Commit ·
f9a78ad
1
Parent(s): 9a89741
Upload folder using huggingface_hub
Browse files- README.md +21 -0
- adapter_config.json +21 -0
- adapter_model.bin +3 -0
- special_tokens_map.json +24 -0
- tokenizer.json +0 -0
- tokenizer.model +3 -0
- tokenizer_config.json +34 -0
- train_args.json +123 -0
- training_args.bin +3 -0
- training_logs.json +1067 -0
README.md
ADDED
|
@@ -0,0 +1,21 @@
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
| 1 |
+
---
|
| 2 |
+
library_name: peft
|
| 3 |
+
---
|
| 4 |
+
## Training procedure
|
| 5 |
+
|
| 6 |
+
|
| 7 |
+
The following `bitsandbytes` quantization config was used during training:
|
| 8 |
+
- quant_method: bitsandbytes
|
| 9 |
+
- load_in_8bit: False
|
| 10 |
+
- load_in_4bit: True
|
| 11 |
+
- llm_int8_threshold: 6.0
|
| 12 |
+
- llm_int8_skip_modules: None
|
| 13 |
+
- llm_int8_enable_fp32_cpu_offload: False
|
| 14 |
+
- llm_int8_has_fp16_weight: False
|
| 15 |
+
- bnb_4bit_quant_type: nf4
|
| 16 |
+
- bnb_4bit_use_double_quant: True
|
| 17 |
+
- bnb_4bit_compute_dtype: float16
|
| 18 |
+
### Framework versions
|
| 19 |
+
|
| 20 |
+
|
| 21 |
+
- PEFT 0.5.0
|
adapter_config.json
ADDED
|
@@ -0,0 +1,21 @@
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
| 1 |
+
{
|
| 2 |
+
"auto_mapping": null,
|
| 3 |
+
"base_model_name_or_path": "models/openllama-3b",
|
| 4 |
+
"bias": "none",
|
| 5 |
+
"fan_in_fan_out": false,
|
| 6 |
+
"inference_mode": true,
|
| 7 |
+
"init_lora_weights": true,
|
| 8 |
+
"layers_pattern": null,
|
| 9 |
+
"layers_to_transform": null,
|
| 10 |
+
"lora_alpha": 16,
|
| 11 |
+
"lora_dropout": 0.1,
|
| 12 |
+
"modules_to_save": null,
|
| 13 |
+
"peft_type": "LORA",
|
| 14 |
+
"r": 64,
|
| 15 |
+
"revision": null,
|
| 16 |
+
"target_modules": [
|
| 17 |
+
"q_proj",
|
| 18 |
+
"v_proj"
|
| 19 |
+
],
|
| 20 |
+
"task_type": "SEQ_CLS"
|
| 21 |
+
}
|
adapter_model.bin
ADDED
|
@@ -0,0 +1,3 @@
|
|
|
|
|
|
|
|
|
|
|
|
|
| 1 |
+
version https://git-lfs.github.com/spec/v1
|
| 2 |
+
oid sha256:747d40123b1dc6fe3bf72f125ec935bc3a758c82c3bbf0b40e965d2eaead25b6
|
| 3 |
+
size 85258545
|
special_tokens_map.json
ADDED
|
@@ -0,0 +1,24 @@
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
| 1 |
+
{
|
| 2 |
+
"bos_token": {
|
| 3 |
+
"content": "<s>",
|
| 4 |
+
"lstrip": false,
|
| 5 |
+
"normalized": true,
|
| 6 |
+
"rstrip": false,
|
| 7 |
+
"single_word": false
|
| 8 |
+
},
|
| 9 |
+
"eos_token": {
|
| 10 |
+
"content": "</s>",
|
| 11 |
+
"lstrip": false,
|
| 12 |
+
"normalized": true,
|
| 13 |
+
"rstrip": false,
|
| 14 |
+
"single_word": false
|
| 15 |
+
},
|
| 16 |
+
"pad_token": "</s>",
|
| 17 |
+
"unk_token": {
|
| 18 |
+
"content": "<unk>",
|
| 19 |
+
"lstrip": false,
|
| 20 |
+
"normalized": true,
|
| 21 |
+
"rstrip": false,
|
| 22 |
+
"single_word": false
|
| 23 |
+
}
|
| 24 |
+
}
|
tokenizer.json
ADDED
|
The diff for this file is too large to render.
See raw diff
|
|
|
tokenizer.model
ADDED
|
@@ -0,0 +1,3 @@
|
|
|
|
|
|
|
|
|
|
|
|
|
| 1 |
+
version https://git-lfs.github.com/spec/v1
|
| 2 |
+
oid sha256:ab1b681ec7fc02fed5edd3026687d7a692a918c4dd8e150ca2e3994a6229843b
|
| 3 |
+
size 534194
|
tokenizer_config.json
ADDED
|
@@ -0,0 +1,34 @@
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
| 1 |
+
{
|
| 2 |
+
"bos_token": {
|
| 3 |
+
"__type": "AddedToken",
|
| 4 |
+
"content": "<s>",
|
| 5 |
+
"lstrip": false,
|
| 6 |
+
"normalized": true,
|
| 7 |
+
"rstrip": false,
|
| 8 |
+
"single_word": false
|
| 9 |
+
},
|
| 10 |
+
"clean_up_tokenization_spaces": false,
|
| 11 |
+
"eos_token": {
|
| 12 |
+
"__type": "AddedToken",
|
| 13 |
+
"content": "</s>",
|
| 14 |
+
"lstrip": false,
|
| 15 |
+
"normalized": true,
|
| 16 |
+
"rstrip": false,
|
| 17 |
+
"single_word": false
|
| 18 |
+
},
|
| 19 |
+
"legacy": null,
|
| 20 |
+
"model_max_length": 2048,
|
| 21 |
+
"pad_token": null,
|
| 22 |
+
"sp_model_kwargs": {},
|
| 23 |
+
"spaces_between_special_tokens": false,
|
| 24 |
+
"tokenizer_class": "LlamaTokenizer",
|
| 25 |
+
"unk_token": {
|
| 26 |
+
"__type": "AddedToken",
|
| 27 |
+
"content": "<unk>",
|
| 28 |
+
"lstrip": false,
|
| 29 |
+
"normalized": true,
|
| 30 |
+
"rstrip": false,
|
| 31 |
+
"single_word": false
|
| 32 |
+
},
|
| 33 |
+
"use_default_system_prompt": true
|
| 34 |
+
}
|
train_args.json
ADDED
|
@@ -0,0 +1,123 @@
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
| 1 |
+
{
|
| 2 |
+
"initial_model_dir": "models/openllama-3b",
|
| 3 |
+
"distribution_id": "code_hard",
|
| 4 |
+
"date_trained": "15/10/2023 07:42:08",
|
| 5 |
+
"output_dir": "models/classify_lora/openllama-3b-code_hard",
|
| 6 |
+
"overwrite_output_dir": false,
|
| 7 |
+
"do_train": false,
|
| 8 |
+
"do_eval": true,
|
| 9 |
+
"do_predict": false,
|
| 10 |
+
"evaluation_strategy": "steps",
|
| 11 |
+
"prediction_loss_only": false,
|
| 12 |
+
"per_device_train_batch_size": 16,
|
| 13 |
+
"per_device_eval_batch_size": 32,
|
| 14 |
+
"per_gpu_train_batch_size": null,
