Commit ·
25e1204
1
Parent(s): ee56f37
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 +1068 -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 sha256:4b1efbfc008a9c08ea960b1cccc3d096d466b5a5a2c81bc407195d28af3caeca
|
| 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_is_correct",
|
| 4 |
+
"date_trained": "15/10/2023 07:00:33",
|
| 5 |
+
"output_dir": "models/classify_lora/openllama-3b-code_is_correct",
|
| 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": 6,
|
| 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_is_correct/runs/Oct15_06-28-45_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_is_correct",
|
| 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_is_correct_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 sha256:87b6e5003a305e01f3573699c55609e986e2ccb1ecdc1a2067eebdb751377a69
|
| 3 |
+
size 5115
|
training_logs.json
ADDED
|
@@ -0,0 +1,1068 @@
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
| 1 |
+
[
|
| 2 |
+
{
|
| 3 |
+
"loss": 0.7294,
|
| 4 |
+
"learning_rate": 7.2e-05,
|
| 5 |
+
"epoch": 0.05,
|
| 6 |
+
"step": 1
|
| 7 |
+
},
|
| 8 |
+
{
|
| 9 |
+
"loss": 0.6837,
|
| 10 |
+
"learning_rate": 7.2e-05,
|
| 11 |
+
"epoch": 0.11,
|
| 12 |
+
"step": 2
|
| 13 |
+
},
|
| 14 |
+
{
|
| 15 |
+
"loss": 0.7207,
|
| 16 |
+
"learning_rate": 7.2e-05,
|
| 17 |
+
"epoch": 0.16,
|
| 18 |
+
"step": 3
|
| 19 |
+
},
|
| 20 |
+
{
|
| 21 |
+
"loss": 0.7014,
|
| 22 |
+
"learning_rate": 7.2e-05,
|
| 23 |
+
"epoch": 0.21,
|
| 24 |
+
"step": 4
|
| 25 |
+
},
|
| 26 |
+
{
|
| 27 |
+
"loss": 0.6994,
|
| 28 |
+
"learning_rate": 7.2e-05,
|
| 29 |
+
"epoch": 0.26,
|
| 30 |
+
"step": 5
|
| 31 |
+
},
|
| 32 |
+
{
|
| 33 |
+
"loss": 0.6877,
|
| 34 |
+
"learning_rate": 7.2e-05,
|
| 35 |
+
"epoch": 0.32,
|
| 36 |
+
"step": 6
|
| 37 |
+
},
|
| 38 |
+
{
|
| 39 |
+
"loss": 0.6923,
|
| 40 |
+
"learning_rate": 7.2e-05,
|
| 41 |
+
"epoch": 0.37,
|
| 42 |
+
"step": 7
|
| 43 |
+
},
|
| 44 |
+
{
|
| 45 |
+
"loss": 0.706,
|
| 46 |
+
"learning_rate": 7.2e-05,
|
| 47 |
+
"epoch": 0.42,
|
| 48 |
+
"step": 8
|
| 49 |
+
},
|
| 50 |
+
{
|
| 51 |
+
"loss": 0.7193,
|
| 52 |
+
"learning_rate": 7.2e-05,
|
| 53 |
+
"epoch": 0.47,
|
| 54 |
+
"step": 9
|
| 55 |
+
},
|
| 56 |
+
{
|
| 57 |
+
"loss": 0.7082,
|
| 58 |
+
"learning_rate": 7.2e-05,
|
| 59 |
+
"epoch": 0.53,
|
| 60 |
+
"step": 10
|
| 61 |
+
},
|
| 62 |
+
{
|
| 63 |
+
"loss": 0.6847,
|
| 64 |
+
"learning_rate": 7.2e-05,
|
| 65 |
+
"epoch": 0.58,
|
| 66 |
+
"step": 11
|
| 67 |
+
},
|
| 68 |
+
{
|
| 69 |
+
"loss": 0.7139,
|
| 70 |
+
"learning_rate": 7.2e-05,
|
| 71 |
+
"epoch": 0.63,
|
| 72 |
+
"step": 12
|
| 73 |
+
},
|
| 74 |
+
{
|
| 75 |
+
"loss": 0.7051,
|
| 76 |
+
"learning_rate": 7.2e-05,
|
| 77 |
+
"epoch": 0.68,
|
| 78 |
+
"step": 13
|
| 79 |
+
},
|
| 80 |
+
{
|
| 81 |
+
"loss": 0.7057,
|
| 82 |
+
"learning_rate": 7.2e-05,
|
| 83 |
+
"epoch": 0.74,
|
| 84 |
+
"step": 14
|
| 85 |
+
},
|
| 86 |
+
{
|
| 87 |
+
"loss": 0.6808,
|
| 88 |
+
"learning_rate": 7.2e-05,
|
| 89 |
+
"epoch": 0.79,
|
| 90 |
+
"step": 15
|
| 91 |
+
},
|
| 92 |
+
{
|
| 93 |
+
"loss": 0.7065,
|
| 94 |
+
"learning_rate": 7.2e-05,
|
| 95 |
+
"epoch": 0.84,
|
| 96 |
+
"step": 16
|
| 97 |
+
},
|
| 98 |
+
{
|
| 99 |
+
"loss": 0.6851,
|
| 100 |
+
"learning_rate": 7.2e-05,
|
| 101 |
+
"epoch": 0.89,
|
| 102 |
+
"step": 17
|
| 103 |
+
},
|
| 104 |
+
{
|
| 105 |
+
"loss": 0.6962,
|
| 106 |
+
"learning_rate": 7.2e-05,
|
| 107 |
+
"epoch": 0.95,
|
| 108 |
+
"step": 18
|
| 109 |
+
},
|
| 110 |
+
{
|
| 111 |
+
"loss": 0.6754,
|
| 112 |
+
"learning_rate": 7.2e-05,
|
| 113 |
+
"epoch": 1.0,
|
| 114 |
+
"step": 19
|
| 115 |
+
},
|
| 116 |
+
{
|
| 117 |
+
"loss": 0.703,
|
| 118 |
+
"learning_rate": 7.2e-05,
|
| 119 |
+
"epoch": 1.05,
|
| 120 |
+
"step": 20
|
| 121 |
+
},
|
| 122 |
+
{
|
| 123 |
+
"loss": 0.6805,
|
| 124 |
+
"learning_rate": 7.2e-05,
|
| 125 |
+
"epoch": 1.11,
|
| 126 |
+
"step": 21
|
| 127 |
+
},
