Commit ·
d274e8a
1
Parent(s): c48d48e
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 +32 -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/llama-7b",
|
| 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": 0.0,
|
| 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 sha256:89c5b426d6361250f2d04c0d2b7f9ec07524efab213aaf70ece46434f0e2dfab
|
| 3 |
+
size 134295057
|
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 sha256:9e556afd44213b6bd1be2b850ebbbd98f5481437a8021afaf58ee7fb1818d347
|
| 3 |
+
size 499723
|
tokenizer_config.json
ADDED
|
@@ -0,0 +1,32 @@
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
| 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 |
+
"model_max_length": 2048,
|
| 20 |
+
"pad_token": null,
|
| 21 |
+
"sp_model_kwargs": {},
|
| 22 |
+
"tokenizer_class": "LlamaTokenizer",
|
| 23 |
+
"unk_token": {
|
| 24 |
+
"__type": "AddedToken",
|
| 25 |
+
"content": "<unk>",
|
| 26 |
+
"lstrip": false,
|
| 27 |
+
"normalized": true,
|
| 28 |
+
"rstrip": false,
|
| 29 |
+
"single_word": false
|
| 30 |
+
},
|
| 31 |
+
"use_default_system_prompt": true
|
| 32 |
+
}
|
train_args.json
ADDED
|
@@ -0,0 +1,123 @@
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
| 1 |
+
{
|
| 2 |
+
"initial_model_dir": "models/llama-7b",
|
| 3 |
+
"distribution_id": "code_is_correct",
|
| 4 |
+
"date_trained": "12/10/2023 05:08:54",
|
| 5 |
+
"output_dir": "models/classify_lora/llama-7b-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": 0.0002,
|
| 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/llama-7b-code_is_correct/runs/Oct12_04-25-08_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-llama-7b-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:d34e69cae4fa459759f9857e7ea9a59d84263df3a1a9e460150262fc2fc71f11
|
| 3 |
+
size 5115
|
training_logs.json
ADDED
|
@@ -0,0 +1,1067 @@
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
| 1 |
+
[
|
| 2 |
+
{
|
| 3 |
+
"loss": 2.319,
|
| 4 |
+
"learning_rate": 0.0002,
|
| 5 |
+
"epoch": 0.05,
|
| 6 |
+
"step": 1
|
| 7 |
+
},
|
| 8 |
+
{
|
| 9 |
+
"loss": 1.1571,
|
| 10 |
+
"learning_rate": 0.0002,
|
| 11 |
+
"epoch": 0.11,
|
| 12 |
+
"step": 2
|
| 13 |
+
},
|
| 14 |
+
{
|
| 15 |
+
"loss": 0.8951,
|
| 16 |
+
"learning_rate": 0.0002,
|
| 17 |
+
"epoch": 0.16,
|
| 18 |
+
"step": 3
|
| 19 |
+
},
|
| 20 |
+
{
|
| 21 |
+
"loss": 0.7019,
|
| 22 |
+
"learning_rate": 0.0002,
|
| 23 |
+
"epoch": 0.21,
|
| 24 |
+
"step": 4
|
| 25 |
+
},
|
| 26 |
+
{
|
| 27 |
+
"loss": 0.6999,
|
| 28 |
+
"learning_rate": 0.0002,
|
| 29 |
+
"epoch": 0.26,
|
| 30 |
+
"step": 5
|
| 31 |
+
},
|
| 32 |
+
{
|
| 33 |
+
"loss": 0.7884,
|
| 34 |
+
"learning_rate": 0.0002,
|
| 35 |
+
"epoch": 0.32,
|
| 36 |
+
"step": 6
|
| 37 |
+
},
|
| 38 |
+
{
|
| 39 |
+
"loss": 0.6625,
|
| 40 |
+
"learning_rate": 0.0002,
|
| 41 |
+
"epoch": 0.37,
|
| 42 |
+
"step": 7
|
| 43 |
+
},
|
| 44 |
+
{
|
| 45 |
+
"loss": 0.6943,
|
| 46 |
+
"learning_rate": 0.0002,
|
| 47 |
+
"epoch": 0.42,
|
| 48 |
+
"step": 8
|
| 49 |
+
},
|
| 50 |
+
{
|
| 51 |
+
"loss": 0.7418,
|
| 52 |
+
"learning_rate": 0.0002,
|
| 53 |
+
"epoch": 0.47,
|
| 54 |
+
"step": 9
|
| 55 |
+
},
|
| 56 |
+
{
|
| 57 |
+
"loss": 0.6976,
|
| 58 |
+
"learning_rate": 0.0002,
|
| 59 |
+
"epoch": 0.53,
|
| 60 |
+
"step": 10
|
| 61 |
+
},
|
| 62 |
+
{
|
| 63 |
+
"loss": 0.7081,
|
| 64 |
+
"learning_rate": 0.0002,
|
| 65 |
+
"epoch": 0.58,
|
| 66 |
+
"step": 11
|
| 67 |
+
},
|
| 68 |
+
{
|
| 69 |
+
"loss": 0.6817,
|
| 70 |
+
"learning_rate": 0.0002,
|
| 71 |
+
"epoch": 0.63,
|
| 72 |
+
"step": 12
|
| 73 |
+
},
|
| 74 |
+
{
|
| 75 |
+
"loss": 0.6799,
|
| 76 |
+
"learning_rate": 0.0002,
|
| 77 |
+
"epoch": 0.68,
|
| 78 |
+
"step": 13
|
| 79 |
+
},
|
| 80 |
+
{
|
| 81 |
+
"loss": 0.7014,
|
| 82 |
+
"learning_rate": 0.0002,
|
| 83 |
+
"epoch": 0.74,
|
| 84 |
+
"step": 14
|
| 85 |
+
},
|
| 86 |
+
{
|
| 87 |
+
"loss": 0.7207,
|
| 88 |
+
"learning_rate": 0.0002,
|
| 89 |
+
"epoch": 0.79,
|
| 90 |
+
"step": 15
|
| 91 |
+
},
|
| 92 |
+
{
|
| 93 |
+
"loss": 0.681,