|
| 15 |
+
"per_gpu_eval_batch_size": null,
|
| 16 |
+
"gradient_accumulation_steps": 1,
|
| 17 |
+
"eval_accumulation_steps": null,
|
| 18 |
+
"eval_delay": 0,
|
| 19 |
+
"learning_rate": 7.2e-05,
|
| 20 |
+
"weight_decay": 0.0,
|
| 21 |
+
"adam_beta1": 0.9,
|
| 22 |
+
"adam_beta2": 0.999,
|
| 23 |
+
"adam_epsilon": 1e-08,
|
| 24 |
+
"max_grad_norm": 0.3,
|
| 25 |
+
"num_train_epochs": 2,
|
| 26 |
+
"max_steps": 100,
|
| 27 |
+
"lr_scheduler_type": "constant",
|
| 28 |
+
"warmup_ratio": 0.03,
|
| 29 |
+
"warmup_steps": 0,
|
| 30 |
+
"log_level": "passive",
|
| 31 |
+
"log_level_replica": "warning",
|
| 32 |
+
"log_on_each_node": true,
|
| 33 |
+
"logging_dir": "models/classify_lora/openllama-3b-code_hard/runs/Oct15_06-35-16_compute-permanent-node-978",
|
| 34 |
+
"logging_strategy": "steps",
|
| 35 |
+
"logging_first_step": false,
|
| 36 |
+
"logging_steps": 1,
|
| 37 |
+
"logging_nan_inf_filter": true,
|
| 38 |
+
"save_strategy": "steps",
|
| 39 |
+
"save_steps": 25,
|
| 40 |
+
"save_total_limit": 0,
|
| 41 |
+
"save_safetensors": false,
|
| 42 |
+
"save_on_each_node": false,
|
| 43 |
+
"no_cuda": false,
|
| 44 |
+
"use_cpu": false,
|
| 45 |
+
"use_mps_device": false,
|
| 46 |
+
"seed": 42,
|
| 47 |
+
"data_seed": null,
|
| 48 |
+
"jit_mode_eval": false,
|
| 49 |
+
"use_ipex": false,
|
| 50 |
+
"bf16": false,
|
| 51 |
+
"fp16": false,
|
| 52 |
+
"fp16_opt_level": "O1",
|
| 53 |
+
"half_precision_backend": "auto",
|
| 54 |
+
"bf16_full_eval": false,
|
| 55 |
+
"fp16_full_eval": false,
|
| 56 |
+
"tf32": null,
|
| 57 |
+
"local_rank": 0,
|
| 58 |
+
"ddp_backend": null,
|
| 59 |
+
"tpu_num_cores": null,
|
| 60 |
+
"tpu_metrics_debug": false,
|
| 61 |
+
"debug": [],
|
| 62 |
+
"dataloader_drop_last": false,
|
| 63 |
+
"eval_steps": 25,
|
| 64 |
+
"dataloader_num_workers": 0,
|
| 65 |
+
"past_index": -1,
|
| 66 |
+
"run_name": "train|models-classify_lora-openllama-3b-code_hard",
|
| 67 |
+
"disable_tqdm": false,
|
| 68 |
+
"remove_unused_columns": false,
|
| 69 |
+
"label_names": null,
|
| 70 |
+
"load_best_model_at_end": false,
|
| 71 |
+
"metric_for_best_model": "eval_code_hard_score",
|
| 72 |
+
"greater_is_better": true,
|
| 73 |
+
"ignore_data_skip": false,
|
| 74 |
+
"sharded_ddp": [],
|
| 75 |
+
"fsdp": [],
|
| 76 |
+
"fsdp_min_num_params": 0,
|
| 77 |
+
"fsdp_config": {
|
| 78 |
+
"min_num_params": 0,
|
| 79 |
+
"xla": false,
|
| 80 |
+
"xla_fsdp_grad_ckpt": false
|
| 81 |
+
},
|
| 82 |
+
"fsdp_transformer_layer_cls_to_wrap": null,
|
| 83 |
+
"deepspeed": "configs/ds_zero_1.json",
|
| 84 |
+
"label_smoothing_factor": 0.0,
|
| 85 |
+
"optim": "paged_adamw_32bit",
|
| 86 |
+
"optim_args": null,
|
| 87 |
+
"adafactor": false,
|
| 88 |
+
"group_by_length": false,
|
| 89 |
+
"length_column_name": "length",
|
| 90 |
+
"report_to": [
|
| 91 |
+
"wandb"
|
| 92 |
+
],
|
| 93 |
+
"ddp_find_unused_parameters": false,
|
| 94 |
+
"ddp_bucket_cap_mb": null,
|
| 95 |
+
"ddp_broadcast_buffers": null,
|
| 96 |
+
"dataloader_pin_memory": true,
|
| 97 |
+
"skip_memory_metrics": true,
|
| 98 |
+
"use_legacy_prediction_loop": false,
|
| 99 |
+
"push_to_hub": false,
|
| 100 |
+
"resume_from_checkpoint": null,
|
| 101 |
+
"hub_model_id": null,
|
| 102 |
+
"hub_strategy": "every_save",
|
| 103 |
+
"hub_token": null,
|
| 104 |
+
"hub_private_repo": false,
|
| 105 |
+
"hub_always_push": false,
|
| 106 |
+
"gradient_checkpointing": false,
|
| 107 |
+
"include_inputs_for_metrics": false,
|
| 108 |
+
"fp16_backend": "auto",
|
| 109 |
+
"push_to_hub_model_id": null,
|
| 110 |
+
"push_to_hub_organization": null,
|
| 111 |
+
"push_to_hub_token": null,
|
| 112 |
+
"_n_gpu": 1,
|
| 113 |
+
"mp_parameters": "",
|
| 114 |
+
"auto_find_batch_size": false,
|
| 115 |
+
"full_determinism": false,
|
| 116 |
+
"torchdynamo": null,
|
| 117 |
+
"ray_scope": "last",
|
| 118 |
+
"ddp_timeout": 1800,
|
| 119 |
+
"torch_compile": false,
|
| 120 |
+
"torch_compile_backend": null,
|
| 121 |
+
"torch_compile_mode": null,
|
| 122 |
+
"dispatch_batches": null
|
| 123 |
+
}
|
training_args.bin
ADDED
|
@@ -0,0 +1,3 @@
|
|
|
|
|
|
|
|
|
|
|
|
|
| 1 |
+
version https://git-lfs.github.com/spec/v1
|
| 2 |
+
oid sha256:7bcc80cd98af14edaa54b80a5d5dcf4a28280e000b9164e0d9ec43e143747cbd
|
| 3 |
+
size 5115
|
training_logs.json
ADDED
|
@@ -0,0 +1,1067 @@
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
| 1 |
+
[
|
| 2 |
+
{
|
| 3 |
+
"loss": 0.6932,
|
| 4 |
+
"learning_rate": 7.2e-05,
|
| 5 |
+
"epoch": 0.02,
|
| 6 |
+
"step": 1
|
| 7 |
+
},
|
| 8 |
+
{
|
| 9 |
+
"loss": 0.6489,
|
| 10 |
+
"learning_rate": 7.2e-05,
|
| 11 |
+
"epoch": 0.04,
|
| 12 |
+
"step": 2
|
| 13 |
+
},
|
| 14 |
+
{
|
| 15 |
+
"loss": 0.5662,
|
| 16 |
+
"learning_rate": 7.2e-05,
|
| 17 |
+
"epoch": 0.05,
|
| 18 |
+
"step": 3
|
| 19 |
+
},
|
| 20 |
+
{
|
| 21 |
+
"loss": 0.5949,
|
| 22 |
+
"learning_rate": 7.2e-05,
|
| 23 |
+
"epoch": 0.07,
|
| 24 |
+