|
| 128 |
+
{
|
| 129 |
+
"loss": 0.6911,
|
| 130 |
+
"learning_rate": 7.2e-05,
|
| 131 |
+
"epoch": 1.16,
|
| 132 |
+
"step": 22
|
| 133 |
+
},
|
| 134 |
+
{
|
| 135 |
+
"loss": 0.6972,
|
| 136 |
+
"learning_rate": 7.2e-05,
|
| 137 |
+
"epoch": 1.21,
|
| 138 |
+
"step": 23
|
| 139 |
+
},
|
| 140 |
+
{
|
| 141 |
+
"loss": 0.7043,
|
| 142 |
+
"learning_rate": 7.2e-05,
|
| 143 |
+
"epoch": 1.26,
|
| 144 |
+
"step": 24
|
| 145 |
+
},
|
| 146 |
+
{
|
| 147 |
+
"loss": 0.6986,
|
| 148 |
+
"learning_rate": 7.2e-05,
|
| 149 |
+
"epoch": 1.32,
|
| 150 |
+
"step": 25
|
| 151 |
+
},
|
| 152 |
+
{
|
| 153 |
+
"eval_code_is_correct_loss": 0.6981677412986755,
|
| 154 |
+
"eval_code_is_correct_score": -0.25274571776390076,
|
| 155 |
+
"eval_code_is_correct_brier_score": 0.25274571776390076,
|
| 156 |
+
"eval_code_is_correct_average_probability": 0.4975625276565552,
|
| 157 |
+
"eval_code_is_correct_accuracy": 0.43,
|
| 158 |
+
"eval_code_is_correct_probabilities": [
|
| 159 |
+
0.5088271498680115,
|
| 160 |
+
0.48198673129081726,
|
| 161 |
+
0.4981909990310669,
|
| 162 |
+
0.5024533271789551,
|
| 163 |
+
0.48107844591140747,
|
| 164 |
+
0.5089389681816101,
|
| 165 |
+
0.4841275215148926,
|
| 166 |
+
0.4966141879558563,
|
| 167 |
+
0.5106652975082397,
|
| 168 |
+
0.5127133131027222,
|
| 169 |
+
0.4725325107574463,
|
| 170 |
+
0.5244400501251221,
|
| 171 |
+
0.511084258556366,
|
| 172 |
+
0.521553635597229,
|
| 173 |
+
0.4971517324447632,
|
| 174 |
+
0.48880547285079956,
|
| 175 |
+
0.4734472632408142,
|
| 176 |
+
0.5228509306907654,
|
| 177 |
+
0.4940826892852783,
|
| 178 |
+
0.47331884503364563,
|
| 179 |
+
0.5112957954406738,
|
| 180 |
+
0.4850262403488159,
|
| 181 |
+
0.4836397171020508,
|
| 182 |
+
0.49904710054397583,
|
| 183 |
+
0.528581440448761,
|
| 184 |
+
0.5303167700767517,
|
| 185 |
+
0.5002449750900269,
|
| 186 |
+
0.5001938343048096,
|
| 187 |
+
0.4752185046672821,
|
| 188 |
+
0.514567494392395,
|
| 189 |
+
0.4778810441493988,
|
| 190 |
+
0.4975382685661316,
|
| 191 |
+
0.5223871469497681,
|
| 192 |
+
0.48589223623275757,
|
| 193 |
+
0.46854928135871887,
|
| 194 |
+
0.5050678849220276,
|
| 195 |
+
0.47625279426574707,
|
| 196 |
+
0.48999258875846863,
|
| 197 |
+
0.4709635376930237,
|
| 198 |
+
0.5040507316589355,
|
| 199 |
+
0.4879027009010315,
|
| 200 |
+
0.5210784077644348,
|
| 201 |
+
0.5200101137161255,
|
| 202 |
+
0.49343761801719666,
|
| 203 |
+
0.47951218485832214,
|
| 204 |
+
0.49272558093070984,
|
| 205 |
+
0.5004433393478394,
|
| 206 |
+
0.5177253484725952,
|
| 207 |
+
0.5200391411781311,
|
| 208 |
+
0.5149074792861938,
|
| 209 |
+
0.5044199824333191,
|
| 210 |
+
0.4863050580024719,
|
| 211 |
+
0.48952627182006836,
|
| 212 |
+
0.48324888944625854,
|
| 213 |
+
0.4773913621902466,
|
| 214 |
+
0.4935683608055115,
|
| 215 |
+
0.5272597074508667,
|
| 216 |
+
0.4823913872241974,
|
| 217 |
+
0.4688487946987152,
|
| 218 |
+
0.4804125130176544,
|
| 219 |
+
0.48838281631469727,
|
| 220 |
+
0.49300506711006165,
|
| 221 |
+
0.4761504828929901,
|
| 222 |
+
0.5559667348861694,
|
| 223 |
+
0.4856325089931488,
|
| 224 |
+
0.4806632697582245,
|
| 225 |
+
0.499020516872406,
|
| 226 |
+
0.5097540020942688,
|
| 227 |
+
0.5151796936988831,
|
| 228 |
+
0.5133408904075623,
|
| 229 |
+
0.4923568367958069,
|
| 230 |
+
0.5121997594833374,
|
| 231 |
+
0.49316778779029846,
|
| 232 |
+
0.4804587960243225,
|
| 233 |
+
0.5247389078140259,
|
| 234 |
+
0.48702362179756165,
|
| 235 |
+
0.48010408878326416,
|
| 236 |
+
0.4811219871044159,
|
| 237 |
+
0.48185238242149353,
|
| 238 |
+
0.4993566572666168,
|
| 239 |
+
0.4789592921733856,
|
| 240 |
+
0.481153279542923,
|
| 241 |
+
0.5070401430130005,
|
| 242 |
+
0.5011581182479858,
|
| 243 |
+
0.4959249794483185,
|
| 244 |
+
0.481794536113739,
|
| 245 |
+
0.4979727566242218,
|
| 246 |
+
0.5089129209518433,
|
| 247 |
+
0.5138169527053833,
|
| 248 |
+
0.48047563433647156,
|
| 249 |
+
0.5165080428123474,