|
| 94 |
+
"learning_rate": 0.0002,
|
| 95 |
+
"epoch": 0.84,
|
| 96 |
+
"step": 16
|
| 97 |
+
},
|
| 98 |
+
{
|
| 99 |
+
"loss": 0.7092,
|
| 100 |
+
"learning_rate": 0.0002,
|
| 101 |
+
"epoch": 0.89,
|
| 102 |
+
"step": 17
|
| 103 |
+
},
|
| 104 |
+
{
|
| 105 |
+
"loss": 0.691,
|
| 106 |
+
"learning_rate": 0.0002,
|
| 107 |
+
"epoch": 0.95,
|
| 108 |
+
"step": 18
|
| 109 |
+
},
|
| 110 |
+
{
|
| 111 |
+
"loss": 0.6796,
|
| 112 |
+
"learning_rate": 0.0002,
|
| 113 |
+
"epoch": 1.0,
|
| 114 |
+
"step": 19
|
| 115 |
+
},
|
| 116 |
+
{
|
| 117 |
+
"loss": 0.7201,
|
| 118 |
+
"learning_rate": 0.0002,
|
| 119 |
+
"epoch": 1.05,
|
| 120 |
+
"step": 20
|
| 121 |
+
},
|
| 122 |
+
{
|
| 123 |
+
"loss": 0.6609,
|
| 124 |
+
"learning_rate": 0.0002,
|
| 125 |
+
"epoch": 1.11,
|
| 126 |
+
"step": 21
|
| 127 |
+
},
|
| 128 |
+
{
|
| 129 |
+
"loss": 0.716,
|
| 130 |
+
"learning_rate": 0.0002,
|
| 131 |
+
"epoch": 1.16,
|
| 132 |
+
"step": 22
|
| 133 |
+
},
|
| 134 |
+
{
|
| 135 |
+
"loss": 0.7015,
|
| 136 |
+
"learning_rate": 0.0002,
|
| 137 |
+
"epoch": 1.21,
|
| 138 |
+
"step": 23
|
| 139 |
+
},
|
| 140 |
+
{
|
| 141 |
+
"loss": 0.7698,
|
| 142 |
+
"learning_rate": 0.0002,
|
| 143 |
+
"epoch": 1.26,
|
| 144 |
+
"step": 24
|
| 145 |
+
},
|
| 146 |
+
{
|
| 147 |
+
"loss": 0.7193,
|
| 148 |
+
"learning_rate": 0.0002,
|
| 149 |
+
"epoch": 1.32,
|
| 150 |
+
"step": 25
|
| 151 |
+
},
|
| 152 |
+
{
|
| 153 |
+
"eval_code_is_correct_loss": 0.7040590047836304,
|
| 154 |
+
"eval_code_is_correct_score": -0.25534066557884216,
|
| 155 |
+
"eval_code_is_correct_brier_score": 0.25534066557884216,
|
| 156 |
+
"eval_code_is_correct_average_probability": 0.4964892268180847,
|
| 157 |
+
"eval_code_is_correct_accuracy": 0.48,
|
| 158 |
+
"eval_code_is_correct_probabilities": [
|
| 159 |
+
0.537757158279419,
|
| 160 |
+
0.45456114411354065,
|
| 161 |
+
0.46773606538772583,
|
| 162 |
+
0.5088943839073181,
|
| 163 |
+
0.44396457076072693,
|
| 164 |
+
0.5441208481788635,
|
| 165 |
+
0.46894216537475586,
|
| 166 |
+
0.5342907905578613,
|
| 167 |
+
0.5333516001701355,
|
| 168 |
+
0.543117105960846,
|
| 169 |
+
0.4494006931781769,
|
| 170 |
+
0.5351378321647644,
|
| 171 |
+
0.5340185761451721,
|
| 172 |
+
0.5496283173561096,
|
| 173 |
+
0.46146854758262634,
|
| 174 |
+
0.46118637919425964,
|
| 175 |
+
0.47490179538726807,
|
| 176 |
+
0.54707932472229,
|
| 177 |
+
0.4630311131477356,
|
| 178 |
+
0.46271002292633057,
|
| 179 |
+
0.553754448890686,
|
| 180 |
+
0.45833271741867065,
|
| 181 |
+
0.44106921553611755,
|
| 182 |
+
0.5405190587043762,
|
| 183 |
+
0.5340151786804199,
|
| 184 |
+
0.538471519947052,
|
| 185 |
+
0.44194644689559937,
|
| 186 |
+
0.4529338479042053,
|
| 187 |
+
0.46126678586006165,
|
| 188 |
+
0.547139585018158,
|
| 189 |
+
0.4498041272163391,
|
| 190 |
+
0.4580332040786743,
|
| 191 |
+
0.5330279469490051,
|
| 192 |
+
0.47573158144950867,
|
| 193 |
+
0.4609244763851166,
|
| 194 |
+
0.5373966693878174,
|
| 195 |
+
0.4560283422470093,
|
| 196 |
+
0.45565375685691833,
|
| 197 |
+
0.4430266320705414,
|
| 198 |
+
0.5496351718902588,
|
| 199 |
+
0.46626290678977966,
|
| 200 |
+
0.5276714563369751,
|
| 201 |
+
0.5368297696113586,
|
| 202 |
+
0.4585987329483032,
|
| 203 |
+
0.4506687819957733,
|
| 204 |
+
0.48509010672569275,
|
| 205 |
+
0.5378000736236572,
|
| 206 |
+
0.5482139587402344,
|
| 207 |
+
0.53798508644104,
|
| 208 |
+
0.5404862761497498,
|
| 209 |
+
0.5462475419044495,
|
| 210 |
+
0.5316811203956604,
|
| 211 |
+
0.545008659362793,
|
| 212 |
+
0.4452662467956543,
|
| 213 |
+
0.46562960743904114,
|
| 214 |
+
0.4650121331214905,
|
| 215 |
+
0.5360617637634277,
|
| 216 |
+
0.4509516656398773,
|
| 217 |
+
0.44952070713043213,
|
| 218 |
+
0.4582957625389099,
|
| 219 |
+
0.4517310857772827,
|
| 220 |
+
0.4437441825866699,
|
| 221 |
+
0.4606005549430847,
|
| 222 |
+
0.547730028629303,
|
| 223 |
+
0.4598681926727295,
|
| 224 |
+
0.46059927344322205,
|
| 225 |
+
0.4745525121688843,
|
| 226 |
+
0.5424997210502625,
|
| 227 |
+
0.5405730605125427,