"step": 4
|
| 25 |
+
},
|
| 26 |
+
{
|
| 27 |
+
"loss": 0.6451,
|
| 28 |
+
"learning_rate": 7.2e-05,
|
| 29 |
+
"epoch": 0.09,
|
| 30 |
+
"step": 5
|
| 31 |
+
},
|
| 32 |
+
{
|
| 33 |
+
"loss": 0.5372,
|
| 34 |
+
"learning_rate": 7.2e-05,
|
| 35 |
+
"epoch": 0.11,
|
| 36 |
+
"step": 6
|
| 37 |
+
},
|
| 38 |
+
{
|
| 39 |
+
"loss": 0.6065,
|
| 40 |
+
"learning_rate": 7.2e-05,
|
| 41 |
+
"epoch": 0.12,
|
| 42 |
+
"step": 7
|
| 43 |
+
},
|
| 44 |
+
{
|
| 45 |
+
"loss": 0.6478,
|
| 46 |
+
"learning_rate": 7.2e-05,
|
| 47 |
+
"epoch": 0.14,
|
| 48 |
+
"step": 8
|
| 49 |
+
},
|
| 50 |
+
{
|
| 51 |
+
"loss": 0.676,
|
| 52 |
+
"learning_rate": 7.2e-05,
|
| 53 |
+
"epoch": 0.16,
|
| 54 |
+
"step": 9
|
| 55 |
+
},
|
| 56 |
+
{
|
| 57 |
+
"loss": 0.5954,
|
| 58 |
+
"learning_rate": 7.2e-05,
|
| 59 |
+
"epoch": 0.18,
|
| 60 |
+
"step": 10
|
| 61 |
+
},
|
| 62 |
+
{
|
| 63 |
+
"loss": 0.5809,
|
| 64 |
+
"learning_rate": 7.2e-05,
|
| 65 |
+
"epoch": 0.19,
|
| 66 |
+
"step": 11
|
| 67 |
+
},
|
| 68 |
+
{
|
| 69 |
+
"loss": 0.5934,
|
| 70 |
+
"learning_rate": 7.2e-05,
|
| 71 |
+
"epoch": 0.21,
|
| 72 |
+
"step": 12
|
| 73 |
+
},
|
| 74 |
+
{
|
| 75 |
+
"loss": 0.6806,
|
| 76 |
+
"learning_rate": 7.2e-05,
|
| 77 |
+
"epoch": 0.23,
|
| 78 |
+
"step": 13
|
| 79 |
+
},
|
| 80 |
+
{
|
| 81 |
+
"loss": 0.5866,
|
| 82 |
+
"learning_rate": 7.2e-05,
|
| 83 |
+
"epoch": 0.25,
|
| 84 |
+
"step": 14
|
| 85 |
+
},
|
| 86 |
+
{
|
| 87 |
+
"loss": 0.6057,
|
| 88 |
+
"learning_rate": 7.2e-05,
|
| 89 |
+
"epoch": 0.26,
|
| 90 |
+
"step": 15
|
| 91 |
+
},
|
| 92 |
+
{
|
| 93 |
+
"loss": 0.6071,
|
| 94 |
+
"learning_rate": 7.2e-05,
|
| 95 |
+
"epoch": 0.28,
|
| 96 |
+
"step": 16
|
| 97 |
+
},
|
| 98 |
+
{
|
| 99 |
+
"loss": 0.484,
|
| 100 |
+
"learning_rate": 7.2e-05,
|
| 101 |
+
"epoch": 0.3,
|
| 102 |
+
"step": 17
|
| 103 |
+
},
|
| 104 |
+
{
|
| 105 |
+
"loss": 0.5456,
|
| 106 |
+
"learning_rate": 7.2e-05,
|
| 107 |
+
"epoch": 0.32,
|
| 108 |
+
"step": 18
|
| 109 |
+
},
|
| 110 |
+
{
|
| 111 |
+
"loss": 0.5866,
|
| 112 |
+
"learning_rate": 7.2e-05,
|
| 113 |
+
"epoch": 0.33,
|
| 114 |
+
"step": 19
|
| 115 |
+
},
|
| 116 |
+
{
|
| 117 |
+
"loss": 0.5685,
|
| 118 |
+
"learning_rate": 7.2e-05,
|
| 119 |
+
"epoch": 0.35,
|
| 120 |
+
"step": 20
|
| 121 |
+
},
|
| 122 |
+
{
|
| 123 |
+
"loss": 0.561,
|
| 124 |
+
"learning_rate": 7.2e-05,
|
| 125 |
+
"epoch": 0.37,
|
| 126 |
+
"step": 21
|
| 127 |
+
},
|
| 128 |
+
{
|
| 129 |
+
"loss": 0.6703,
|
| 130 |
+
"learning_rate": 7.2e-05,
|
| 131 |
+
"epoch": 0.39,
|
| 132 |
+
"step": 22
|
| 133 |
+
},
|
| 134 |
+
{
|
| 135 |
+
"loss": 0.5549,
|
| 136 |
+
"learning_rate": 7.2e-05,
|
| 137 |
+
"epoch": 0.4,
|
| 138 |
+
"step": 23
|
| 139 |
+
},
|
| 140 |
+
{
|
| 141 |
+
"loss": 0.5696,
|
| 142 |
+
"learning_rate": 7.2e-05,
|
| 143 |
+
"epoch": 0.42,
|
| 144 |
+
"step": 24
|
| 145 |
+
},
|
| 146 |
+
{
|
| 147 |
+
"loss": 0.6033,
|
| 148 |
+
"learning_rate": 7.2e-05,
|
| 149 |
+
"epoch": 0.44,
|
| 150 |
+
"step": 25
|
| 151 |
+
},
|
| 152 |
+
{
|
| 153 |
+
"eval_code_hard_loss": 0.5722607374191284,
|
| 154 |
+
"eval_code_hard_score": -0.20501708984375,
|
| 155 |
+
"eval_code_hard_brier_score": 0.20501708984375,
|
| 156 |
+
"eval_code_hard_average_probability": 0.5791293978691101,
|
| 157 |
+
"eval_code_hard_accuracy": 0.72,
|
| 158 |
+
"eval_code_hard_probabilities": [
|
| 159 |
+
0.5183141827583313,
|
| 160 |
+
0.5020153522491455,
|
| 161 |
+
0.4965677261352539,
|
| 162 |
+
0.9698847532272339,
|
| 163 |
+
0.9712179899215698,
|
| 164 |
+
0.9694114327430725,
|
| 165 |
+
0.5282353162765503,
|
| 166 |
+
0.537986695766449,
|
| 167 |
+
0.539186954498291,
|
| 168 |
+
0.5343082547187805,
|
| 169 |
+
0.5109472274780273,
|
| 170 |
+
0.5189101099967957,
|
| 171 |
+
0.9725697040557861,
|
| 172 |
+
0.9698337316513062,
|
| 173 |
+
0.9678789973258972,
|
| 174 |
+
0.4937111735343933,
|
| 175 |
+
0.4989865720272064,
|
| 176 |
+
0.47172561287879944,
|
| 177 |
+
0.9673305153846741,
|
| 178 |
+
0.974334180355072,
|
| 179 |
+
0.9758511781692505,
|
| 180 |
+
0.5213514566421509,
|
| 181 |
+
0.5089063048362732,
|
| 182 |
+
0.5075064301490784,
|
| 183 |
+
0.528186023235321,
|
| 184 |
+
0.5266287922859192,
|
| 185 |
+
0.5207107067108154,
|
| 186 |
+
0.6874252557754517,
|
| 187 |
+
0.5859951376914978,
|
| 188 |
+
0.5912247896194458,
|
| 189 |
+
0.4992046058177948,
|
| 190 |
+
0.49837902188301086,
|
| 191 |
+
0.48732590675354004,
|
| 192 |
+
0.5029973983764648,
|
| 193 |
+
0.5089282989501953,
|
| 194 |
+
0.5102050304412842,
|
| 195 |
+
0.49744459986686707,
|
| 196 |
+
0.5039994120597839,
|
| 197 |
+
0.5496022701263428,
|
| 198 |
+
0.4445340037345886,
|
| 199 |
+
0.4402852952480316,
|
| 200 |
+
0.4423198997974396,
|
| 201 |
+
0.5244624018669128,
|
| 202 |
+
0.5203747749328613,
|
| 203 |
+
0.510670006275177,
|
| 204 |
+
0.5222905874252319,
|
| 205 |
+
0.49989885091781616,
|
| 206 |
+
0.5032917261123657,
|
| 207 |
+