|
| 250 |
+
0.5102388262748718,
|
| 251 |
+
0.515179455280304,
|
| 252 |
+
0.5202771425247192,
|
| 253 |
+
0.49016904830932617,
|
| 254 |
+
0.517183244228363,
|
| 255 |
+
0.50786954164505,
|
| 256 |
+
0.5065057873725891,
|
| 257 |
+
0.47355517745018005,
|
| 258 |
+
0.4693579375743866
|
| 259 |
+
],
|
| 260 |
+
"eval_code_is_correct_runtime": 29.2492,
|
| 261 |
+
"eval_code_is_correct_samples_per_second": 3.419,
|
| 262 |
+
"eval_code_is_correct_steps_per_second": 0.068,
|
| 263 |
+
"epoch": 1.32,
|
| 264 |
+
"step": 25
|
| 265 |
+
},
|
| 266 |
+
{
|
| 267 |
+
"loss": 0.696,
|
| 268 |
+
"learning_rate": 7.2e-05,
|
| 269 |
+
"epoch": 1.37,
|
| 270 |
+
"step": 26
|
| 271 |
+
},
|
| 272 |
+
{
|
| 273 |
+
"loss": 0.6871,
|
| 274 |
+
"learning_rate": 7.2e-05,
|
| 275 |
+
"epoch": 1.42,
|
| 276 |
+
"step": 27
|
| 277 |
+
},
|
| 278 |
+
{
|
| 279 |
+
"loss": 0.6972,
|
| 280 |
+
"learning_rate": 7.2e-05,
|
| 281 |
+
"epoch": 1.47,
|
| 282 |
+
"step": 28
|
| 283 |
+
},
|
| 284 |
+
{
|
| 285 |
+
"loss": 0.6813,
|
| 286 |
+
"learning_rate": 7.2e-05,
|
| 287 |
+
"epoch": 1.53,
|
| 288 |
+
"step": 29
|
| 289 |
+
},
|
| 290 |
+
{
|
| 291 |
+
"loss": 0.6828,
|
| 292 |
+
"learning_rate": 7.2e-05,
|
| 293 |
+
"epoch": 1.58,
|
| 294 |
+
"step": 30
|
| 295 |
+
},
|
| 296 |
+
{
|
| 297 |
+
"loss": 0.6826,
|
| 298 |
+
"learning_rate": 7.2e-05,
|
| 299 |
+
"epoch": 1.63,
|
| 300 |
+
"step": 31
|
| 301 |
+
},
|
| 302 |
+
{
|
| 303 |
+
"loss": 0.6938,
|
| 304 |
+
"learning_rate": 7.2e-05,
|
| 305 |
+
"epoch": 1.68,
|
| 306 |
+
"step": 32
|
| 307 |
+
},
|
| 308 |
+
{
|
| 309 |
+
"loss": 0.6955,
|
| 310 |
+
"learning_rate": 7.2e-05,
|
| 311 |
+
"epoch": 1.74,
|
| 312 |
+
"step": 33
|
| 313 |
+
},
|
| 314 |
+
{
|
| 315 |
+
"loss": 0.6808,
|
| 316 |
+
"learning_rate": 7.2e-05,
|
| 317 |
+
"epoch": 1.79,
|
| 318 |
+
"step": 34
|
| 319 |
+
},
|
| 320 |
+
{
|
| 321 |
+
"loss": 0.6886,
|
| 322 |
+
"learning_rate": 7.2e-05,
|
| 323 |
+
"epoch": 1.84,
|
| 324 |
+
"step": 35
|
| 325 |
+
},
|
| 326 |
+
{
|
| 327 |
+
"loss": 0.6998,
|
| 328 |
+
"learning_rate": 7.2e-05,
|
| 329 |
+
"epoch": 1.89,
|
| 330 |
+
"step": 36
|
| 331 |
+
},
|
| 332 |
+
{
|
| 333 |
+
"loss": 0.6769,
|
| 334 |
+
"learning_rate": 7.2e-05,
|
| 335 |
+
"epoch": 1.95,
|
| 336 |
+
"step": 37
|
| 337 |
+
},
|
| 338 |
+
{
|
| 339 |
+
"loss": 0.7146,
|
| 340 |
+
"learning_rate": 7.2e-05,
|
| 341 |
+
"epoch": 2.0,
|
| 342 |
+
"step": 38
|
| 343 |
+
},
|
| 344 |
+
{
|
| 345 |
+
"loss": 0.6642,
|
| 346 |
+
"learning_rate": 7.2e-05,
|
| 347 |
+
"epoch": 2.05,
|
| 348 |
+
"step": 39
|
| 349 |
+
},
|
| 350 |
+
|
| 351 |
+
|
| 352 |
+
"loss": 0.684,
|
| 353 |
+
"learning_rate": 7.2e-05,
|
| 354 |
+
"epoch": 2.11,
|
| 355 |
+
"step": 40
|
| 356 |
+
},
|
| 357 |
+
{
|
| 358 |
+
"loss": 0.6962,
|
| 359 |
+
"learning_rate": 7.2e-05,
|
| 360 |
+
"epoch": 2.16,
|
| 361 |
+
"step": 41
|
| 362 |
+
},
|
| 363 |
+
{
|
| 364 |
+
"loss": 0.7021,
|
| 365 |
+
"learning_rate": 7.2e-05,
|
| 366 |
+
"epoch": 2.21,
|
| 367 |
+
"step": 42
|
| 368 |
+
},
|
| 369 |
+
{
|
| 370 |
+
"loss": 0.702,
|
| 371 |
+
"learning_rate": 7.2e-05,
|
| 372 |
+
"epoch": 2.26,
|
| 373 |
+
"step": 43
|
| 374 |
+
},
|
| 375 |
+
{
|
| 376 |
+
"loss": 0.6701,
|
| 377 |
+
"learning_rate": 7.2e-05,
|
| 378 |
+
"epoch": 2.32,
|
| 379 |
+
"step": 44
|
| 380 |
+
},
|
| 381 |
+
{
|
| 382 |
+
"loss": 0.7221,
|
| 383 |
+
"learning_rate": 7.2e-05,
|
| 384 |
+
"epoch": 2.37,
|
| 385 |
+
"step": 45
|
| 386 |
+
},
|
| 387 |
+
{
|
| 388 |
+
"loss": 0.6895,
|
| 389 |
+
"learning_rate": 7.2e-05,
|
| 390 |
+
"epoch": 2.42,
|
| 391 |
+
"step": 46
|
| 392 |
+
},
|
| 393 |
+
{
|
| 394 |
+
"loss": 0.6897,
|
| 395 |
+
"learning_rate": 7.2e-05,
|
| 396 |
+
"epoch": 2.47,
|
| 397 |
+
"step": 47
|
| 398 |
+
},
|
| 399 |
+
{
|
| 400 |
+
"loss": 0.7001,
|
| 401 |
+
"learning_rate": 7.2e-05,
|
| 402 |
+
"epoch": 2.53,
|
| 403 |
+
"step": 48
|
| 404 |
+
},
|
| 405 |
+
{