|
| 228 |
+
0.5450511574745178,
|
| 229 |
+
0.4547448456287384,
|
| 230 |
+
0.5489879846572876,
|
| 231 |
+
0.5354413986206055,
|
| 232 |
+
0.4355437159538269,
|
| 233 |
+
0.5529230833053589,
|
| 234 |
+
0.44915857911109924,
|
| 235 |
+
0.44940751791000366,
|
| 236 |
+
0.4544369876384735,
|
| 237 |
+
0.45526307821273804,
|
| 238 |
+
0.5274561047554016,
|
| 239 |
+
0.44203782081604004,
|
| 240 |
+
0.4448660910129547,
|
| 241 |
+
0.520516574382782,
|
| 242 |
+
0.5427135825157166,
|
| 243 |
+
0.4694376587867737,
|
| 244 |
+
0.45909956097602844,
|
| 245 |
+
0.46876445412635803,
|
| 246 |
+
0.5469857454299927,
|
| 247 |
+
0.5509116053581238,
|
| 248 |
+
0.4485166072845459,
|
| 249 |
+
0.5540351867675781,
|
| 250 |
+
0.5366657972335815,
|
| 251 |
+
0.5381718277931213,
|
| 252 |
+
0.539725661277771,
|
| 253 |
+
0.5412744879722595,
|
| 254 |
+
0.5397677421569824,
|
| 255 |
+
0.5319379568099976,
|
| 256 |
+
0.5362846255302429,
|
| 257 |
+
0.45564961433410645,
|
| 258 |
+
0.44395512342453003
|
| 259 |
+
],
|
| 260 |
+
"eval_code_is_correct_runtime": 38.6775,
|
| 261 |
+
"eval_code_is_correct_samples_per_second": 2.585,
|
| 262 |
+
"eval_code_is_correct_steps_per_second": 0.052,
|
| 263 |
+
"epoch": 1.32,
|
| 264 |
+
"step": 25
|
| 265 |
+
},
|
| 266 |
+
{
|
| 267 |
+
"loss": 0.6893,
|
| 268 |
+
"learning_rate": 0.0002,
|
| 269 |
+
"epoch": 1.37,
|
| 270 |
+
"step": 26
|
| 271 |
+
},
|
| 272 |
+
{
|
| 273 |
+
"loss": 0.6885,
|
| 274 |
+
"learning_rate": 0.0002,
|
| 275 |
+
"epoch": 1.42,
|
| 276 |
+
"step": 27
|
| 277 |
+
},
|
| 278 |
+
{
|
| 279 |
+
"loss": 0.6805,
|
| 280 |
+
"learning_rate": 0.0002,
|
| 281 |
+
"epoch": 1.47,
|
| 282 |
+
"step": 28
|
| 283 |
+
},
|
| 284 |
+
{
|
| 285 |
+
"loss": 0.7027,
|
| 286 |
+
"learning_rate": 0.0002,
|
| 287 |
+
"epoch": 1.53,
|
| 288 |
+
"step": 29
|
| 289 |
+
},
|
| 290 |
+
{
|
| 291 |
+
"loss": 0.7283,
|
| 292 |
+
"learning_rate": 0.0002,
|
| 293 |
+
"epoch": 1.58,
|
| 294 |
+
"step": 30
|
| 295 |
+
},
|
| 296 |
+
{
|
| 297 |
+
"loss": 0.6708,
|
| 298 |
+
"learning_rate": 0.0002,
|
| 299 |
+
"epoch": 1.63,
|
| 300 |
+
"step": 31
|
| 301 |
+
},
|
| 302 |
+
{
|
| 303 |
+
"loss": 0.7083,
|
| 304 |
+
"learning_rate": 0.0002,
|
| 305 |
+
"epoch": 1.68,
|
| 306 |
+
"step": 32
|
| 307 |
+
},
|
| 308 |
+
{
|
| 309 |
+
"loss": 0.7336,
|
| 310 |
+
"learning_rate": 0.0002,
|
| 311 |
+
"epoch": 1.74,
|
| 312 |
+
"step": 33
|
| 313 |
+
},
|
| 314 |
+
{
|
| 315 |
+
"loss": 0.7028,
|
| 316 |
+
"learning_rate": 0.0002,
|
| 317 |
+
"epoch": 1.79,
|
| 318 |
+
"step": 34
|
| 319 |
+
},
|
| 320 |
+
{
|
| 321 |
+
"loss": 0.695,
|
| 322 |
+
"learning_rate": 0.0002,
|
| 323 |
+
"epoch": 1.84,
|
| 324 |
+
"step": 35
|
| 325 |
+
},
|
| 326 |
+
{
|
| 327 |
+
"loss": 0.6997,
|
| 328 |
+
"learning_rate": 0.0002,
|
| 329 |
+
"epoch": 1.89,
|
| 330 |
+
"step": 36
|
| 331 |
+
},
|
| 332 |
+
{
|
| 333 |
+
"loss": 0.6743,
|
| 334 |
+
"learning_rate": 0.0002,
|
| 335 |
+
"epoch": 1.95,
|
| 336 |
+
"step": 37
|
| 337 |
+
},
|
| 338 |
+
{
|
| 339 |
+
"loss": 0.7388,
|
| 340 |
+
"learning_rate": 0.0002,
|
| 341 |
+
"epoch": 2.0,
|
| 342 |
+
"step": 38
|
| 343 |
+
},
|
| 344 |
+
{
|
| 345 |
+
"loss": 0.6464,
|
| 346 |
+
"learning_rate": 0.0002,
|
| 347 |
+
"epoch": 2.05,
|
| 348 |
+
"step": 39
|
| 349 |
+
},
|
| 350 |
+
{
|
| 351 |
+
"loss": 0.6772,
|
| 352 |
+
"learning_rate": 0.0002,
|
| 353 |
+
"epoch": 2.11,
|
| 354 |
+
"step": 40
|
| 355 |
+
},
|
| 356 |
+
{
|
| 357 |
+
"loss": 0.7009,
|
| 358 |
+
"learning_rate": 0.0002,
|
| 359 |
+
"epoch": 2.16,
|
| 360 |
+
"step": 41
|
| 361 |
+
},
|
| 362 |
+
{
|
| 363 |
+
"loss": 0.7021,
|
| 364 |
+
"learning_rate": 0.0002,
|
| 365 |
+
"epoch": 2.21,
|
| 366 |
+
"step": 42
|
| 367 |
+
},
|
| 368 |
+
{
|
| 369 |
+
"loss": 0.7051,
|
| 370 |
+
"learning_rate": 0.0002,
|
| 371 |
+
"epoch": 2.26,
|
| 372 |
+
"step": 43
|
| 373 |
+
},
|
| 374 |
+
{
|
| 375 |
+
"loss": 0.6816,
|
| 376 |
+
"learning_rate": 0.0002,
|
| 377 |
+
"epoch": 2.32,
|
| 378 |
+
"step": 44
|
| 379 |
+
},
|
| 380 |