0.5169360041618347,
|
| 208 |
+
0.4835692048072815,
|
| 209 |
+
0.49604979157447815,
|
| 210 |
+
0.48959624767303467,
|
| 211 |
+
0.474219411611557,
|
| 212 |
+
0.39609599113464355,
|
| 213 |
+
0.965065598487854,
|
| 214 |
+
0.9651297926902771,
|
| 215 |
+
0.9658015966415405,
|
| 216 |
+
0.4790042042732239,
|
| 217 |
+
0.5015706419944763,
|
| 218 |
+
0.5015318393707275,
|
| 219 |
+
0.9697352051734924,
|
| 220 |
+
0.969887912273407,
|
| 221 |
+
0.9698120951652527,
|
| 222 |
+
0.5096650123596191,
|
| 223 |
+
0.515565037727356,
|
| 224 |
+
0.50871741771698,
|
| 225 |
+
0.5117788910865784,
|
| 226 |
+
0.47279125452041626,
|
| 227 |
+
0.5129478573799133,
|
| 228 |
+
0.513483464717865,
|
| 229 |
+
0.45352092385292053,
|
| 230 |
+
0.4934040606021881,
|
| 231 |
+
0.5126722455024719,
|
| 232 |
+
0.5389389991760254,
|
| 233 |
+
0.517637312412262,
|
| 234 |
+
0.535071849822998,
|
| 235 |
+
0.5183783769607544,
|
| 236 |
+
0.5209357142448425,
|
| 237 |
+
0.501355767250061,
|
| 238 |
+
0.49917733669281006,
|
| 239 |
+
0.4997139871120453,
|
| 240 |
+
0.4962727725505829,
|
| 241 |
+
0.4922448694705963,
|
| 242 |
+
0.49315178394317627,
|
| 243 |
+
0.5435120463371277,
|
| 244 |
+
0.5534088611602783,
|
| 245 |
+
0.5032275319099426,
|
| 246 |
+
0.4997408390045166,
|
| 247 |
+
0.5043116807937622,
|
| 248 |
+
0.5037261843681335,
|
| 249 |
+
0.5321357846260071,
|
| 250 |
+
0.5255170464515686,
|
| 251 |
+
0.528411328792572,
|
| 252 |
+
0.5442461371421814,
|
| 253 |
+
0.5378051996231079,
|
| 254 |
+
0.5345613956451416,
|
| 255 |
+
0.5022143125534058,
|
| 256 |
+
0.5025732517242432,
|
| 257 |
+
0.5056655406951904,
|
| 258 |
+
0.497100830078125
|
| 259 |
+
],
|
| 260 |
+
"eval_code_hard_runtime": 53.2886,
|
| 261 |
+
"eval_code_hard_samples_per_second": 1.877,
|
| 262 |
+
"eval_code_hard_steps_per_second": 0.038,
|
| 263 |
+
"epoch": 0.44,
|
| 264 |
+
"step": 25
|
| 265 |
+
},
|
| 266 |
+
{
|
| 267 |
+
"loss": 0.537,
|
| 268 |
+
"learning_rate": 7.2e-05,
|
| 269 |
+
"epoch": 0.46,
|
| 270 |
+
"step": 26
|
| 271 |
+
},
|
| 272 |
+
{
|
| 273 |
+
"loss": 0.56,
|
| 274 |
+
"learning_rate": 7.2e-05,
|
| 275 |
+
"epoch": 0.47,
|
| 276 |
+
"step": 27
|
| 277 |
+
},
|
| 278 |
+
{
|
| 279 |
+
"loss": 0.6412,
|
| 280 |
+
"learning_rate": 7.2e-05,
|
| 281 |
+
"epoch": 0.49,
|
| 282 |
+
"step": 28
|
| 283 |
+
},
|
| 284 |
+
{
|
| 285 |
+
"loss": 0.5359,
|
| 286 |
+
"learning_rate": 7.2e-05,
|
| 287 |
+
"epoch": 0.51,
|
| 288 |
+
"step": 29
|
| 289 |
+
},
|
| 290 |
+
{
|
| 291 |
+
"loss": 0.5129,
|
| 292 |
+
"learning_rate": 7.2e-05,
|
| 293 |
+
"epoch": 0.53,
|
| 294 |
+
"step": 30
|
| 295 |
+
},
|
| 296 |
+
{
|
| 297 |
+
"loss": 0.5104,
|
| 298 |
+
"learning_rate": 7.2e-05,
|
| 299 |
+
"epoch": 0.54,
|
| 300 |
+
"step": 31
|
| 301 |
+
},
|
| 302 |
+
{
|
| 303 |
+
"loss": 0.4987,
|
| 304 |
+
"learning_rate": 7.2e-05,
|
| 305 |
+
"epoch": 0.56,
|
| 306 |
+
"step": 32
|
| 307 |
+
},
|
| 308 |
+
{
|
| 309 |
+
"loss": 0.4875,
|
| 310 |
+
"learning_rate": 7.2e-05,
|
| 311 |
+
"epoch": 0.58,
|
| 312 |
+
"step": 33
|
| 313 |
+
},
|
| 314 |
+
{
|
| 315 |
+
"loss": 0.4075,
|
| 316 |
+
"learning_rate": 7.2e-05,
|
| 317 |
+
"epoch": 0.6,
|
| 318 |
+
"step": 34
|
| 319 |
+
},
|
| 320 |
+
{
|
| 321 |
+
"loss": 0.4638,
|
| 322 |
+
"learning_rate": 7.2e-05,
|
| 323 |
+
"epoch": 0.61,
|
| 324 |
+
"step": 35
|
| 325 |
+
},
|
| 326 |
+
{
|
| 327 |
+
"loss": 0.4016,
|
| 328 |
+
"learning_rate": 7.2e-05,
|
| 329 |
+
"epoch": 0.63,
|
| 330 |
+
"step": 36
|
| 331 |
+
},
|
| 332 |
+
{
|
| 333 |
+
"loss": 0.5376,
|
| 334 |
+
"learning_rate": 7.2e-05,
|
| 335 |
+
"epoch": 0.65,
|
| 336 |
+
"step": 37
|
| 337 |
+
},
|
| 338 |
+
{
|
| 339 |
+
"loss": 0.5298,
|
| 340 |
+
"learning_rate": 7.2e-05,
|
| 341 |
+
"epoch": 0.67,
|
| 342 |
+
"step": 38
|
| 343 |
+
},
|
| 344 |
+
{
|
| 345 |
+
"loss": 0.4862,
|
| 346 |
+
"learning_rate": 7.2e-05,
|
| 347 |
+
"epoch": 0.68,
|
| 348 |
+
"step": 39
|
| 349 |
+
},
|
| 350 |
+
{
|
| 351 |
+
"loss": 0.5096,
|
| 352 |
+
"learning_rate": 7.2e-05,
|
| 353 |
+
"epoch": 0.7,
|
| 354 |
+
"step": 40
|
| 355 |
+
},
|
| 356 |
+
{
|
| 357 |
+
"loss": 0.4659,
|
| 358 |
+
"learning_rate": 7.2e-05,
|
| 359 |
+
"epoch": 0.72,
|
| 360 |
+
"step": 41
|
| 361 |
+
},
|
| 362 |
+
{
|
| 363 |
+
"loss": 0.4683,
|
| 364 |
+
"learning_rate": 7.2e-05,
|
| 365 |
+
"epoch": 0.74,
|
| 366 |
+
"step": 42
|
| 367 |
+
},
|
| 368 |
+
{
|
| 369 |
+
"loss": 0.427,
|
| 370 |
+
"learning_rate": 7.2e-05,
|
| 371 |
+
"epoch": 0.75,
|
| 372 |
+
"step": 43
|
| 373 |
+
},
|
| 374 |
+
{
|
| 375 |
+
"loss": 0.4062,
|
| 376 |
+
"learning_rate": 7.2e-05,
|
| 377 |
+
"epoch": 0.77,
|
| 378 |
+
"step": 44
|
| 379 |
+
},
|
| 380 |
+
{
|
| 381 |
+
"loss": 0.4772,
|
| 382 |
+
"learning_rate": 7.2e-05,
|
| 383 |
+
"epoch": 0.79,
|
| 384 |
+
"step": 45
|
| 385 |
+
},
|
| 386 |
+
{
|
| 387 |
+
"loss": 0.3566,
|
| 388 |
+
"learning_rate": 7.2e-05,
|
| 389 |
+
"epoch": 0.81,