|
| 406 |
+
"loss": 0.6845,
|
| 407 |
+
"learning_rate": 7.2e-05,
|
| 408 |
+
"epoch": 2.58,
|
| 409 |
+
"step": 49
|
| 410 |
+
},
|
| 411 |
+
{
|
| 412 |
+
"loss": 0.702,
|
| 413 |
+
"learning_rate": 7.2e-05,
|
| 414 |
+
"epoch": 2.63,
|
| 415 |
+
"step": 50
|
| 416 |
+
},
|
| 417 |
+
{
|
| 418 |
+
"eval_code_is_correct_loss": 0.6996689438819885,
|
| 419 |
+
"eval_code_is_correct_score": -0.25365304946899414,
|
| 420 |
+
"eval_code_is_correct_brier_score": 0.25365304946899414,
|
| 421 |
+
"eval_code_is_correct_average_probability": 0.4967172145843506,
|
| 422 |
+
"eval_code_is_correct_accuracy": 0.43,
|
| 423 |
+
"eval_code_is_correct_probabilities": [
|
| 424 |
+
0.5074375867843628,
|
| 425 |
+
0.4723518490791321,
|
| 426 |
+
0.48632118105888367,
|
| 427 |
+
0.49848517775535583,
|
| 428 |
+
0.476654589176178,
|
| 429 |
+
0.5107546448707581,
|
| 430 |
+
0.4704548716545105,
|
| 431 |
+
0.49364662170410156,
|
| 432 |
+
0.49723079800605774,
|
| 433 |
+
0.49965834617614746,
|
| 434 |
+
0.4738099277019501,
|
| 435 |
+
0.5174873471260071,
|
| 436 |
+
0.5036610960960388,
|
| 437 |
+
0.5156582593917847,
|
| 438 |
+
0.49093103408813477,
|
| 439 |
+
0.49312105774879456,
|
| 440 |
+
0.4914981424808502,
|
| 441 |
+
0.5185192823410034,
|
| 442 |
+
0.5077788233757019,
|
| 443 |
+
0.4805068373680115,
|
| 444 |
+
0.4961990416049957,
|
| 445 |
+
0.49203917384147644,
|
| 446 |
+
0.4657118618488312,
|
| 447 |
+
0.4971868097782135,
|
| 448 |
+
0.5209845304489136,
|
| 449 |
+
0.5242859721183777,
|
| 450 |
+
0.5050302147865295,
|
| 451 |
+
0.5064529776573181,
|
| 452 |
+
0.47357138991355896,
|
| 453 |
+
0.5292315483093262,
|
| 454 |
+
0.48264941573143005,
|
| 455 |
+
0.5002627968788147,
|
| 456 |
+
0.5115101933479309,
|
| 457 |
+
0.4869055151939392,
|
| 458 |
+
0.46029502153396606,
|
| 459 |
+
0.49682366847991943,
|
| 460 |
+
0.4798851013183594,
|
| 461 |
+
0.4962373971939087,
|
| 462 |
+
0.4617874026298523,
|
| 463 |
+
0.5052241683006287,
|
| 464 |
+
0.4935094118118286,
|
| 465 |
+
0.5108048319816589,
|
| 466 |
+
0.5389617085456848,
|
| 467 |
+
0.49815618991851807,
|
| 468 |
+
0.47433826327323914,
|
| 469 |
+
0.49846169352531433,
|
| 470 |
+
0.49559274315834045,
|
| 471 |
+
0.5151790976524353,
|
| 472 |
+
0.5131077170372009,
|
| 473 |
+
0.5054615139961243,
|
| 474 |
+
0.4999138414859772,
|
| 475 |
+
0.5005282163619995,
|
| 476 |
+
0.4700585603713989,
|
| 477 |
+
0.4792395234107971,
|
| 478 |
+
0.484203577041626,
|
| 479 |
+
0.5069689750671387,
|
| 480 |
+
0.5312584042549133,
|
| 481 |
+
0.48872625827789307,
|
| 482 |
+
0.4664309024810791,
|
| 483 |
+
0.4882381856441498,
|
| 484 |
+
0.49106600880622864,
|
| 485 |
+
0.5087692737579346,
|
| 486 |
+
0.4675412178039551,
|
| 487 |
+
0.573962390422821,
|
| 488 |
+
0.4947565197944641,
|
| 489 |
+
0.4820893108844757,
|
| 490 |
+
0.5046402215957642,
|
| 491 |
+
0.5246191620826721,
|
| 492 |
+
0.5146574378013611,
|
| 493 |
+
0.5106992125511169,
|
| 494 |
+
0.5022192597389221,
|
| 495 |
+
0.5226336717605591,
|
| 496 |
+
0.5060445070266724,
|
| 497 |
+
0.481097012758255,
|
| 498 |
+
0.5169151425361633,
|
| 499 |
+
0.4949752390384674,
|
| 500 |
+
0.4639112055301666,
|
| 501 |
+
0.4817794859409332,
|
| 502 |
+
0.48367029428482056,
|
| 503 |
+
0.4781152904033661,
|
| 504 |
+
0.47685888409614563,
|
| 505 |
+
0.48095741868019104,
|
| 506 |
+
0.491726815700531,
|
| 507 |
+
0.5017462968826294,
|
| 508 |
+
0.4904364347457886,
|
| 509 |
+
0.4698854982852936,
|
| 510 |
+
0.5083165168762207,
|
| 511 |
+
0.5017600655555725,
|
| 512 |
+
0.5108470916748047,
|
| 513 |
+
0.48250630497932434,
|
| 514 |
+
0.5158757567405701,
|
| 515 |
+
0.5019074082374573,
|
| 516 |
+
0.5176712274551392,
|
| 517 |
+
0.5154513716697693,
|
| 518 |
+
0.503238320350647,
|
| 519 |
+
0.5167521834373474,
|
| 520 |
+
0.4993845224380493,
|
| 521 |
+
0.4867921769618988,
|
| 522 |
+
0.47787371277809143,
|
| 523 |
+
0.46018683910369873
|
| 524 |
+
],
