2B-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/substitutivity_original_Estonian/Qwen3.5-2B-Base_substitutivity_splits_original_features_train_substitutivity_splits_original_features_test2/checkpoint-830/adapter_config.json b/substitutivity_original_Estonian/Qwen3.5-2B-Base_substitutivity_splits_original_features_train_substitutivity_splits_original_features_test2/checkpoint-830/adapter_config.json deleted file mode 100644 index 6a095bb3bfe249ce8e15df6ae999e88fa1e17162..0000000000000000000000000000000000000000 --- a/substitutivity_original_Estonian/Qwen3.5-2B-Base_substitutivity_splits_original_features_train_substitutivity_splits_original_features_test2/checkpoint-830/adapter_config.json +++ /dev/null @@ -1,46 +0,0 @@ -{ - "alora_invocation_tokens": null, - "alpha_pattern": {}, - "arrow_config": null, - "auto_mapping": null, - "base_model_name_or_path": "Qwen/Qwen3.5-2B-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": 0.05731002492184622, - "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": [ - "o_proj", - "q_proj", - "k_proj", - "down_proj", - "up_proj", - "v_proj", - "gate_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/substitutivity_original_Estonian/Qwen3.5-2B-Base_substitutivity_splits_original_features_train_substitutivity_splits_original_features_test2/checkpoint-830/adapter_model.safetensors b/substitutivity_original_Estonian/Qwen3.5-2B-Base_substitutivity_splits_original_features_train_substitutivity_splits_original_features_test2/checkpoint-830/adapter_model.safetensors deleted file mode 100644 index a6b826405e9579efe197f78234c36f4dc456d1fc..0000000000000000000000000000000000000000 --- a/substitutivity_original_Estonian/Qwen3.5-2B-Base_substitutivity_splits_original_features_train_substitutivity_splits_original_features_test2/checkpoint-830/adapter_model.safetensors +++ /dev/null @@ -1,3 +0,0 @@ -version https://git-lfs.github.com/spec/v1 -oid sha256:7d04573369962e5a3cdd3d6c4aaa9fe20f688e0d2f1d148a031f386f6303d55b -size 174613392 diff --git a/substitutivity_original_Estonian/Qwen3.5-2B-Base_substitutivity_splits_original_features_train_substitutivity_splits_original_features_test2/checkpoint-830/chat_template.jinja b/substitutivity_original_Estonian/Qwen3.5-2B-Base_substitutivity_splits_original_features_train_substitutivity_splits_original_features_test2/checkpoint-830/chat_template.jinja deleted file mode 100644 index 0ef09f214eaa6d9bca297988afc1454b5827b2c7..0000000000000000000000000000000000000000 --- a/substitutivity_original_Estonian/Qwen3.5-2B-Base_substitutivity_splits_original_features_train_substitutivity_splits_original_features_test2/checkpoint-830/chat_template.jinja +++ /dev/null @@ -1,154 +0,0 @@ -{%- set image_count = namespace(value=0) %} -{%- set video_count = namespace(value=0) %} -{%- macro render_content(content, do_vision_count, is_system_content=false) %} - {%- if content is string %} - {{- content }} - {%- elif content is iterable and content is not mapping %} - {%- for item in content %} - {%- if 'image' in item or 'image_url' in item or item.type == 'image' %} - {%- if is_system_content %} - {{- raise_exception('System message cannot contain images.') }} - {%- endif %} - {%- if do_vision_count %} - {%- set image_count.value = image_count.value + 1 %} - {%- endif %} - {%- if add_vision_id %} - {{- 'Picture ' ~ image_count.value ~ ': ' }} - {%- endif %} - {{- '<|vision_start|><|image_pad|><|vision_end|>' }} - {%- elif 'video' in item or item.type == 'video' %} - {%- if is_system_content %} - {{- raise_exception('System message cannot contain videos.') }} - {%- endif %} - {%- if do_vision_count %} - {%- set video_count.value = video_count.value + 1 %} - {%- endif %} - {%- if add_vision_id %} - {{- 'Video ' ~ video_count.value ~ ': ' }} - {%- endif %} - {{- '<|vision_start|><|video_pad|><|vision_end|>' }} - {%- elif 'text' in item %} - {{- item.text }} - {%- else %} - {{- raise_exception('Unexpected item type in content.') }} - {%- endif %} - {%- endfor %} - {%- elif content is none or content is undefined %} - {{- '' }} - {%- else %} - {{- raise_exception('Unexpected content type.') }} - {%- endif %} -{%- endmacro %} -{%- if not messages %} - {{- raise_exception('No messages provided.') }} -{%- endif %} -{%- if tools and tools is iterable and tools is not mapping %} - {{- '<|im_start|>system\n' }} - {{- "# Tools\n\nYou have access to the following functions:\n\n" }} - {%- for tool in tools %} - {{- "\n" }} - {{- tool | tojson }} - {%- endfor %} - {{- "\n" }} - {{- '\n\nIf you choose to call a function ONLY reply in the following format with NO suffix:\n\n\n\n\nvalue_1\n\n\nThis is the value for the second parameter\nthat can span\nmultiple lines\n\n\n\n\n\nReminder:\n- Function calls MUST follow the specified format: an inner block must be nested within XML tags\n- Required parameters MUST be specified\n- You may provide optional reasoning for your function call in natural language BEFORE the function call, but NOT after\n- If there is no function call available, answer the question like normal with your current knowledge and do not tell the user about function calls\n' }} - {%- if messages[0].role == 'system' %} - {%- set content = render_content(messages[0].content, false, true)|trim %} - {%- if content %} - {{- '\n\n' + content }} - {%- endif %} - {%- endif %} - {{- '<|im_end|>\n' }} -{%- else %} - {%- if messages[0].role == 'system' %} - {%- set content = render_content(messages[0].content, false, true)|trim %} - {{- '<|im_start|>system\n' + 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" %} - {%- set content = render_content(message.content, false)|trim %} - {%- if not(content.startswith('') and content.endswith('')) %} - {%- set ns.multi_step_tool = false %} - {%- set ns.last_query_index = index %} - {%- endif %} - {%- endif %} -{%- endfor %} -{%- if ns.multi_step_tool %} - {{- raise_exception('No user query found in messages.') }} -{%- endif %} -{%- for message in messages %} - {%- set content = render_content(message.content, true)|trim %} - {%- if message.role == "system" %} - {%- if not loop.first %} - {{- raise_exception('System message must be at the beginning.') }} - {%- endif %} - {%- elif message.role == "user" %} - {{- '<|im_start|>' + message.role + '\n' + content + '<|im_end|>' + '\n' }} - {%- elif message.role == "assistant" %} - {%- set reasoning_content = '' %} - {%- if message.reasoning_content is string %} - {%- set reasoning_content = message.reasoning_content %} - {%- else %} - {%- if '' in content %} - {%- set reasoning_content = content.split('')[0].rstrip('\n').split('')[-1].lstrip('\n') %} - {%- set content = content.split('')[-1].lstrip('\n') %} - {%- endif %} - {%- endif %} - {%- set reasoning_content = reasoning_content|trim %} - {%- if loop.index0 > ns.last_query_index %} - {{- '<|im_start|>' + message.role + '\n\n' + reasoning_content + '\n\n\n' + content }} - {%- else %} - {{- '<|im_start|>' + message.role + '\n' + content }} - {%- endif %} - {%- if message.tool_calls and message.tool_calls is iterable and message.tool_calls is not mapping %} - {%- for tool_call in message.tool_calls %} - {%- if tool_call.function is defined %} - {%- set tool_call = tool_call.function %} - {%- endif %} - {%- if loop.first %} - {%- if content|trim %} - {{- '\n\n\n\n' }} - {%- else %} - {{- '\n\n' }} - {%- endif %} - {%- else %} - {{- '\n\n\n' }} - {%- endif %} - {%- if tool_call.arguments is defined %} - {%- for args_name, args_value in tool_call.arguments|items %} - {{- '\n' }} - {%- set args_value = args_value | tojson | safe if args_value is mapping or (args_value is sequence and args_value is not string) else args_value | string %} - {{- args_value }} - {{- '\n\n' }} - {%- endfor %} - {%- endif %} - {{- '\n' }} - {%- endfor %} - {%- endif %} - {{- '<|im_end|>\n' }} - {%- elif message.role == "tool" %} - {%- if loop.previtem and loop.previtem.role != "tool" %} - {{- '<|im_start|>user' }} - {%- endif %} - {{- '\n\n' }} - {{- content }} - {{- '\n' }} - {%- if not loop.last and loop.nextitem.role != "tool" %} - {{- '<|im_end|>\n' }} - {%- elif loop.last %} - {{- '<|im_end|>\n' }} - {%- endif %} - {%- else %} - {{- raise_exception('Unexpected message role.') }} - {%- endif %} -{%- endfor %} -{%- if add_generation_prompt %} - {{- '<|im_start|>assistant\n' }} - {%- if enable_thinking is defined and enable_thinking is true %} - {{- '\n' }} - {%- else %} - {{- '\n\n\n\n' }} - {%- endif %} -{%- endif %} \ No newline at end of file diff --git a/substitutivity_original_Estonian/Qwen3.5-2B-Base_substitutivity_splits_original_features_train_substitutivity_splits_original_features_test2/checkpoint-830/tokenizer.json b/substitutivity_original_Estonian/Qwen3.5-2B-Base_substitutivity_splits_original_features_train_substitutivity_splits_original_features_test2/checkpoint-830/tokenizer.json deleted file mode 100644 index 67741b04f23bfdb46501f748ce27865ec82eccfb..0000000000000000000000000000000000000000 --- a/substitutivity_original_Estonian/Qwen3.5-2B-Base_substitutivity_splits_original_features_train_substitutivity_splits_original_features_test2/checkpoint-830/tokenizer.json +++ /dev/null @@ -1,3 +0,0 @@ -version https://git-lfs.github.com/spec/v1 -oid sha256:87a7830d63fcf43bf241c3c5242e96e62dd3fdc29224ca26fed8ea333db72de4 -size 19989343 diff --git a/substitutivity_original_Estonian/Qwen3.5-2B-Base_substitutivity_splits_original_features_train_substitutivity_splits_original_features_test2/checkpoint-830/tokenizer_config.json b/substitutivity_original_Estonian/Qwen3.5-2B-Base_substitutivity_splits_original_features_train_substitutivity_splits_original_features_test2/checkpoint-830/tokenizer_config.json deleted file mode 100644 index b4a37b2a6fd3ab3317cd7bac72855be1a843b2bb..0000000000000000000000000000000000000000 --- a/substitutivity_original_Estonian/Qwen3.5-2B-Base_substitutivity_splits_original_features_train_substitutivity_splits_original_features_test2/checkpoint-830/tokenizer_config.json +++ /dev/null @@ -1,31 +0,0 @@ -{ - "add_prefix_space": false, - "audio_bos_token": "<|audio_start|>", - "audio_eos_token": "<|audio_end|>", - "audio_token": "<|audio_pad|>", - "backend": "tokenizers", - "bos_token": null, - "clean_up_tokenization_spaces": false, - "eos_token": "<|endoftext|>", - "errors": "replace", - "image_token": "<|image_pad|>", - "is_local": false, - "model_max_length": 262144, - "model_specific_special_tokens": { - "audio_bos_token": "<|audio_start|>", - "audio_eos_token": "<|audio_end|>", - "audio_token": "<|audio_pad|>", - "image_token": "<|image_pad|>", - "video_token": "<|video_pad|>", - "vision_bos_token": "<|vision_start|>", - "vision_eos_token": "<|vision_end|>" - }, - "pad_token": "<|endoftext|>", - "pretokenize_regex": "(?i:'s|'t|'re|'ve|'m|'ll|'d)|[^\\r\\n\\p{L}\\p{N}]?[\\p{L}\\p{M}]+|\\p{N}| ?[^\\s\\p{L}\\p{M}\\p{N}]+[\\r\\n]*|\\s*[\\r\\n]+|\\s+(?!