|
| 390 |
+
"step": 46
|
| 391 |
+
},
|
| 392 |
+
{
|
| 393 |
+
"loss": 0.4675,
|
| 394 |
+
"learning_rate": 7.2e-05,
|
| 395 |
+
"epoch": 0.82,
|
| 396 |
+
"step": 47
|
| 397 |
+
},
|
| 398 |
+
{
|
| 399 |
+
"loss": 0.4146,
|
| 400 |
+
"learning_rate": 7.2e-05,
|
| 401 |
+
"epoch": 0.84,
|
| 402 |
+
"step": 48
|
| 403 |
+
},
|
| 404 |
+
{
|
| 405 |
+
"loss": 0.4207,
|
| 406 |
+
"learning_rate": 7.2e-05,
|
| 407 |
+
"epoch": 0.86,
|
| 408 |
+
"step": 49
|
| 409 |
+
},
|
| 410 |
+
{
|
| 411 |
+
"loss": 0.3584,
|
| 412 |
+
"learning_rate": 7.2e-05,
|
| 413 |
+
"epoch": 0.88,
|
| 414 |
+
"step": 50
|
| 415 |
+
},
|
| 416 |
+
{
|
| 417 |
+
"eval_code_hard_loss": 0.34301868081092834,
|
| 418 |
+
"eval_code_hard_score": -0.11159273982048035,
|
| 419 |
+
"eval_code_hard_brier_score": 0.11159273982048035,
|
| 420 |
+
"eval_code_hard_average_probability": 0.7284997701644897,
|
| 421 |
+
"eval_code_hard_accuracy": 0.86,
|
| 422 |
+
"eval_code_hard_probabilities": [
|
| 423 |
+
0.619510293006897,
|
| 424 |
+
0.6040221452713013,
|
| 425 |
+
0.5958019495010376,
|
| 426 |
+
0.9492889046669006,
|
| 427 |
+
0.9177025556564331,
|
| 428 |
+
0.9289835095405579,
|
| 429 |
+
0.8927240371704102,
|
| 430 |
+
0.9211738705635071,
|
| 431 |
+
0.9168058037757874,
|
| 432 |
+
0.7706602215766907,
|
| 433 |
+
0.7466457486152649,
|
| 434 |
+
0.8367064595222473,
|
| 435 |
+
0.9997155070304871,
|
| 436 |
+
0.9995123147964478,
|
| 437 |
+
0.9982605576515198,
|
| 438 |
+
0.4536142945289612,
|
| 439 |
+
0.5211822986602783,
|
| 440 |
+
0.4859596788883209,
|
| 441 |
+
0.9396563172340393,
|
| 442 |
+
0.9814699292182922,
|
| 443 |
+
0.9855539202690125,
|
| 444 |
+
0.9185433387756348,
|
| 445 |
+
0.9342010617256165,
|
| 446 |
+
0.9058659076690674,
|
| 447 |
+
0.8620021939277649,
|
| 448 |
+
0.8177945017814636,
|
| 449 |
+
0.9235488176345825,
|
| 450 |
+
0.9499996304512024,
|
| 451 |
+
0.9158856868743896,
|
| 452 |
+
0.9167284369468689,
|
| 453 |
+
0.4879337251186371,
|
| 454 |
+
0.43694543838500977,
|
| 455 |
+
0.4652228355407715,
|
| 456 |
+
0.896939218044281,
|
| 457 |
+
0.8988721966743469,
|
| 458 |
+
0.9008588790893555,
|
| 459 |
+
0.6782506108283997,
|
| 460 |
+
0.6183739304542542,
|
| 461 |
+
0.7814574837684631,
|
| 462 |
+
0.740852952003479,
|
| 463 |
+
0.6149900555610657,
|
| 464 |
+
0.6410658359527588,
|
| 465 |
+
0.828216552734375,
|
| 466 |
+
0.7742620706558228,
|
| 467 |
+
0.7144817113876343,
|
| 468 |
+
0.5286781191825867,
|
| 469 |
+
0.4971938729286194,
|
| 470 |
+
0.506715714931488,
|
| 471 |
+
0.7012039422988892,
|
| 472 |
+
0.3713959753513336,
|
| 473 |
+
0.4451392889022827,
|
| 474 |
+
0.5213230848312378,
|
| 475 |
+
0.495220422744751,
|
| 476 |
+
0.47032222151756287,
|
| 477 |
+
0.995394766330719,
|
| 478 |
+
0.9957503080368042,
|
| 479 |
+
0.9961684346199036,
|
| 480 |
+
0.39281144738197327,
|
| 481 |
+
0.5100150108337402,
|
| 482 |
+
0.47976943850517273,
|
| 483 |
+
0.9990038275718689,
|
| 484 |
+
0.9989190101623535,
|
| 485 |
+
0.9989215135574341,
|
| 486 |
+
0.5996841192245483,
|
| 487 |
+
0.5701720118522644,
|
| 488 |
+
0.58901047706604,
|
| 489 |
+
0.9164248108863831,
|
| 490 |
+
0.9526078104972839,
|
| 491 |
+
0.9073719382286072,
|
| 492 |
+
0.7042233943939209,
|
| 493 |
+
0.5729785561561584,
|
| 494 |
+
0.5182032585144043,
|
| 495 |
+
0.8634644150733948,
|
| 496 |
+
0.9283011555671692,
|
| 497 |
+
0.9007238745689392,
|
| 498 |
+
0.7439146637916565,
|
| 499 |
+
0.6496825814247131,
|
| 500 |
+
0.6525353789329529,
|
| 501 |
+
0.5044839382171631,
|
| 502 |
+
0.4995957612991333,
|
| 503 |
+
0.5234301090240479,
|
| 504 |
+
0.5146706104278564,
|
| 505 |
+
0.6355378031730652,
|
| 506 |
+
0.5025534629821777,
|
| 507 |
+
0.9130427837371826,
|
| 508 |
+
0.8997641205787659,
|
| 509 |
+
0.5774441957473755,
|
| 510 |
+
0.519719123840332,
|
| 511 |
+
0.5050529837608337,
|
| 512 |
+
0.5099009275436401,
|
| 513 |
+
0.8784401416778564,
|
| 514 |
+
0.886016845703125,
|
| 515 |
+
0.8805269598960876,
|
| 516 |
+
0.8014693856239319,
|
| 517 |
+
0.7851292490959167,
|
| 518 |
+
0.8030146956443787,
|
| 519 |
+
0.5338838696479797,
|
| 520 |
+
0.5891290903091431,
|
| 521 |
+
0.3894357979297638,
|
| 522 |
+
0.5362494587898254
|
| 523 |
+
],
|
| 524 |
+
"eval_code_hard_runtime": 53.3003,
|
| 525 |
+
"eval_code_hard_samples_per_second": 1.876,
|
| 526 |
+
"eval_code_hard_steps_per_second": 0.038,
|
| 527 |
+
"epoch": 0.88,
|
| 528 |
+
"step": 50
|
| 529 |
+
},
|
| 530 |
+
{
|
| 531 |
+
"loss": 0.4189,
|
| 532 |
+
"learning_rate": 7.2e-05,
|
| 533 |
+
"epoch": 0.89,
|
| 534 |
+
"step": 51
|
| 535 |
+
},
|
| 536 |
+
{
|
| 537 |
+
"loss": 0.3505,
|
| 538 |
+
"learning_rate": 7.2e-05,
|
| 539 |
+
"epoch": 0.91,
|
| 540 |
+
"step": 52
|
| 541 |
+
},
|
| 542 |
+
{
|
| 543 |
+
"loss": 0.3945,
|
| 544 |
+
"learning_rate": 7.2e-05,
|
| 545 |
+
"epoch": 0.93,
|
| 546 |
+
"step": 53
|
| 547 |
+
},
|
| 548 |
+
{
|
| 549 |
+
"loss": 0.4302,