|
| 525 |
+
"eval_code_is_correct_runtime": 29.2407,
|
| 526 |
+
"eval_code_is_correct_samples_per_second": 3.42,
|
| 527 |
+
"eval_code_is_correct_steps_per_second": 0.068,
|
| 528 |
+
"epoch": 2.63,
|
| 529 |
+
"step": 50
|
| 530 |
+
},
|
| 531 |
+
{
|
| 532 |
+
"loss": 0.6957,
|
| 533 |
+
"learning_rate": 7.2e-05,
|
| 534 |
+
"epoch": 2.68,
|
| 535 |
+
"step": 51
|
| 536 |
+
},
|
| 537 |
+
{
|
| 538 |
+
"loss": 0.6726,
|
| 539 |
+
"learning_rate": 7.2e-05,
|
| 540 |
+
"epoch": 2.74,
|
| 541 |
+
"step": 52
|
| 542 |
+
},
|
| 543 |
+
{
|
| 544 |
+
"loss": 0.6883,
|
| 545 |
+
"learning_rate": 7.2e-05,
|
| 546 |
+
"epoch": 2.79,
|
| 547 |
+
"step": 53
|
| 548 |
+
},
|
| 549 |
+
{
|
| 550 |
+
"loss": 0.6894,
|
| 551 |
+
"learning_rate": 7.2e-05,
|
| 552 |
+
"epoch": 2.84,
|
| 553 |
+
"step": 54
|
| 554 |
+
},
|
| 555 |
+
{
|
| 556 |
+
"loss": 0.6743,
|
| 557 |
+
"learning_rate": 7.2e-05,
|
| 558 |
+
"epoch": 2.89,
|
| 559 |
+
"step": 55
|
| 560 |
+
},
|
| 561 |
+
{
|
| 562 |
+
"loss": 0.6833,
|
| 563 |
+
"learning_rate": 7.2e-05,
|
| 564 |
+
"epoch": 2.95,
|
| 565 |
+
"step": 56
|
| 566 |
+
},
|
| 567 |
+
{
|
| 568 |
+
"loss": 0.6786,
|
| 569 |
+
"learning_rate": 7.2e-05,
|
| 570 |
+
"epoch": 3.0,
|
| 571 |
+
"step": 57
|
| 572 |
+
},
|
| 573 |
+
{
|
| 574 |
+
"loss": 0.699,
|
| 575 |
+
"learning_rate": 7.2e-05,
|
| 576 |
+
"epoch": 3.05,
|
| 577 |
+
"step": 58
|
| 578 |
+
},
|
| 579 |
+
{
|
| 580 |
+
"loss": 0.6791,
|
| 581 |
+
"learning_rate": 7.2e-05,
|
| 582 |
+
"epoch": 3.11,
|
| 583 |
+
"step": 59
|
| 584 |
+
},
|
| 585 |
+
{
|
| 586 |
+
"loss": 0.6751,
|
| 587 |
+
"learning_rate": 7.2e-05,
|
| 588 |
+
"epoch": 3.16,
|
| 589 |
+
"step": 60
|
| 590 |
+
},
|
| 591 |
+
{
|
| 592 |
+
"loss": 0.6812,
|
| 593 |
+
"learning_rate": 7.2e-05,
|
| 594 |
+
"epoch": 3.21,
|
| 595 |
+
"step": 61
|
| 596 |
+
},
|
| 597 |
+
{
|
| 598 |
+
"loss": 0.6724,
|
| 599 |
+
"learning_rate": 7.2e-05,
|
| 600 |
+
"epoch": 3.26,
|
| 601 |
+
"step": 62
|
| 602 |
+
},
|
| 603 |
+
{
|
| 604 |
+
"loss": 0.6829,
|
| 605 |
+
"learning_rate": 7.2e-05,
|
| 606 |
+
"epoch": 3.32,
|
| 607 |
+
"step": 63
|
| 608 |
+
},
|
| 609 |
+
{
|
| 610 |
+
"loss": 0.6842,
|
| 611 |
+
"learning_rate": 7.2e-05,
|
| 612 |
+
"epoch": 3.37,
|
| 613 |
+
"step": 64
|
| 614 |
+
},
|
| 615 |
+
{
|
| 616 |
+
"loss": 0.6806,
|
| 617 |
+
"learning_rate": 7.2e-05,
|
| 618 |
+
"epoch": 3.42,
|
| 619 |
+
"step": 65
|
| 620 |
+
},
|
| 621 |
+
{
|
| 622 |
+
"loss": 0.6905,
|
| 623 |
+
"learning_rate": 7.2e-05,
|
| 624 |
+
"epoch": 3.47,
|
| 625 |
+
"step": 66
|
| 626 |
+
},
|
| 627 |
+
{
|
| 628 |
+
"loss": 0.6845,
|
| 629 |
+
"learning_rate": 7.2e-05,
|
| 630 |
+
"epoch": 3.53,
|
| 631 |
+
"step": 67
|
| 632 |
+
},
|
| 633 |
+
{
|
| 634 |
+
"loss": 0.6722,
|
| 635 |
+
"learning_rate": 7.2e-05,
|
| 636 |
+
"epoch": 3.58,
|
| 637 |
+
"step": 68
|
| 638 |
+
},
|
| 639 |
+
{
|
| 640 |
+
"loss": 0.6976,
|
| 641 |
+
"learning_rate": 7.2e-05,
|
| 642 |
+
"epoch": 3.63,
|
| 643 |
+
"step": 69
|
| 644 |
+
},
|
| 645 |
+
{
|
| 646 |
+
"loss": 0.6757,
|
| 647 |
+
"learning_rate": 7.2e-05,
|
| 648 |
+
"epoch": 3.68,
|
| 649 |
+
"step": 70
|
| 650 |
+
},
|
| 651 |
+
{
|
| 652 |
+
"loss": 0.6638,
|
| 653 |
+
"learning_rate": 7.2e-05,
|
| 654 |
+
"epoch": 3.74,
|
| 655 |
+
"step": 71
|
| 656 |
+
},
|
| 657 |
+
{
|
| 658 |
+
"loss": 0.6589,
|
| 659 |
+
"learning_rate": 7.2e-05,
|
| 660 |
+
"epoch": 3.79,
|
| 661 |
+
"step": 72
|
| 662 |
+
},
|
| 663 |
+
{
|
| 664 |
+
"loss": 0.6792,
|
| 665 |
+
"learning_rate": 7.2e-05,
|
| 666 |
+
"epoch": 3.84,
|
| 667 |
+
"step": 73
|
| 668 |
+
},
|
| 669 |
+
{
|
| 670 |
+
"loss": 0.6868,
|
| 671 |
+
"learning_rate": 7.2e-05,
|
| 672 |
+
"epoch": 3.89,
|
| 673 |
+
"step": 74
|
| 674 |
+
},
|
| 675 |
+
{
|
| 676 |
+
"loss": 0.6743,
|
| 677 |
+
"learning_rate": 7.2e-05,
|
| 678 |
+
"epoch": 3.95,
|
| 679 |
+
"step": 75
|
| 680 |
+
},
|
| 681 |
+
{
|
| 682 |
+