+
{
|
| 381 |
+
"loss": 0.7047,
|
| 382 |
+
"learning_rate": 0.0002,
|
| 383 |
+
"epoch": 2.37,
|
| 384 |
+
"step": 45
|
| 385 |
+
},
|
| 386 |
+
{
|
| 387 |
+
"loss": 0.683,
|
| 388 |
+
"learning_rate": 0.0002,
|
| 389 |
+
"epoch": 2.42,
|
| 390 |
+
"step": 46
|
| 391 |
+
},
|
| 392 |
+
{
|
| 393 |
+
"loss": 0.6875,
|
| 394 |
+
"learning_rate": 0.0002,
|
| 395 |
+
"epoch": 2.47,
|
| 396 |
+
"step": 47
|
| 397 |
+
},
|
| 398 |
+
{
|
| 399 |
+
"loss": 0.6723,
|
| 400 |
+
"learning_rate": 0.0002,
|
| 401 |
+
"epoch": 2.53,
|
| 402 |
+
"step": 48
|
| 403 |
+
},
|
| 404 |
+
{
|
| 405 |
+
"loss": 0.6727,
|
| 406 |
+
"learning_rate": 0.0002,
|
| 407 |
+
"epoch": 2.58,
|
| 408 |
+
"step": 49
|
| 409 |
+
},
|
| 410 |
+
{
|
| 411 |
+
"loss": 0.6683,
|
| 412 |
+
"learning_rate": 0.0002,
|
| 413 |
+
"epoch": 2.63,
|
| 414 |
+
"step": 50
|
| 415 |
+
},
|
| 416 |
+
{
|
| 417 |
+
"eval_code_is_correct_loss": 0.6990968585014343,
|
| 418 |
+
"eval_code_is_correct_score": -0.2527521252632141,
|
| 419 |
+
"eval_code_is_correct_brier_score": 0.2527521252632141,
|
| 420 |
+
"eval_code_is_correct_average_probability": 0.5049085021018982,
|
| 421 |
+
"eval_code_is_correct_accuracy": 0.52,
|
| 422 |
+
"eval_code_is_correct_probabilities": [
|
| 423 |
+
0.4113783538341522,
|
| 424 |
+
0.5813317894935608,
|
| 425 |
+
0.5584416389465332,
|
| 426 |
+
0.41934269666671753,
|
| 427 |
+
0.5603463053703308,
|
| 428 |
+
0.4234032928943634,
|
| 429 |
+
0.5691850781440735,
|
| 430 |
+
0.45391106605529785,
|
| 431 |
+
0.41783419251441956,
|
| 432 |
+
0.3891553580760956,
|
| 433 |
+
0.5918279886245728,
|
| 434 |
+
0.39812296628952026,
|
| 435 |
+
0.3982398509979248,
|
| 436 |
+
0.4082736074924469,
|
| 437 |
+
0.5677108764648438,
|
| 438 |
+
0.6011993885040283,
|
| 439 |
+
0.5879720449447632,
|
| 440 |
+
0.3799493908882141,
|
| 441 |
+
0.6072862148284912,
|
| 442 |
+
0.5974151492118835,
|
| 443 |
+
0.41310346126556396,
|
| 444 |
+
0.563809871673584,
|
| 445 |
+
0.5413035750389099,
|
| 446 |
+
0.4107910394668579,
|
| 447 |
+
0.39933568239212036,
|
| 448 |
+
0.4007815718650818,
|
| 449 |
+
0.5928361415863037,
|
| 450 |
+
0.5920814871788025,
|
| 451 |
+
0.5649393200874329,
|
| 452 |
+
0.4397546648979187,
|
| 453 |
+
0.5982930660247803,
|
| 454 |
+
0.5736469030380249,
|
| 455 |
+
0.38247016072273254,
|
| 456 |
+
0.6029582023620605,
|
| 457 |
+
0.6050556302070618,
|
| 458 |
+
0.4481448829174042,
|
| 459 |
+
0.5958505272865295,
|
| 460 |
+
0.5872756838798523,
|
| 461 |
+
0.5685001015663147,
|
| 462 |
+
0.4170364737510681,
|
| 463 |
+
0.5979400873184204,
|
| 464 |
+
0.41077008843421936,
|
| 465 |
+
0.4539501667022705,
|
| 466 |
+
0.5808403491973877,
|
| 467 |
+
0.5866948962211609,
|
| 468 |
+
0.5965925455093384,
|
| 469 |
+
0.40257659554481506,
|
| 470 |
+
0.41161125898361206,
|
| 471 |
+
0.3845840096473694,
|
| 472 |
+
0.4100808799266815,
|
| 473 |
+
0.41701117157936096,
|
| 474 |
+
0.44743281602859497,
|
| 475 |
+
0.3830299377441406,
|
| 476 |
+
0.5776842832565308,
|
| 477 |
+
0.602536141872406,
|
| 478 |
+
0.6121425628662109,
|
| 479 |
+
0.4026949107646942,
|
| 480 |
+
0.6027580499649048,
|
| 481 |
+
0.5705916285514832,
|
| 482 |
+
0.5934571623802185,
|
| 483 |
+
0.5883546471595764,
|
| 484 |
+
0.5897589325904846,
|
| 485 |
+
0.5788743495941162,
|
| 486 |
+
0.42781656980514526,
|
| 487 |
+
0.6336899399757385,
|
| 488 |
+
0.5650004148483276,
|
| 489 |
+
0.5743370056152344,
|
| 490 |
+
0.4256901144981384,
|
| 491 |
+
0.4294798672199249,
|
| 492 |
+
0.4059537351131439,
|
| 493 |
+
0.5902615189552307,
|
| 494 |
+
0.4481872022151947,
|
| 495 |
+
0.46126502752304077,
|
| 496 |
+
0.5778518915176392,
|
| 497 |
+
0.409288614988327,
|
| 498 |
+
0.5872939229011536,
|
| 499 |
+
0.553527295589447,
|
| 500 |
+
0.6006361246109009,
|
| 501 |
+
0.6073405146598816,
|
| 502 |
+
0.3934749662876129,
|
| 503 |
+
0.5734230279922485,
|
| 504 |
+
0.5781282782554626,
|
| 505 |
+
0.4140267074108124,
|
| 506 |
+
0.4379793703556061,
|
| 507 |
+
0.5872884392738342,
|
| 508 |
+
0.5996562242507935,