|
| 550 |
+
"learning_rate": 7.2e-05,
|
| 551 |
+
"epoch": 0.95,
|
| 552 |
+
"step": 54
|
| 553 |
+
},
|
| 554 |
+
{
|
| 555 |
+
"loss": 0.3197,
|
| 556 |
+
"learning_rate": 7.2e-05,
|
| 557 |
+
"epoch": 0.96,
|
| 558 |
+
"step": 55
|
| 559 |
+
},
|
| 560 |
+
{
|
| 561 |
+
"loss": 0.3941,
|
| 562 |
+
"learning_rate": 7.2e-05,
|
| 563 |
+
"epoch": 0.98,
|
| 564 |
+
"step": 56
|
| 565 |
+
},
|
| 566 |
+
{
|
| 567 |
+
"loss": 0.2922,
|
| 568 |
+
"learning_rate": 7.2e-05,
|
| 569 |
+
"epoch": 1.0,
|
| 570 |
+
"step": 57
|
| 571 |
+
},
|
| 572 |
+
{
|
| 573 |
+
"loss": 0.4051,
|
| 574 |
+
"learning_rate": 7.2e-05,
|
| 575 |
+
"epoch": 1.02,
|
| 576 |
+
"step": 58
|
| 577 |
+
},
|
| 578 |
+
{
|
| 579 |
+
"loss": 0.3588,
|
| 580 |
+
"learning_rate": 7.2e-05,
|
| 581 |
+
"epoch": 1.04,
|
| 582 |
+
"step": 59
|
| 583 |
+
},
|
| 584 |
+
{
|
| 585 |
+
"loss": 0.3388,
|
| 586 |
+
"learning_rate": 7.2e-05,
|
| 587 |
+
"epoch": 1.05,
|
| 588 |
+
"step": 60
|
| 589 |
+
},
|
| 590 |
+
{
|
| 591 |
+
"loss": 0.3442,
|
| 592 |
+
"learning_rate": 7.2e-05,
|
| 593 |
+
"epoch": 1.07,
|
| 594 |
+
"step": 61
|
| 595 |
+
},
|
| 596 |
+
{
|
| 597 |
+
"loss": 0.2811,
|
| 598 |
+
"learning_rate": 7.2e-05,
|
| 599 |
+
"epoch": 1.09,
|
| 600 |
+
"step": 62
|
| 601 |
+
},
|
| 602 |
+
{
|
| 603 |
+
"loss": 0.2847,
|
| 604 |
+
"learning_rate": 7.2e-05,
|
| 605 |
+
"epoch": 1.11,
|
| 606 |
+
"step": 63
|
| 607 |
+
},
|
| 608 |
+
{
|
| 609 |
+
"loss": 0.3596,
|
| 610 |
+
"learning_rate": 7.2e-05,
|
| 611 |
+
"epoch": 1.12,
|
| 612 |
+
"step": 64
|
| 613 |
+
},
|
| 614 |
+
{
|
| 615 |
+
"loss": 0.4253,
|
| 616 |
+
"learning_rate": 7.2e-05,
|
| 617 |
+
"epoch": 1.14,
|
| 618 |
+
"step": 65
|
| 619 |
+
},
|
| 620 |
+
{
|
| 621 |
+
"loss": 0.2929,
|
| 622 |
+
"learning_rate": 7.2e-05,
|
| 623 |
+
"epoch": 1.16,
|
| 624 |
+
"step": 66
|
| 625 |
+
},
|
| 626 |
+
{
|
| 627 |
+
"loss": 0.2756,
|
| 628 |
+
"learning_rate": 7.2e-05,
|
| 629 |
+
"epoch": 1.18,
|
| 630 |
+
"step": 67
|
| 631 |
+
},
|
| 632 |
+
{
|
| 633 |
+
"loss": 0.3144,
|
| 634 |
+
"learning_rate": 7.2e-05,
|
| 635 |
+
"epoch": 1.19,
|
| 636 |
+
"step": 68
|
| 637 |
+
},
|
| 638 |
+
{
|
| 639 |
+
"loss": 0.2823,
|
| 640 |
+
"learning_rate": 7.2e-05,
|
| 641 |
+
"epoch": 1.21,
|
| 642 |
+
"step": 69
|
| 643 |
+
},
|
| 644 |
+
{
|
| 645 |
+
"loss": 0.2129,
|
| 646 |
+
"learning_rate": 7.2e-05,
|
| 647 |
+
"epoch": 1.23,
|
| 648 |
+
"step": 70
|
| 649 |
+
},
|
| 650 |
+
{
|
| 651 |
+
"loss": 0.2722,
|
| 652 |
+
"learning_rate": 7.2e-05,
|
| 653 |
+
"epoch": 1.25,
|
| 654 |
+
"step": 71
|
| 655 |
+
},
|
| 656 |
+
{
|
| 657 |
+
"loss": 0.2964,
|
| 658 |
+
"learning_rate": 7.2e-05,
|
| 659 |
+
"epoch": 1.26,
|
| 660 |
+
"step": 72
|
| 661 |
+
},
|
| 662 |
+
{
|
| 663 |
+
"loss": 0.3038,
|
| 664 |
+
"learning_rate": 7.2e-05,
|
| 665 |
+
"epoch": 1.28,
|
| 666 |
+
"step": 73
|
| 667 |
+
},
|
| 668 |
+
{
|
| 669 |
+
"loss": 0.2398,
|
| 670 |
+
"learning_rate": 7.2e-05,
|
| 671 |
+
"epoch": 1.3,
|
| 672 |
+
"step": 74
|
| 673 |
+
},
|
| 674 |
+
{
|
| 675 |
+
"loss": 0.2923,
|
| 676 |
+
"learning_rate": 7.2e-05,
|
| 677 |
+
"epoch": 1.32,
|
| 678 |
+
"step": 75
|
| 679 |
+
},
|
| 680 |
+
{
|
| 681 |
+
"eval_code_hard_loss": 0.24489852786064148,
|
| 682 |
+
"eval_code_hard_score": -0.08631329238414764,
|
| 683 |
+
"eval_code_hard_brier_score": 0.08631329238414764,
|
| 684 |
+
"eval_code_hard_average_probability": 0.8100240230560303,
|
| 685 |
+
"eval_code_hard_accuracy": 0.85,
|
| 686 |
+
"eval_code_hard_probabilities": [
|
| 687 |
+
0.7177616953849792,
|
| 688 |
+
0.795508623123169,
|
| 689 |
+
0.7857334613800049,
|
| 690 |
+
0.9991870522499084,
|
| 691 |
+
0.9987917542457581,
|
| 692 |
+
0.9994052648544312,
|
| 693 |
+
0.9846364259719849,
|
| 694 |
+
0.9944690465927124,
|
| 695 |
+
0.9703215956687927,
|
| 696 |
+
0.778681755065918,
|
| 697 |
+
0.971013069152832,
|
| 698 |
+
0.9788711071014404,
|
| 699 |
+
0.9999998807907104,
|
| 700 |
+
1.0,
|
| 701 |
+
1.0,
|
| 702 |
+
0.35785171389579773,
|
| 703 |
+
0.4882768392562866,
|
| 704 |
+
0.6737111806869507,
|
| 705 |
+
0.9996606111526489,
|
| 706 |
+
0.9999253749847412,
|
| 707 |
+
0.9999637603759766,
|
| 708 |
+
0.9994186162948608,
|
| 709 |
+
0.9990422129631042,
|
| 710 |
+
0.9974905252456665,
|
| 711 |
+
0.9995825886726379,
|
| 712 |
+
0.9953965544700623,
|
| 713 |
+
0.9997794032096863,
|
| 714 |
+
0.9983587861061096,
|
| 715 |
+
0.9997310042381287,
|
| 716 |
+
0.9997231364250183,
|
| 717 |
+
0.4219929277896881,
|
| 718 |
+
0.2606010437011719,
|
| 719 |
+
0.6489297151565552,
|
| 720 |
+
0.9547426104545593,
|
| 721 |
+
0.9553850889205933,
|
| 722 |
+
0.950897216796875,
|
| 723 |
+
0.9628025889396667,
|
| 724 |
+
0.9581644535064697,
|
| 725 |
+
0.8295655846595764,
|
| 726 |
+
0.9866873621940613,
|
| 727 |
+
0.9280973672866821,
|
| 728 |
+
0.9797408580780029,
|
| 729 |
+
0.9973568916320801,
|
| 730 |
+
0.9953936338424683,