"eval_code_is_correct_loss": 0.7115283012390137,
|
| 683 |
+
"eval_code_is_correct_score": -0.25936323404312134,
|
| 684 |
+
"eval_code_is_correct_brier_score": 0.25936323404312134,
|
| 685 |
+
"eval_code_is_correct_average_probability": 0.492844820022583,
|
| 686 |
+
"eval_code_is_correct_accuracy": 0.45,
|
| 687 |
+
""eval_code_is_correct_probabilities": [
|
| 688 |
+
0.5315108895301819,
|
| 689 |
+
0.4404357671737671,
|
| 690 |
+
0.4393325448036194,
|
| 691 |
+
0.5136876702308655,
|
| 692 |
+
0.4317817986011505,
|
| 693 |
+
0.5407170057296753,
|
| 694 |
+
0.42210498452186584,
|
| 695 |
+
0.5032879114151001,
|
| 696 |
+
0.499985009431839,
|
| 697 |
+
0.5063377022743225,
|
| 698 |
+
0.44853413105010986,
|
| 699 |
+
0.5249595642089844,
|
| 700 |
+
0.519929051399231,
|
| 701 |
+
0.5444207191467285,
|
| 702 |
+
0.46772903203964233,
|
| 703 |
+
0.4709188640117645,
|
| 704 |
+
0.49579933285713196,
|
| 705 |
+
0.5452953577041626,
|
| 706 |
+
0.5071123838424683,
|
| 707 |
+
0.4668845534324646,
|
| 708 |
+
0.5050070285797119,
|
| 709 |
+
0.4854079782962799,
|
| 710 |
+
0.4207446873188019,
|
| 711 |
+
0.5136836171150208,
|
| 712 |
+
0.5283437371253967,
|
| 713 |
+
0.5321181416511536,
|
| 714 |
+
0.4974518120288849,
|
| 715 |
+
0.4980766177177429,
|
| 716 |
+
0.43917903304100037,
|
| 717 |
+
0.5816516876220703,
|
| 718 |
+
0.4618088901042938,
|
| 719 |
+
0.48619911074638367,
|
| 720 |
+
0.5236653089523315,
|
| 721 |
+
0.45910876989364624,
|
| 722 |
+
0.40605857968330383,
|
| 723 |
+
0.4992276430130005,
|
| 724 |
+
0.45045509934425354,
|
| 725 |
+
0.47411617636680603,
|
| 726 |
+
0.40248948335647583,
|
| 727 |
+
0.5305703282356262,
|
| 728 |
+
0.4813811182975769,
|
| 729 |
+
0.5272884368896484,
|
| 730 |
+
0.6161030530929565,
|
| 731 |
+
0.46347883343696594,
|
| 732 |
+
0.43354177474975586,
|
| 733 |
+
0.4877569079399109,
|
| 734 |
+
0.5148259997367859,
|
| 735 |
+
0.5279054045677185,
|
| 736 |
+
0.5294895172119141,
|
| 737 |
+
0.5120590329170227,
|
| 738 |
+
0.5139973163604736,
|
| 739 |
+
0.5520778298377991,
|
| 740 |
+
0.45969536900520325,
|
| 741 |
+
0.4304928183555603,
|
| 742 |
+
0.48464471101760864,
|
| 743 |
+
0.5018194913864136,
|
| 744 |
+
0.5791583061218262,
|
| 745 |
+
0.47863584756851196,
|
| 746 |
+
0.41838565468788147,
|
| 747 |
+
0.4741626977920532,
|
| 748 |
+
0.4651537537574768,
|
| 749 |
+
0.5022154450416565,
|
| 750 |
+
0.4011237621307373,
|
| 751 |
+
0.6317132115364075,
|
| 752 |
+
0.4900273084640503,
|
| 753 |
+
0.44324833154678345,
|
| 754 |
+
0.4855543375015259,
|
| 755 |
+
0.5688998103141785,
|
| 756 |
+
0.532179057598114,
|
| 757 |
+
0.5357102155685425,
|
| 758 |
+
0.48324429988861084,
|
| 759 |
+
0.5526444911956787,
|
| 760 |
+
0.539372444152832,
|
| 761 |
+
0.44041797518730164,
|
| 762 |
+
0.5359059572219849,
|
| 763 |
+
0.46845659613609314,
|
| 764 |
+
0.42640483379364014,
|
| 765 |
+
0.4501952528953552,
|
| 766 |
+
0.4579535722732544,
|
| 767 |
+
0.4656803011894226,
|
| 768 |
+
0.43683329224586487,
|
| 769 |
+
0.448000431060791,
|
| 770 |
+
0.482930064201355,
|
| 771 |
+
0.5224722623825073,
|
| 772 |
+
0.4616173803806305,
|
| 773 |
+
0.42225685715675354,
|
| 774 |
+
0.48842281103134155,
|
| 775 |
+
0.5094820261001587,
|
| 776 |
+
0.5359870195388794,
|
| 777 |
+
0.45143479108810425,
|
| 778 |
+
0.5495825409889221,
|
| 779 |
+
0.5185818672180176,
|
| 780 |
+
0.5359978675842285,
|
| 781 |
+
0.5442188382148743,
|
| 782 |
+
0.5404256582260132,
|
| 783 |
+
0.5551626682281494,
|
| 784 |
+
0.5226612091064453,
|
| 785 |
+
0.4831567108631134,
|
| 786 |
+
0.4785938560962677,
|
| 787 |
+
0.41752859950065613
|
| 788 |
+
],
|
| 789 |
+
"eval_code_is_correct_runtime": 29.2291,
|
| 790 |
+
"eval_code_is_correct_samples_per_second": 3.421,
|
| 791 |
+
"eval_code_is_correct_steps_per_second": 0.068,
|
| 792 |
+
"epoch": 3.95,
|
| 793 |
+
"step": 75
|
| 794 |
+
},
|
| 795 |
+
{
|
| 796 |
+
"loss": 0.6737,
|
| 797 |
+