|
| 509 |
+
0.6247859001159668,
|
| 510 |
+
0.42214199900627136,
|
| 511 |
+
0.4056937098503113,
|
| 512 |
+
0.5941511392593384,
|
| 513 |
+
0.4240345358848572,
|
| 514 |
+
0.3906824290752411,
|
| 515 |
+
0.4487050473690033,
|
| 516 |
+
0.4064308702945709,
|
| 517 |
+
0.4991621673107147,
|
| 518 |
+
0.3930267095565796,
|
| 519 |
+
0.4262542128562927,
|
| 520 |
+
0.39416787028312683,
|
| 521 |
+
0.6034715175628662,
|
| 522 |
+
0.5542858839035034
|
| 523 |
+
],
|
| 524 |
+
"eval_code_is_correct_runtime": 38.6577,
|
| 525 |
+
"eval_code_is_correct_samples_per_second": 2.587,
|
| 526 |
+
"eval_code_is_correct_steps_per_second": 0.052,
|
| 527 |
+
"epoch": 2.63,
|
| 528 |
+
"step": 50
|
| 529 |
+
},
|
| 530 |
+
{
|
| 531 |
+
"loss": 0.6992,
|
| 532 |
+
"learning_rate": 0.0002,
|
| 533 |
+
"epoch": 2.68,
|
| 534 |
+
"step": 51
|
| 535 |
+
},
|
| 536 |
+
{
|
| 537 |
+
"loss": 0.7576,
|
| 538 |
+
"learning_rate": 0.0002,
|
| 539 |
+
"epoch": 2.74,
|
| 540 |
+
"step": 52
|
| 541 |
+
},
|
| 542 |
+
{
|
| 543 |
+
"loss": 0.6925,
|
| 544 |
+
"learning_rate": 0.0002,
|
| 545 |
+
"epoch": 2.79,
|
| 546 |
+
"step": 53
|
| 547 |
+
},
|
| 548 |
+
{
|
| 549 |
+
"loss": 0.6884,
|
| 550 |
+
"learning_rate": 0.0002,
|
| 551 |
+
"epoch": 2.84,
|
| 552 |
+
"step": 54
|
| 553 |
+
},
|
| 554 |
+
{
|
| 555 |
+
"loss": 0.6555,
|
| 556 |
+
"learning_rate": 0.0002,
|
| 557 |
+
"epoch": 2.89,
|
| 558 |
+
"step": 55
|
| 559 |
+
},
|
| 560 |
+
{
|
| 561 |
+
"loss": 0.7276,
|
| 562 |
+
"learning_rate": 0.0002,
|
| 563 |
+
"epoch": 2.95,
|
| 564 |
+
"step": 56
|
| 565 |
+
},
|
| 566 |
+
{
|
| 567 |
+
"loss": 0.6827,
|
| 568 |
+
"learning_rate": 0.0002,
|
| 569 |
+
"epoch": 3.0,
|
| 570 |
+
"step": 57
|
| 571 |
+
},
|
| 572 |
+
{
|
| 573 |
+
"loss": 0.6291,
|
| 574 |
+
"learning_rate": 0.0002,
|
| 575 |
+
"epoch": 3.05,
|
| 576 |
+
"step": 58
|
| 577 |
+
},
|
| 578 |
+
{
|
| 579 |
+
"loss": 0.7595,
|
| 580 |
+
"learning_rate": 0.0002,
|
| 581 |
+
"epoch": 3.11,
|
| 582 |
+
"step": 59
|
| 583 |
+
},
|
| 584 |
+
{
|
| 585 |
+
"loss": 0.6753,
|
| 586 |
+
"learning_rate": 0.0002,
|
| 587 |
+
"epoch": 3.16,
|
| 588 |
+
"step": 60
|
| 589 |
+
},
|
| 590 |
+
{
|
| 591 |
+
"loss": 0.7555,
|
| 592 |
+
"learning_rate": 0.0002,
|
| 593 |
+
"epoch": 3.21,
|
| 594 |
+
"step": 61
|
| 595 |
+
},
|
| 596 |
+
{
|
| 597 |
+
"loss": 0.6965,
|
| 598 |
+
"learning_rate": 0.0002,
|
| 599 |
+
"epoch": 3.26,
|
| 600 |
+
"step": 62
|
| 601 |
+
},
|
| 602 |
+
{
|
| 603 |
+
"loss": 0.6775,
|
| 604 |
+
"learning_rate": 0.0002,
|
| 605 |
+
"epoch": 3.32,
|
| 606 |
+
"step": 63
|
| 607 |
+
},
|
| 608 |
+
{
|
| 609 |
+
"loss": 0.6841,
|
| 610 |
+
"learning_rate": 0.0002,
|
| 611 |
+
"epoch": 3.37,
|
| 612 |
+
"step": 64
|
| 613 |
+
},
|
| 614 |
+
{
|
| 615 |
+
"loss": 0.6706,
|
| 616 |
+
"learning_rate": 0.0002,
|
| 617 |
+
"epoch": 3.42,
|
| 618 |
+
"step": 65
|
| 619 |
+
},
|
| 620 |
+
{
|
| 621 |
+
"loss": 0.7128,
|
| 622 |
+
"learning_rate": 0.0002,
|
| 623 |
+
"epoch": 3.47,
|
| 624 |
+
"step": 66
|
| 625 |
+
},
|
| 626 |
+
{
|
| 627 |
+
"loss": 0.7036,
|
| 628 |
+
"learning_rate": 0.0002,
|
| 629 |
+
"epoch": 3.53,
|
| 630 |
+
"step": 67
|
| 631 |
+
},
|
| 632 |
+
{
|
| 633 |
+
"loss": 0.67,
|
| 634 |
+
"learning_rate": 0.0002,
|
| 635 |
+
"epoch": 3.58,
|
| 636 |
+
"step": 68
|
| 637 |
+
},
|
| 638 |
+
{
|
| 639 |
+
"loss": 0.6816,
|
| 640 |
+
"learning_rate": 0.0002,
|
| 641 |
+
"epoch": 3.63,
|
| 642 |
+
"step": 69
|
| 643 |
+
},
|
| 644 |
+
{
|
| 645 |
+
"loss": 0.6635,
|
| 646 |
+
"learning_rate": 0.0002,
|
| 647 |
+
"epoch": 3.68,
|
| 648 |
+
"step": 70
|
| 649 |
+
},
|
| 650 |
+
{
|
| 651 |
+
"loss": 0.6457,
|
| 652 |
+
"learning_rate": 0.0002,
|
| 653 |
+
"epoch": 3.74,
|
| 654 |
+
"step": 71
|
| 655 |
+
},
|
| 656 |
+
{
|
| 657 |
+
"loss": 0.6369,
|
| 658 |
+
"learning_rate": 0.0002,
|
| 659 |
+
"epoch": 3.79,
|
| 660 |
+
"step": 72
|
| 661 |
+
},
|
| 662 |
+
{
|
| 663 |
+
"loss": 0.6955,
|
| 664 |
+
"learning_rate": 0.0002,
|
| 665 |
+
"epoch": 3.84,
|
| 666 |