\\S)|\\s+", - "split_special_tokens": false, - "tokenizer_class": "TokenizersBackend", - "unk_token": null, - "video_token": "<|video_pad|>", - "vision_bos_token": "<|vision_start|>", - "vision_eos_token": "<|vision_end|>" -} diff --git a/substitutivity_original_Estonian/Qwen3.5-2B-Base_substitutivity_splits_original_features_train_substitutivity_splits_original_features_test2/checkpoint-830/trainer_state.json b/substitutivity_original_Estonian/Qwen3.5-2B-Base_substitutivity_splits_original_features_train_substitutivity_splits_original_features_test2/checkpoint-830/trainer_state.json deleted file mode 100644 index 9a1ca37386db34f48df761cf5b998d30ef3d5f2c..0000000000000000000000000000000000000000 --- a/substitutivity_original_Estonian/Qwen3.5-2B-Base_substitutivity_splits_original_features_train_substitutivity_splits_original_features_test2/checkpoint-830/trainer_state.json +++ /dev/null @@ -1,238 +0,0 @@ -{ - "best_global_step": null, - "best_metric": null, - "best_model_checkpoint": null, - "epoch": 2.0, - "eval_steps": 500, - "global_step": 830, - "is_hyper_param_search": false, - "is_local_process_zero": true, - "is_world_process_zero": true, - "log_history": [ - { - "entropy": 2.2933380508422854, - "epoch": 0.12062726176115803, - "grad_norm": 2.8075337409973145, - "learning_rate": 1.5531017374629602e-05, - "loss": 2.1429234313964844, - "mean_token_accuracy": 0.5779684272408485, - "num_tokens": 121428.0, - "step": 50 - }, - { - "entropy": 1.1315062403678895, - "epoch": 0.24125452352231605, - "grad_norm": 1.9415371417999268, - "learning_rate": 3.1378994287516954e-05, - "loss": 1.0461698913574218, - "mean_token_accuracy": 0.7166991305351257, - "num_tokens": 243966.0, - "step": 100 - }, - { - "entropy": 0.9537658470869065, - "epoch": 0.3618817852834741, - "grad_norm": 1.8353157043457031, - "learning_rate": 4.72269712004043e-05, - "loss": 0.8774301147460938, - "mean_token_accuracy": 0.7506789201498032, - "num_tokens": 366649.0, - "step": 150 - }, - { - "entropy": 0.8955666983127594, - "epoch": 0.4825090470446321, - "grad_norm": 1.5419880151748657, - "learning_rate": 6.307494811329165e-05, - "loss": 0.8234302520751953, - "mean_token_accuracy": 0.7616239994764328, - "num_tokens": 488367.0, - "step": 200 - }, - { - "entropy": 0.8689469587802887, - "epoch": 0.6031363088057901, - "grad_norm": 1.5661484003067017, - "learning_rate": 7.8922925026179e-05, - "loss": 0.7961890411376953, - "mean_token_accuracy": 0.767715499997139, - "num_tokens": 603131.0, - "step": 250 - }, - { - "entropy": 0.8425652319192887, - "epoch": 0.7237635705669482, - "grad_norm": 1.5694905519485474, - "learning_rate": 9.477090193906636e-05, - "loss": 0.7718594360351563, - "mean_token_accuracy": 0.7736722755432129, - "num_tokens": 726412.0, - "step": 300 - }, - { - "entropy": 0.843294187784195, - "epoch": 0.8443908323281062, - "grad_norm": 1.1118028163909912, - "learning_rate": 0.00011061887885195371, - "loss": 0.7801297760009765, - "mean_token_accuracy": 0.7714760398864746, - "num_tokens": 846516.0, - "step": 350 - }, - { - "entropy": 0.8213214933872223, - "epoch": 0.9650180940892642, - "grad_norm": 0.9202918410301208, - "learning_rate": 0.00012646685576484106, - "loss": 0.7514453887939453, - "mean_token_accuracy": 0.7774044269323349, - "num_tokens": 958978.0, - "step": 400 - }, - { - "epoch": 1.0, - "eval_entropy": 0.818668900915746, - "eval_mean_token_accuracy": 0.7533496866065464, - "eval_not_syn_loss": 0.7798981070518494, - "eval_not_syn_runtime": 53.8744, - "eval_not_syn_samples_per_second": 26.358, - "eval_not_syn_steps_per_second": 3.304, - "eval_num_tokens": 993806.0, - "step": 415 - }, - { - "epoch": 1.0, - "eval_entropy": 0.7627468125873738, - "eval_mean_token_accuracy": 0.8024111574285486, - "eval_num_tokens": 993806.0, - "eval_syn_loss": 0.7218216061592102, - "eval_syn_runtime": 55.3239, - "eval_syn_samples_per_second": 25.667, - "eval_syn_steps_per_second": 3.217, - "step": 415 - }, - { - "entropy": 0.7802230679627621, - "epoch": 1.0844390832328106, - "grad_norm": 