"learning_rate": 7.2e-05,
|
| 798 |
+
"epoch": 4.0,
|
| 799 |
+
"step": 76
|
| 800 |
+
},
|
| 801 |
+
{
|
| 802 |
+
"loss": 0.6813,
|
| 803 |
+
"learning_rate": 7.2e-05,
|
| 804 |
+
"epoch": 4.05,
|
| 805 |
+
"step": 77
|
| 806 |
+
},
|
| 807 |
+
{
|
| 808 |
+
"loss": 0.6845,
|
| 809 |
+
"learning_rate": 7.2e-05,
|
| 810 |
+
"epoch": 4.11,
|
| 811 |
+
"step": 78
|
| 812 |
+
},
|
| 813 |
+
{
|
| 814 |
+
"loss": 0.68,
|
| 815 |
+
"learning_rate": 7.2e-05,
|
| 816 |
+
"epoch": 4.16,
|
| 817 |
+
"step": 79
|
| 818 |
+
},
|
| 819 |
+
{
|
| 820 |
+
"loss": 0.6638,
|
| 821 |
+
"learning_rate": 7.2e-05,
|
| 822 |
+
"epoch": 4.21,
|
| 823 |
+
"step": 80
|
| 824 |
+
},
|
| 825 |
+
{
|
| 826 |
+
"loss": 0.7058,
|
| 827 |
+
"learning_rate": 7.2e-05,
|
| 828 |
+
"epoch": 4.26,
|
| 829 |
+
"step": 81
|
| 830 |
+
},
|
| 831 |
+
{
|
| 832 |
+
"loss": 0.6644,
|
| 833 |
+
"learning_rate": 7.2e-05,
|
| 834 |
+
"epoch": 4.32,
|
| 835 |
+
"step": 82
|
| 836 |
+
},
|
| 837 |
+
{
|
| 838 |
+
"loss": 0.6641,
|
| 839 |
+
"learning_rate": 7.2e-05,
|
| 840 |
+
"epoch": 4.37,
|
| 841 |
+
"step": 83
|
| 842 |
+
},
|
| 843 |
+
{
|
| 844 |
+
"loss": 0.665,
|
| 845 |
+
"learning_rate": 7.2e-05,
|
| 846 |
+
"epoch": 4.42,
|
| 847 |
+
"step": 84
|
| 848 |
+
},
|
| 849 |
+
{
|
| 850 |
+
"loss": 0.6785,
|
| 851 |
+
"learning_rate": 7.2e-05,
|
| 852 |
+
"epoch": 4.47,
|
| 853 |
+
"step": 85
|
| 854 |
+
},
|
| 855 |
+
{
|
| 856 |
+
"loss": 0.6354,
|
| 857 |
+
"learning_rate": 7.2e-05,
|
| 858 |
+
"epoch": 4.53,
|
| 859 |
+
"step": 86
|
| 860 |
+
},
|
| 861 |
+
{
|
| 862 |
+
"loss": 0.6546,
|
| 863 |
+
"learning_rate": 7.2e-05,
|
| 864 |
+
"epoch": 4.58,
|
| 865 |
+
"step": 87
|
| 866 |
+
},
|
| 867 |
+
{
|
| 868 |
+
"loss": 0.6678,
|
| 869 |
+
"learning_rate": 7.2e-05,
|
| 870 |
+
"epoch": 4.63,
|
| 871 |
+
"step": 88
|
| 872 |
+
},
|
| 873 |
+
{
|
| 874 |
+
"loss": 0.6342,
|
| 875 |
+
"learning_rate": 7.2e-05,
|
| 876 |
+
"epoch": 4.68,
|
| 877 |
+
"step": 89
|
| 878 |
+
},
|
| 879 |
+
{
|
| 880 |
+
"loss": 0.6919,
|
| 881 |
+
"learning_rate": 7.2e-05,
|
| 882 |
+
"epoch": 4.74,
|
| 883 |
+
"step": 90
|
| 884 |
+
},
|
| 885 |
+
{
|
| 886 |
+
"loss": 0.6494,
|
| 887 |
+
"learning_rate": 7.2e-05,
|
| 888 |
+
"epoch": 4.79,
|
| 889 |
+
"step": 91
|
| 890 |
+
},
|
| 891 |
+
{
|
| 892 |
+
"loss": 0.6597,
|
| 893 |
+
"learning_rate": 7.2e-05,
|
| 894 |
+
"epoch": 4.84,
|
| 895 |
+
"step": 92
|
| 896 |
+
},
|
| 897 |
+
{
|
| 898 |
+
"loss": 0.6769,
|
| 899 |
+
"learning_rate": 7.2e-05,
|
| 900 |
+
"epoch": 4.89,
|
| 901 |
+
"step": 93
|
| 902 |
+
},
|
| 903 |
+
{
|
| 904 |
+
"loss": 0.6669,
|
| 905 |
+
"learning_rate": 7.2e-05,
|
| 906 |
+
"epoch": 4.95,
|
| 907 |
+
"step": 94
|
| 908 |
+
},
|
| 909 |
+
{
|
| 910 |
+
"loss": 0.6585,
|
| 911 |
+
"learning_rate": 7.2e-05,
|
| 912 |
+
"epoch": 5.0,
|
| 913 |
+
"step": 95
|
| 914 |
+
},
|
| 915 |
+
{
|
| 916 |
+
"loss": 0.6861,
|
| 917 |
+
"learning_rate": 7.2e-05,
|
| 918 |
+
"epoch": 5.05,
|
| 919 |
+
"step": 96
|
| 920 |
+
},
|
| 921 |
+
{
|
| 922 |
+
"loss": 0.6498,
|
| 923 |
+
"learning_rate": 7.2e-05,
|
| 924 |
+
"epoch": 5.11,
|
| 925 |
+
"step": 97
|
| 926 |
+
},
|
| 927 |
+
{
|
| 928 |
+
"loss": 0.6466,
|
| 929 |
+
"learning_rate": 7.2e-05,
|
| 930 |
+
"epoch": 5.16,
|
| 931 |
+
"step": 98
|
| 932 |
+
},
|
| 933 |
+
{
|
| 934 |
+
"loss": 0.6421,
|
| 935 |
+
"learning_rate": 7.2e-05,
|
| 936 |
+
"epoch": 5.21,
|
| 937 |
+
"step": 99
|
| 938 |
+
},
|
| 939 |
+
{
|
| 940 |
+
"loss": 0.6568,
|
| 941 |
+
"learning_rate": 7.2e-05,
|
| 942 |
+
"epoch": 5.26,
|
| 943 |
+
"step": 100
|
| 944 |
+
},
|
| 945 |
+
{
|
| 946 |
+
"eval_code_is_correct_loss": 0.7250129580497742,
|
| 947 |
+
"eval_code_is_correct_score": -0.2655547261238098,
|
| 948 |
+
"eval_code_is_correct_brier_score": 0.2655547261238098,
|
| 949 |
+
"eval_code_is_correct_average_probability": 0.49155572056770325,
|
| 950 |
+
"eval_code_is_correct_accuracy": 0.48,
|