+
"step": 73
|
| 667 |
+
},
|
| 668 |
+
{
|
| 669 |
+
"loss": 0.6937,
|
| 670 |
+
"learning_rate": 0.0002,
|
| 671 |
+
"epoch": 3.89,
|
| 672 |
+
"step": 74
|
| 673 |
+
},
|
| 674 |
+
{
|
| 675 |
+
"loss": 0.6842,
|
| 676 |
+
"learning_rate": 0.0002,
|
| 677 |
+
"epoch": 3.95,
|
| 678 |
+
"step": 75
|
| 679 |
+
},
|
| 680 |
+
{
|
| 681 |
+
"eval_code_is_correct_loss": 0.7020106315612793,
|
| 682 |
+
"eval_code_is_correct_score": -0.2538911998271942,
|
| 683 |
+
"eval_code_is_correct_brier_score": 0.2538911998271942,
|
| 684 |
+
"eval_code_is_correct_average_probability": 0.502829372882843,
|
| 685 |
+
"eval_code_is_correct_accuracy": 0.45,
|
| 686 |
+
"eval_code_is_correct_probabilities": [
|
| 687 |
+
0.5088492035865784,
|
| 688 |
+
0.4741206467151642,
|
| 689 |
+
0.4018367826938629,
|
| 690 |
+
0.49836266040802,
|
| 691 |
+
0.3702163100242615,
|
| 692 |
+
0.6138777732849121,
|
| 693 |
+
0.43586498498916626,
|
| 694 |
+
0.6494307518005371,
|
| 695 |
+
0.5666269063949585,
|
| 696 |
+
0.5525698065757751,
|
| 697 |
+
0.3989929258823395,
|
| 698 |
+
0.49194008111953735,
|
| 699 |
+
0.5326002836227417,
|
| 700 |
+
0.5366830825805664,
|
| 701 |
+
0.42133161425590515,
|
| 702 |
+
0.42769843339920044,
|
| 703 |
+
0.44886454939842224,
|
| 704 |
+
0.4730355739593506,
|
| 705 |
+
0.48204565048217773,
|
| 706 |
+
0.4801846444606781,
|
| 707 |
+
0.5868195295333862,
|
| 708 |
+
0.4399736523628235,
|
| 709 |
+
0.3782220780849457,
|
| 710 |
+
0.5690118074417114,
|
| 711 |
+
0.5845021605491638,
|
| 712 |
+
0.5171434283256531,
|
| 713 |
+
0.41184958815574646,
|
| 714 |
+
0.4574517011642456,
|
| 715 |
+
0.41264650225639343,
|
| 716 |
+
0.6442864537239075,
|
| 717 |
+
0.48700496554374695,
|
| 718 |
+
0.3954983949661255,
|
| 719 |
+
0.4781593382358551,
|
| 720 |
+
0.43342939019203186,
|
| 721 |
+
0.4707704186439514,
|
| 722 |
+
0.5405588746070862,
|
| 723 |
+
0.4150758981704712,
|
| 724 |
+
0.4007004499435425,
|
| 725 |
+
0.34886103868484497,
|
| 726 |
+
0.5174517631530762,
|
| 727 |
+
0.47847726941108704,
|
| 728 |
+
0.548160970211029,
|
| 729 |
+
0.7044894099235535,
|
| 730 |
+
0.45586660504341125,
|
| 731 |
+
0.4582586884498596,
|
| 732 |
+
0.5384201407432556,
|
| 733 |
+
0.5183372497558594,
|
| 734 |
+
0.5092693567276001,
|
| 735 |
+
0.5427851676940918,
|
| 736 |
+
0.5732136964797974,
|
| 737 |
+
0.5735348463058472,
|
| 738 |
+
0.6746758818626404,
|
| 739 |
+
0.49393337965011597,
|
| 740 |
+
0.42578810453414917,
|
| 741 |
+
0.4929383099079132,
|
| 742 |
+
0.4470900893211365,
|
| 743 |
+
0.574955940246582,
|
| 744 |
+
0.4823959767818451,
|
| 745 |
+
0.3653642237186432,
|
| 746 |
+
0.46867260336875916,
|
| 747 |
+
0.42166757583618164,
|
| 748 |
+
0.47335392236709595,
|
| 749 |
+
0.4112188518047333,
|
| 750 |
+
0.6188163757324219,
|
| 751 |
+
0.5855951905250549,
|
| 752 |
+
0.47489026188850403,
|
| 753 |
+
0.45155712962150574,
|
| 754 |
+
0.6093342900276184,
|
| 755 |
+
0.605854868888855,
|
| 756 |
+
0.562972903251648,
|
| 757 |
+
0.4357713460922241,
|
| 758 |
+
0.598077118396759,
|
| 759 |
+
0.5654999613761902,
|
| 760 |
+
0.43106329441070557,
|
| 761 |
+
0.5422348976135254,
|
| 762 |
+
0.40416285395622253,
|
| 763 |
+
0.4005366861820221,
|
| 764 |
+
0.46095961332321167,
|
| 765 |
+
0.490922749042511,
|
| 766 |
+
0.49807679653167725,
|
| 767 |
+
0.40843698382377625,
|
| 768 |
+
0.4329848885536194,
|
| 769 |
+
0.5059437155723572,
|
| 770 |
+
0.6385719180107117,
|
| 771 |
+
0.49556964635849,
|
| 772 |
+
0.43445929884910583,
|
| 773 |
+
0.49239978194236755,
|
| 774 |
+
0.6199828386306763,
|
| 775 |
+
0.5497725605964661,
|
| 776 |
+
0.40616798400878906,
|
| 777 |
+
0.6228035092353821,
|
| 778 |
+
0.5643633008003235,
|
| 779 |
+
0.5394735932350159,
|
| 780 |
+
0.5994738340377808,
|
| 781 |
+
0.7553389668464661,
|
| 782 |
+
0.5502137541770935,
|
| 783 |
+
0.579784631729126,
|
| 784 |
+
0.5687956809997559,
|
| 785 |
+
0.5465372800827026,
|
| 786 |
+
0.3481208384037018
|
| 787 |
+
],
|
| 788 |
+
"eval_code_is_correct_runtime": 38.6558,
|
| 789 |