|
| 731 |
+
0.9935207962989807,
|
| 732 |
+
0.5510905385017395,
|
| 733 |
+
0.49630528688430786,
|
| 734 |
+
0.4779393970966339,
|
| 735 |
+
0.8343468308448792,
|
| 736 |
+
0.43401989340782166,
|
| 737 |
+
0.3891345262527466,
|
| 738 |
+
0.4458746612071991,
|
| 739 |
+
0.6388282179832458,
|
| 740 |
+
0.8165429830551147,
|
| 741 |
+
0.9992231130599976,
|
| 742 |
+
0.9993274211883545,
|
| 743 |
+
0.9994113445281982,
|
| 744 |
+
0.41094741225242615,
|
| 745 |
+
0.5163617134094238,
|
| 746 |
+
0.488020658493042,
|
| 747 |
+
0.999996542930603,
|
| 748 |
+
0.9999948740005493,
|
| 749 |
+
0.9999948740005493,
|
| 750 |
+
0.6086482405662537,
|
| 751 |
+
0.6147478222846985,
|
| 752 |
+
0.6657524108886719,
|
| 753 |
+
0.9949613809585571,
|
| 754 |
+
0.9949374198913574,
|
| 755 |
+
0.9945763349533081,
|
| 756 |
+
0.7646587491035461,
|
| 757 |
+
0.817954957485199,
|
| 758 |
+
0.6410450339317322,
|
| 759 |
+
0.9997159838676453,
|
| 760 |
+
0.9999265670776367,
|
| 761 |
+
0.9998390674591064,
|
| 762 |
+
0.9317758083343506,
|
| 763 |
+
0.8353602290153503,
|
| 764 |
+
0.8241412043571472,
|
| 765 |
+
0.501185417175293,
|
| 766 |
+
0.4842207431793213,
|
| 767 |
+
0.6102815270423889,
|
| 768 |
+
0.4974021911621094,
|
| 769 |
+
0.6316254138946533,
|
| 770 |
+
0.4609917998313904,
|
| 771 |
+
0.9999159574508667,
|
| 772 |
+
0.999816358089447,
|
| 773 |
+
0.6341696381568909,
|
| 774 |
+
0.5534595847129822,
|
| 775 |
+
0.5341170430183411,
|
| 776 |
+
0.5336964130401611,
|
| 777 |
+
0.9986201524734497,
|
| 778 |
+
0.9990990161895752,
|
| 779 |
+
0.9992678761482239,
|
| 780 |
+
0.9444563984870911,
|
| 781 |
+
0.9397106170654297,
|
| 782 |
+
0.9423726201057434,
|
| 783 |
+
0.6565976142883301,
|
| 784 |
+
0.8141821026802063,
|
| 785 |
+
0.16233490407466888,
|
| 786 |
+
0.6133077144622803
|
| 787 |
+
],
|
| 788 |
+
"eval_code_hard_runtime": 53.2911,
|
| 789 |
+
"eval_code_hard_samples_per_second": 1.876,
|
| 790 |
+
"eval_code_hard_steps_per_second": 0.038,
|
| 791 |
+
"epoch": 1.32,
|
| 792 |
+
"step": 75
|
| 793 |
+
},
|
| 794 |
+
{
|
| 795 |
+
"loss": 0.3058,
|
| 796 |
+
"learning_rate": 7.2e-05,
|
| 797 |
+
"epoch": 1.33,
|
| 798 |
+
"step": 76
|
| 799 |
+
},
|
| 800 |
+
{
|
| 801 |
+
"loss": 0.3695,
|
| 802 |
+
"learning_rate": 7.2e-05,
|
| 803 |
+
"epoch": 1.35,
|
| 804 |
+
"step": 77
|
| 805 |
+
},
|
| 806 |
+
{
|
| 807 |
+
"loss": 0.1812,
|
| 808 |
+
"learning_rate": 7.2e-05,
|
| 809 |
+
"epoch": 1.37,
|
| 810 |
+
"step": 78
|
| 811 |
+
},
|
| 812 |
+
{
|
| 813 |
+
"loss": 0.2968,
|
| 814 |
+
"learning_rate": 7.2e-05,
|
| 815 |
+
"epoch": 1.39,
|
| 816 |
+
"step": 79
|
| 817 |
+
},
|
| 818 |
+
{
|
| 819 |
+
"loss": 0.3167,
|
| 820 |
+
"learning_rate": 7.2e-05,
|
| 821 |
+
"epoch": 1.4,
|
| 822 |
+
"step": 80
|
| 823 |
+
},
|
| 824 |
+
{
|
| 825 |
+
"loss": 0.488,
|
| 826 |
+
"learning_rate": 7.2e-05,
|
| 827 |
+
"epoch": 1.42,
|
| 828 |
+
"step": 81
|
| 829 |
+
},
|
| 830 |
+
{
|
| 831 |
+
"loss": 0.3406,
|
| 832 |
+
"learning_rate": 7.2e-05,
|
| 833 |
+
"epoch": 1.44,
|
| 834 |
+
"step": 82
|
| 835 |
+
},
|
| 836 |
+
{
|
| 837 |
+
"loss": 0.3298,
|
| 838 |
+
"learning_rate": 7.2e-05,
|
| 839 |
+
"epoch": 1.46,
|
| 840 |
+
"step": 83
|
| 841 |
+
},
|
| 842 |
+
{
|
| 843 |
+
"loss": 0.3004,
|
| 844 |
+
"learning_rate": 7.2e-05,
|
| 845 |
+
"epoch": 1.47,
|
| 846 |
+
"step": 84
|
| 847 |
+
},
|
| 848 |
+
{
|
| 849 |
+
"loss": 0.1752,
|
| 850 |
+
"learning_rate": 7.2e-05,
|
| 851 |
+
"epoch": 1.49,
|
| 852 |
+
"step": 85
|
| 853 |
+
},
|
| 854 |
+
{
|
| 855 |
+
"loss": 0.2629,
|
| 856 |
+
"learning_rate": 7.2e-05,
|
| 857 |
+
"epoch": 1.51,
|
| 858 |
+
"step": 86
|
| 859 |
+
},
|
| 860 |
+
{
|
| 861 |
+
"loss": 0.2125,
|
| 862 |
+
"learning_rate": 7.2e-05,
|
| 863 |
+
"epoch": 1.53,
|
| 864 |
+
"step": 87
|
| 865 |
+
},
|
| 866 |
+
{
|
| 867 |
+
"loss": 0.2029,
|
| 868 |
+
"learning_rate": 7.2e-05,
|
| 869 |
+
"epoch": 1.54,
|
| 870 |
+
"step": 88
|
| 871 |
+
},
|
| 872 |
+
{
|
| 873 |
+
"loss": 0.2783,
|
| 874 |
+
"learning_rate": 7.2e-05,
|
| 875 |
+
"epoch": 1.56,
|
| 876 |
+
"step": 89
|
| 877 |
+
},
|
| 878 |
+
{
|
| 879 |
+
"loss": 0.207,
|
| 880 |
+
"learning_rate": 7.2e-05,
|
| 881 |
+
"epoch": 1.58,
|
| 882 |
+
"step": 90
|
| 883 |
+
},
|
| 884 |
+
{
|
| 885 |
+
"loss": 0.2904,
|
| 886 |
+
"learning_rate": 7.2e-05,
|
| 887 |
+
"epoch": 1.6,
|
| 888 |
+
"step": 91
|
| 889 |
+
},
|
| 890 |
+
{
|
| 891 |
+
"loss": 0.1934,
|
| 892 |
+
"learning_rate": 7.2e-05,
|
| 893 |
+
"epoch": 1.61,
|
| 894 |
+
"step": 92
|
| 895 |
+
},
|
| 896 |
+
{
|
| 897 |
+
"loss": 0.3917,
|
| 898 |
+
"learning_rate": 7.2e-05,
|
| 899 |
+
"epoch": 1.63,
|
| 900 |
+
"step": 93
|
| 901 |
+
},
|
| 902 |
+
{
|
| 903 |
+
"loss": 0.194,
|
| 904 |
+
"learning_rate": 7.2e-05,
|
| 905 |
+
"epoch": 1.65,
|
| 906 |
+
"step": 94
|
| 907 |
+
},
|
| 908 |
+
{
|
| 909 |
+
"loss": 0.2205,
|
| 910 |
+
"learning_rate": 7.2e-05,
|
| 911 |
+