1.0145268440246582, - "learning_rate": 0.00013151131541121804, - "loss": 0.7111608123779297, - "mean_token_accuracy": 0.7845737542768921, - "num_tokens": 1080301.0, - "step": 450 - }, - { - "entropy": 0.7558828103542328, - "epoch": 1.2050663449939687, - "grad_norm": 1.0477591753005981, - "learning_rate": 0.0001313741160170038, - "loss": 0.6933037567138672, - "mean_token_accuracy": 0.7889221370220184, - "num_tokens": 1201215.0, - "step": 500 - }, - { - "entropy": 0.7593123841285706, - "epoch": 1.3256936067551266, - "grad_norm": 0.899161159992218, - "learning_rate": 0.00013112089697387701, - "loss": 0.6962493133544921, - "mean_token_accuracy": 0.7877359414100646, - "num_tokens": 1322895.0, - "step": 550 - }, - { - "entropy": 0.7460986822843552, - "epoch": 1.4463208685162847, - "grad_norm": 1.1061116456985474, - "learning_rate": 0.00013075210608885215, - "loss": 0.6817412567138672, - "mean_token_accuracy": 0.7904874366521836, - "num_tokens": 1436749.0, - "step": 600 - }, - { - "entropy": 0.735516265630722, - "epoch": 1.5669481302774426, - "grad_norm": 0.9925405979156494, - "learning_rate": 0.00013026839555278796, - "loss": 0.6690705871582031, - "mean_token_accuracy": 0.7930250608921051, - "num_tokens": 1555400.0, - "step": 650 - }, - { - "entropy": 0.7302519899606704, - "epoch": 1.6875753920386007, - "grad_norm": 0.800025463104248, - "learning_rate": 0.0001296706207870159, - "loss": 0.6714833831787109, - "mean_token_accuracy": 0.7946307575702667, - "num_tokens": 1676892.0, - "step": 700 - }, - { - "entropy": 0.7321263110637665, - "epoch": 1.8082026537997589, - "grad_norm": 1.0500682592391968, - "learning_rate": 0.00012895983893056423, - "loss": 0.6672868347167968, - "mean_token_accuracy": 0.7947645837068558, - "num_tokens": 1802689.0, - "step": 750 - }, - { - "entropy": 0.7075755137205124, - "epoch": 1.9288299155609168, - "grad_norm": 0.9248411655426025, - "learning_rate": 0.00012813730697065298, - "loss": 0.6466568756103516, - "mean_token_accuracy": 0.798747689127922, - "num_tokens": 1920804.0, - "step": 800 - }, - { - "epoch": 2.0, - "eval_entropy": 0.7426781811740961, - "eval_mean_token_accuracy": 0.7693507239389955, - "eval_not_syn_loss": 0.7139139175415039, - "eval_not_syn_runtime": 54.4286, - "eval_not_syn_samples_per_second": 26.089, - "eval_not_syn_steps_per_second": 3.27, - "eval_num_tokens": 1987612.0, - "step": 830 - }, - { - "epoch": 2.0, - "eval_entropy": 0.6937930352902144, - "eval_mean_token_accuracy": 0.8180048633157537, - "eval_num_tokens": 1987612.0, - "eval_syn_loss": 0.6489660143852234, - "eval_syn_runtime": 56.1525, - "eval_syn_samples_per_second": 25.288, - "eval_syn_steps_per_second": 3.17, - "step": 830 - } - ], - "logging_steps": 50, - "max_steps": 4150, - "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": 4.103380324314163e+16, - "train_batch_size": 8, - "trial_name": null, - "trial_params": null -} diff --git a/substitutivity_original_Estonian/Qwen3.5-2B-Base_substitutivity_splits_original_features_train_substitutivity_splits_original_features_test2/checkpoint-830/training_args.bin b/substitutivity_original_Estonian/Qwen3.5-2B-Base_substitutivity_splits_original_features_train_substitutivity_splits_original_features_test2/checkpoint-830/training_args.bin deleted file mode 100644 index f3d8098f65400413dd700f32ada5649470871017..0000000000000000000000000000000000000000 --- a/substitutivity_original_Estonian/Qwen3.5-2B-Base_substitutivity_splits_original_features_train_substitutivity_splits_original_features_test2/checkpoint-830/training_args.bin +++ /dev/null @@ -1,3 +0,0 @@ -version https://git-lfs.github.com/spec/v1 -oid sha256:617af0f41414927baaaf158949585bb75eb4b05ff78baa5216c860ef0b431009 -size 6033