+
"eval_code_is_correct_samples_per_second": 2.587,
|
| 790 |
+
"eval_code_is_correct_steps_per_second": 0.052,
|
| 791 |
+
"epoch": 3.95,
|
| 792 |
+
"step": 75
|
| 793 |
+
},
|
| 794 |
+
{
|
| 795 |
+
"loss": 0.6694,
|
| 796 |
+
"learning_rate": 0.0002,
|
| 797 |
+
"epoch": 4.0,
|
| 798 |
+
"step": 76
|
| 799 |
+
},
|
| 800 |
+
{
|
| 801 |
+
"loss": 0.6634,
|
| 802 |
+
"learning_rate": 0.0002,
|
| 803 |
+
"epoch": 4.05,
|
| 804 |
+
"step": 77
|
| 805 |
+
},
|
| 806 |
+
{
|
| 807 |
+
"loss": 0.6602,
|
| 808 |
+
"learning_rate": 0.0002,
|
| 809 |
+
"epoch": 4.11,
|
| 810 |
+
"step": 78
|
| 811 |
+
},
|
| 812 |
+
{
|
| 813 |
+
"loss": 0.6396,
|
| 814 |
+
"learning_rate": 0.0002,
|
| 815 |
+
"epoch": 4.16,
|
| 816 |
+
"step": 79
|
| 817 |
+
},
|
| 818 |
+
{
|
| 819 |
+
"loss": 0.7378,
|
| 820 |
+
"learning_rate": 0.0002,
|
| 821 |
+
"epoch": 4.21,
|
| 822 |
+
"step": 80
|
| 823 |
+
},
|
| 824 |
+
{
|
| 825 |
+
"loss": 0.8125,
|
| 826 |
+
"learning_rate": 0.0002,
|
| 827 |
+
"epoch": 4.26,
|
| 828 |
+
"step": 81
|
| 829 |
+
},
|
| 830 |
+
{
|
| 831 |
+
"loss": 0.5817,
|
| 832 |
+
"learning_rate": 0.0002,
|
| 833 |
+
"epoch": 4.32,
|
| 834 |
+
"step": 82
|
| 835 |
+
},
|
| 836 |
+
{
|
| 837 |
+
"loss": 0.6169,
|
| 838 |
+
"learning_rate": 0.0002,
|
| 839 |
+
"epoch": 4.37,
|
| 840 |
+
"step": 83
|
| 841 |
+
},
|
| 842 |
+
{
|
| 843 |
+
"loss": 0.6157,
|
| 844 |
+
"learning_rate": 0.0002,
|
| 845 |
+
"epoch": 4.42,
|
| 846 |
+
"step": 84
|
| 847 |
+
},
|
| 848 |
+
{
|
| 849 |
+
"loss": 0.6109,
|
| 850 |
+
"learning_rate": 0.0002,
|
| 851 |
+
"epoch": 4.47,
|
| 852 |
+
"step": 85
|
| 853 |
+
},
|
| 854 |
+
{
|
| 855 |
+
"loss": 0.5057,
|
| 856 |
+
"learning_rate": 0.0002,
|
| 857 |
+
"epoch": 4.53,
|
| 858 |
+
"step": 86
|
| 859 |
+
},
|
| 860 |
+
{
|
| 861 |
+
"loss": 0.5264,
|
| 862 |
+
"learning_rate": 0.0002,
|
| 863 |
+
"epoch": 4.58,
|
| 864 |
+
"step": 87
|
| 865 |
+
},
|
| 866 |
+
{
|
| 867 |
+
"loss": 0.6092,
|
| 868 |
+
"learning_rate": 0.0002,
|
| 869 |
+
"epoch": 4.63,
|
| 870 |
+
"step": 88
|
| 871 |
+
},
|
| 872 |
+
{
|
| 873 |
+
"loss": 0.5337,
|
| 874 |
+
"learning_rate": 0.0002,
|
| 875 |
+
"epoch": 4.68,
|
| 876 |
+
"step": 89
|
| 877 |
+
},
|
| 878 |
+
{
|
| 879 |
+
"loss": 0.6412,
|
| 880 |
+
"learning_rate": 0.0002,
|
| 881 |
+
"epoch": 4.74,
|
| 882 |
+
"step": 90
|
| 883 |
+
},
|
| 884 |
+
{
|
| 885 |
+
"loss": 0.5222,
|
| 886 |
+
"learning_rate": 0.0002,
|
| 887 |
+
"epoch": 4.79,
|
| 888 |
+
"step": 91
|
| 889 |
+
},
|
| 890 |
+
{
|
| 891 |
+
"loss": 0.612,
|
| 892 |
+
"learning_rate": 0.0002,
|
| 893 |
+
"epoch": 4.84,
|
| 894 |
+
"step": 92
|
| 895 |
+
},
|
| 896 |
+
{
|
| 897 |
+
"loss": 0.6778,
|
| 898 |
+
"learning_rate": 0.0002,
|
| 899 |
+
"epoch": 4.89,
|
| 900 |
+
"step": 93
|
| 901 |
+
},
|
| 902 |
+
{
|
| 903 |
+
"loss": 0.673,
|
| 904 |
+
"learning_rate": 0.0002,
|
| 905 |
+
"epoch": 4.95,
|
| 906 |
+
"step": 94
|
| 907 |
+
},
|
| 908 |
+
{
|
| 909 |
+
"loss": 0.5896,
|
| 910 |
+
"learning_rate": 0.0002,
|
| 911 |
+
"epoch": 5.0,
|
| 912 |
+
"step": 95
|
| 913 |
+
},
|
| 914 |
+
{
|
| 915 |
+
"loss": 0.6669,
|
| 916 |
+
"learning_rate": 0.0002,
|
| 917 |
+
"epoch": 5.05,
|
| 918 |
+
"step": 96
|
| 919 |
+
},
|
| 920 |
+
{
|
| 921 |
+
"loss": 0.7279,
|
| 922 |
+
"learning_rate": 0.0002,
|
| 923 |
+
"epoch": 5.11,
|
| 924 |
+
"step": 97
|
| 925 |
+
},
|
| 926 |
+
{
|
| 927 |
+
"loss": 0.4678,
|
| 928 |
+
"learning_rate": 0.0002,
|
| 929 |
+
"epoch": 5.16,
|
| 930 |
+
"step": 98
|
| 931 |
+
},
|
| 932 |
+
{
|
| 933 |
+
"loss": 0.5572,
|
| 934 |
+
"learning_rate": 0.0002,
|
| 935 |
+
"epoch": 5.21,
|
| 936 |
+
"step": 99
|
| 937 |
+
},
|
| 938 |
+
{
|
| 939 |
+
"loss": 0.5173,
|
| 940 |
+
"learning_rate": 0.0002,
|
| 941 |
+
"epoch": 5.26,
|
| 942 |
+
"step": 100
|
| 943 |
+
},
|
| 944 |
+
{
|
| 945 |
+
"eval_code_is_correct_loss": 0.6044353246688843,
|
| 946 |
+
"eval_code_is_correct_score": -0.20567844808101654,
|
| 947 |
+
"eval_code_is_correct_brier_score": 0.20567844808101654,
|
| 948 |
+