"epoch": 1.67,
|
| 912 |
+
"step": 95
|
| 913 |
+
},
|
| 914 |
+
{
|
| 915 |
+
"loss": 0.1669,
|
| 916 |
+
"learning_rate": 7.2e-05,
|
| 917 |
+
"epoch": 1.68,
|
| 918 |
+
"step": 96
|
| 919 |
+
},
|
| 920 |
+
{
|
| 921 |
+
"loss": 0.2966,
|
| 922 |
+
"learning_rate": 7.2e-05,
|
| 923 |
+
"epoch": 1.7,
|
| 924 |
+
"step": 97
|
| 925 |
+
},
|
| 926 |
+
{
|
| 927 |
+
"loss": 0.2516,
|
| 928 |
+
"learning_rate": 7.2e-05,
|
| 929 |
+
"epoch": 1.72,
|
| 930 |
+
"step": 98
|
| 931 |
+
},
|
| 932 |
+
{
|
| 933 |
+
"loss": 0.2276,
|
| 934 |
+
"learning_rate": 7.2e-05,
|
| 935 |
+
"epoch": 1.74,
|
| 936 |
+
"step": 99
|
| 937 |
+
},
|
| 938 |
+
{
|
| 939 |
+
"loss": 0.1814,
|
| 940 |
+
"learning_rate": 7.2e-05,
|
| 941 |
+
"epoch": 1.75,
|
| 942 |
+
"step": 100
|
| 943 |
+
},
|
| 944 |
+
{
|
| 945 |
+
"eval_code_hard_loss": 0.2296922653913498,
|
| 946 |
+
"eval_code_hard_score": -0.08056112378835678,
|
| 947 |
+
"eval_code_hard_brier_score": 0.08056112378835678,
|
| 948 |
+
"eval_code_hard_average_probability": 0.8206403255462646,
|
| 949 |
+
"eval_code_hard_accuracy": 0.84,
|
| 950 |
+
"eval_code_hard_probabilities": [
|
| 951 |
+
0.7861270308494568,
|
| 952 |
+
0.8348590731620789,
|
| 953 |
+
0.8349205851554871,
|
| 954 |
+
0.9997000694274902,
|
| 955 |
+
0.9995880722999573,
|
| 956 |
+
0.9998056292533875,
|
| 957 |
+
0.9992101192474365,
|
| 958 |
+
0.9999064207077026,
|
| 959 |
+
0.9989945292472839,
|
| 960 |
+
0.8230730891227722,
|
| 961 |
+
0.9760466814041138,
|
| 962 |
+
0.9709921479225159,
|
| 963 |
+
1.0,
|
| 964 |
+
1.0,
|
| 965 |
+
1.0,
|
| 966 |
+
0.460991233587265,
|
| 967 |
+
0.6684326529502869,
|
| 968 |
+
0.957171618938446,
|
| 969 |
+
0.999259889125824,
|
| 970 |
+
0.9999682903289795,
|
| 971 |
+
0.9999852180480957,
|
| 972 |
+
0.9999778270721436,
|
| 973 |
+
0.9999755620956421,
|
| 974 |
+
0.9997803568840027,
|
| 975 |
+
0.9999879598617554,
|
| 976 |
+
0.9971315860748291,
|
| 977 |
+
0.9999914169311523,
|
| 978 |
+
0.999528169631958,
|
| 979 |
+
0.9998747110366821,
|
| 980 |
+
0.9998290538787842,
|
| 981 |
+
0.4468780755996704,
|
| 982 |
+
0.2516433894634247,
|
| 983 |
+
0.6790833473205566,
|
| 984 |
+
0.9911589026451111,
|
| 985 |
+
0.9917107820510864,
|
| 986 |
+
0.9915744662284851,
|
| 987 |
+
0.9812775254249573,
|
| 988 |
+
0.9739009141921997,
|
| 989 |
+
0.7593258023262024,
|
| 990 |
+
0.9627566933631897,
|
| 991 |
+
0.8079734444618225,
|
| 992 |
+
0.943427562713623,
|
| 993 |
+
0.9999816417694092,
|
| 994 |
+
0.9999567270278931,
|
| 995 |
+
0.9999591112136841,
|
| 996 |
+
0.5994269847869873,
|
| 997 |
+
0.48037776350975037,
|
| 998 |
+
0.4851032495498657,
|
| 999 |
+
0.9079890847206116,
|
| 1000 |
+
0.4937061071395874,
|
| 1001 |
+
0.4938817620277405,
|
| 1002 |
+
0.4849710762500763,
|
| 1003 |
+
0.7558672428131104,
|
| 1004 |
+
0.8851609230041504,
|
| 1005 |
+
0.9924831986427307,
|
| 1006 |
+
0.9915366172790527,
|
| 1007 |
+
0.9959695339202881,
|
| 1008 |
+
0.4088115990161896,
|
| 1009 |
+
0.5256107449531555,
|
| 1010 |
+
0.4217263460159302,
|
| 1011 |
+
0.999980092048645,
|
| 1012 |
+
0.999962568283081,
|
| 1013 |
+
0.9999624490737915,
|
| 1014 |
+
0.7795777916908264,
|
| 1015 |
+
0.6225330829620361,
|
| 1016 |
+
0.70358806848526,
|
| 1017 |
+
0.9998552799224854,
|
| 1018 |
+
0.9998051524162292,
|
| 1019 |
+
0.9998724460601807,
|
| 1020 |
+
0.40740594267845154,
|
| 1021 |
+
0.7388888597488403,
|
| 1022 |
+
0.4775027632713318,
|
| 1023 |
+
0.9999308586120605,
|
| 1024 |
+
0.9999823570251465,
|
| 1025 |
+
0.9999585151672363,
|
| 1026 |
+
0.9347867369651794,
|
| 1027 |
+
0.5755781531333923,
|
| 1028 |
+
0.6215885877609253,
|
| 1029 |
+
0.5002992153167725,
|
| 1030 |
+
0.5065785050392151,
|
| 1031 |
+
0.699110209941864,
|
| 1032 |
+
0.472739577293396,
|
| 1033 |
+
0.6727705597877502,
|
| 1034 |
+
0.4803294241428375,
|
| 1035 |
+
0.9999979734420776,
|
| 1036 |
+
0.9999915361404419,
|
| 1037 |
+
0.8218505382537842,
|
| 1038 |
+
0.5483625531196594,
|
| 1039 |
+
0.5385692119598389,
|
| 1040 |
+
0.49461352825164795,
|
| 1041 |
+
0.9993742108345032,
|
| 1042 |
+
0.99980229139328,
|
| 1043 |
+
0.9998019337654114,
|
| 1044 |
+
0.993417501449585,
|
| 1045 |
+
0.9916191101074219,
|
| 1046 |
+
0.9953900575637817,
|
| 1047 |
+
0.7289006114006042,
|
| 1048 |
+
0.8122584223747253,
|
| 1049 |
+
0.2769394814968109,
|
| 1050 |
+
0.6622205972671509
|
| 1051 |
+
],
|
| 1052 |
+
"eval_code_hard_runtime": 53.2227,
|
| 1053 |
+
"eval_code_hard_samples_per_second": 1.879,
|
| 1054 |
+
"eval_code_hard_steps_per_second": 0.038,
|
| 1055 |
+
"epoch": 1.75,
|
| 1056 |
+
"step": 100
|
| 1057 |
+
},
|
| 1058 |
+
{
|
| 1059 |
+
"train_runtime": 4004.592,
|
| 1060 |
+
"train_samples_per_second": 0.799,
|
| 1061 |
+
"train_steps_per_second": 0.025,
|
| 1062 |
+
"total_flos": 0.0,
|
| 1063 |
+
"train_loss": 0.417534539103508,
|
| 1064 |
+
"epoch": 1.75,
|
| 1065 |
+
"step": 100
|
| 1066 |
+
}
|
| 1067 |
+
]]
|