| 951 |
+
"eval_code_is_correct_probabilities": [
|
| 952 |
+
0.5083851218223572,
|
| 953 |
+
0.4273935854434967,
|
| 954 |
+
0.4364035129547119,
|
| 955 |
+
0.4872526526451111,
|
| 956 |
+
0.37440311908721924,
|
| 957 |
+
0.5888470411300659,
|
| 958 |
+
0.37513232231140137,
|
| 959 |
+
0.48127123713493347,
|
| 960 |
+
0.47687897086143494,
|
| 961 |
+
0.4585302472114563,
|
| 962 |
+
0.4116853177547455,
|
| 963 |
+
0.5032776594161987,
|
| 964 |
+
0.5315262675285339,
|
| 965 |
+
0.5401375889778137,
|
| 966 |
+
0.47433847188949585,
|
| 967 |
+
0.48878082633018494,
|
| 968 |
+
0.5708949565887451,
|
| 969 |
+
0.5461510419845581,
|
| 970 |
+
0.5810272097587585,
|
| 971 |
+
0.4907934367656708,
|
| 972 |
+
0.4523153305053711,
|
| 973 |
+
0.5213923454284668,
|
| 974 |
+
0.40194064378738403,
|
| 975 |
+
0.4633757770061493,
|
| 976 |
+
0.4978197515010834,
|
| 977 |
+
0.5100975632667542,
|
| 978 |
+
0.5173598527908325,
|
| 979 |
+
0.5476617813110352,
|
| 980 |
+
0.45146241784095764,
|
| 981 |
+
0.6570491194725037,
|
| 982 |
+
0.453940212726593,
|
| 983 |
+
0.5251180529594421,
|
| 984 |
+
0.48246118426322937,
|
| 985 |
+
0.4712594449520111,
|
| 986 |
+
0.3543213903903961,
|
| 987 |
+
0.4387568533420563,
|
| 988 |
+
0.45936518907546997,
|
| 989 |
+
0.48073822259902954,
|
| 990 |
+
0.3076059818267822,
|
| 991 |
+
0.5172088742256165,
|
| 992 |
+
0.5109153985977173,
|
| 993 |
+
0.49946558475494385,
|
| 994 |
+
0.7792166471481323,
|
| 995 |
+
0.43944960832595825,
|
| 996 |
+
0.42077937722206116,
|
| 997 |
+
0.5631754398345947,
|
| 998 |
+
0.493571937084198,
|
| 999 |
+
0.49611571431159973,
|
| 1000 |
+
0.5187433362007141,
|
| 1001 |
+
0.5067903995513916,
|
| 1002 |
+
0.48328694701194763,
|
| 1003 |
+
0.6678159236907959,
|
| 1004 |
+
0.3781387507915497,
|
| 1005 |
+
0.3635239005088806,
|
| 1006 |
+
0.5364617109298706,
|
| 1007 |
+
0.5524997115135193,
|
| 1008 |
+
0.632897675037384,
|
| 1009 |
+
0.511653482913971,
|
| 1010 |
+
0.3425663709640503,
|
| 1011 |
+
0.5403125882148743,
|
| 1012 |
+
0.47932958602905273,
|
| 1013 |
+
0.5451420545578003,
|
| 1014 |
+
0.32550284266471863,
|
| 1015 |
+
0.7307556867599487,
|
| 1016 |
+
0.5360798835754395,
|
| 1017 |
+
0.36980098485946655,
|
| 1018 |
+
0.5202561616897583,
|
| 1019 |
+
0.6240019798278809,
|
| 1020 |
+
0.5508140921592712,
|
| 1021 |
+
0.5679302215576172,
|
| 1022 |
+
0.5044476389884949,
|
| 1023 |
+
0.5724747776985168,
|
| 1024 |
+
0.5710216760635376,
|
| 1025 |
+
0.3993741571903229,
|
| 1026 |
+
0.5445007681846619,
|
| 1027 |
+
0.4687022268772125,
|
| 1028 |
+
0.38320469856262207,
|
| 1029 |
+
0.4335331320762634,
|
| 1030 |
+
0.44064751267433167,
|
| 1031 |
+
0.3625722825527191,
|
| 1032 |
+
0.42068955302238464,
|
| 1033 |
+
0.3999265134334564,
|
| 1034 |
+
0.38937661051750183,
|
| 1035 |
+
0.5059962272644043,
|
| 1036 |
+
0.48507365584373474,
|
| 1037 |
+
0.3598549962043762,
|
| 1038 |
+
0.5178597569465637,
|
| 1039 |
+
0.518267810344696,
|
| 1040 |
+
0.5441962480545044,
|
| 1041 |
+
0.4130031168460846,
|
| 1042 |
+
0.613496720790863,
|
| 1043 |
+
0.45683199167251587,
|
| 1044 |
+
0.534511148929596,
|
| 1045 |
+
0.5284443497657776,
|
| 1046 |
+
0.6089046001434326,
|
| 1047 |
+
0.5780519843101501,
|
| 1048 |
+
0.544437050819397,
|
| 1049 |
+
0.41231974959373474,
|
| 1050 |
+
0.5470799207687378,
|
| 1051 |
+
0.34541189670562744
|
| 1052 |
+
],
|
| 1053 |
+
"eval_code_is_correct_runtime": 29.2342,
|
| 1054 |
+
"eval_code_is_correct_samples_per_second": 3.421,
|
| 1055 |
+
"eval_code_is_correct_steps_per_second": 0.068,
|
| 1056 |
+
"epoch": 5.26,
|
| 1057 |
+
"step": 100
|
| 1058 |
+
},
|
| 1059 |
+
{
|
| 1060 |
+
"train_runtime": 1903.4626,
|
| 1061 |
+
"train_samples_per_second": 1.681,
|
| 1062 |
+
"train_steps_per_second": 0.053,
|
| 1063 |
+
"total_flos": 0.0,
|
| 1064 |
+
"train_loss": 0.6841741448640823,
|
| 1065 |
+
"epoch": 5.26,
|
| 1066 |
+
"step": 100
|
| 1067 |
+
}
|
| 1068 |
+
]
|