"eval_code_is_correct_average_probability": 0.5809673070907593,
|
| 949 |
+
"eval_code_is_correct_accuracy": 0.69,
|
| 950 |
+
"eval_code_is_correct_probabilities": [
|
| 951 |
+
0.495758056640625,
|
| 952 |
+
0.6722157001495361,
|
| 953 |
+
0.5509936213493347,
|
| 954 |
+
0.30042609572410583,
|
| 955 |
+
0.4185842275619507,
|
| 956 |
+
0.6230117678642273,
|
| 957 |
+
0.6212649941444397,
|
| 958 |
+
0.6848446726799011,
|
| 959 |
+
0.6345605850219727,
|
| 960 |
+
0.3738310635089874,
|
| 961 |
+
0.4916575253009796,
|
| 962 |
+
0.4045080244541168,
|
| 963 |
+
0.43291226029396057,
|
| 964 |
+
0.4618946313858032,
|
| 965 |
+
0.6218451857566833,
|
| 966 |
+
0.3830638527870178,
|
| 967 |
+
0.6171413064002991,
|
| 968 |
+
0.2532573342323303,
|
| 969 |
+
0.6897477507591248,
|
| 970 |
+
0.6743988990783691,
|
| 971 |
+
0.7978313565254211,
|
| 972 |
+
0.6350521445274353,
|
| 973 |
+
0.5316161513328552,
|
| 974 |
+
0.347903311252594,
|
| 975 |
+
0.6238374710083008,
|
| 976 |
+
0.298748254776001,
|
| 977 |
+
0.5980588793754578,
|
| 978 |
+
0.6492370963096619,
|
| 979 |
+
0.5917383432388306,
|
| 980 |
+
0.6613048911094666,
|
| 981 |
+
0.6900146007537842,
|
| 982 |
+
0.5732241272926331,
|
| 983 |
+
0.23556573688983917,
|
| 984 |
+
0.5519108772277832,
|
| 985 |
+
0.7326077222824097,
|
| 986 |
+
0.856022834777832,
|
| 987 |
+
0.5683114528656006,
|
| 988 |
+
0.6112112402915955,
|
| 989 |
+
0.43082374334335327,
|
| 990 |
+
0.2740175426006317,
|
| 991 |
+
0.710313618183136,
|
| 992 |
+
0.3746766149997711,
|
| 993 |
+
0.9831528067588806,
|
| 994 |
+
0.7283727526664734,
|
| 995 |
+
0.6630771160125732,
|
| 996 |
+
0.7669262886047363,
|
| 997 |
+
0.310622900724411,
|
| 998 |
+
0.4299599826335907,
|
| 999 |
+
0.40406081080436707,
|
| 1000 |
+
0.569300651550293,
|
| 1001 |
+
0.3967253267765045,
|
| 1002 |
+
0.9761428833007812,
|
| 1003 |
+
0.25975796580314636,
|
| 1004 |
+
0.5678868889808655,
|
| 1005 |
+
0.6639503836631775,
|
| 1006 |
+
0.5604276657104492,
|
| 1007 |
+
0.5109429955482483,
|
| 1008 |
+
0.7717424631118774,
|
| 1009 |
+
0.5076764225959778,
|
| 1010 |
+
0.480741024017334,
|
| 1011 |
+
0.6118683218955994,
|
| 1012 |
+
0.7221248149871826,
|
| 1013 |
+
0.42638152837753296,
|
| 1014 |
+
0.9544664621353149,
|
| 1015 |
+
0.7895866632461548,
|
| 1016 |
+
0.5485232472419739,
|
| 1017 |
+
0.5473383665084839,
|
| 1018 |
+
0.5949488878250122,
|
| 1019 |
+
0.6287338137626648,
|
| 1020 |
+
0.3621536195278168,
|
| 1021 |
+
0.5074862837791443,
|
| 1022 |
+
0.8851379156112671,
|
| 1023 |
+
0.8405280113220215,
|
| 1024 |
+
0.6945863962173462,
|
| 1025 |
+
0.5270545482635498,
|
| 1026 |
+
0.5763978362083435,
|
| 1027 |
+
0.6713876724243164,
|
| 1028 |
+
0.7237599492073059,
|
| 1029 |
+
0.8137936592102051,
|
| 1030 |
+
0.33156439661979675,
|
| 1031 |
+
0.6127635836601257,
|
| 1032 |
+
0.44015809893608093,
|
| 1033 |
+
0.2821398079395294,
|
| 1034 |
+
0.6921646595001221,
|
| 1035 |
+
0.6725201606750488,
|
| 1036 |
+
0.6315755844116211,
|
| 1037 |
+
0.7930103540420532,
|
| 1038 |
+
0.7124245166778564,
|
| 1039 |
+
0.3897824287414551,
|
| 1040 |
+
0.5918102860450745,
|
| 1041 |
+
0.7035940289497375,
|
| 1042 |
+
0.36644646525382996,
|
| 1043 |
+
0.4447966516017914,
|
| 1044 |
+
0.5399353504180908,
|
| 1045 |
+
0.9951468110084534,
|
| 1046 |
+
0.3498772382736206,
|
| 1047 |
+
0.765509307384491,
|
| 1048 |
+
0.6127597689628601,
|
| 1049 |
+
0.8720724582672119,
|
| 1050 |
+
0.5690097212791443
|
| 1051 |
+
],
|
| 1052 |
+
"eval_code_is_correct_runtime": 38.7216,
|
| 1053 |
+
"eval_code_is_correct_samples_per_second": 2.583,
|
| 1054 |
+
"eval_code_is_correct_steps_per_second": 0.052,
|
| 1055 |
+
"epoch": 5.26,
|
| 1056 |
+
"step": 100
|
| 1057 |
+
},
|
| 1058 |
+
{
|
| 1059 |
+
"train_runtime": 2619.7779,
|
| 1060 |
+
"train_samples_per_second": 1.221,
|
| 1061 |
+
"train_steps_per_second": 0.038,
|
| 1062 |
+
"total_flos": 0.0,
|
| 1063 |
+
"train_loss": 0.6988758665323257,
|
| 1064 |
+
"epoch": 5.26,
|
| 1065 |
+
"step": 100
|
| 1066 |
+
}
|
| 1067 |
+
]
|