Commit ·
990f704
1
Parent(s): 8a98df1
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-30b",
|
| 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 sha256:4af317572ebb393ec91a722a322143b919d9daa61f89132f876608f2db28b966
|
| 3 |
+
size 409081617
|
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-30b",
|
| 3 |
+
"distribution_id": "code_is_correct",
|
| 4 |
+
"date_trained": "12/10/2023 22:59:53",
|
| 5 |
+
"output_dir": "models/classify_lora/llama-30b-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": 8,
|
| 13 |
+
"per_device_eval_batch_size": 8,
|
| 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-30b-code_is_correct/runs/Oct12_21-11-21_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": 1,
|
| 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-30b-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 sha256:83376142b6acdfdbd765c8cbb554223ef940aeb5b420941689ee78401b6a810c
|
| 3 |
+
size 5115
|
training_logs.json
ADDED
|
@@ -0,0 +1,1067 @@
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
| 1 |
+
[
|
| 2 |
+
{
|
| 3 |
+
"loss": 0.687,
|
| 4 |
+
"learning_rate": 0.0002,
|
| 5 |
+
"epoch": 0.05,
|
| 6 |
+
"step": 1
|
| 7 |
+
},
|
| 8 |
+
{
|
| 9 |
+
"loss": 0.722,
|
| 10 |
+
"learning_rate": 0.0002,
|
| 11 |
+
"epoch": 0.11,
|
| 12 |
+
"step": 2
|
| 13 |
+
},
|
| 14 |
+
{
|
| 15 |
+
"loss": 0.7011,
|
| 16 |
+
"learning_rate": 0.0002,
|
| 17 |
+
"epoch": 0.16,
|
| 18 |
+
"step": 3
|
| 19 |
+
},
|
| 20 |
+
{
|
| 21 |
+
"loss": 0.6987,
|
| 22 |
+
"learning_rate": 0.0002,
|
| 23 |
+
"epoch": 0.21,
|
| 24 |
+
"step": 4
|
| 25 |
+
},
|
| 26 |
+
{
|
| 27 |
+
"loss": 0.6776,
|
| 28 |
+
"learning_rate": 0.0002,
|
| 29 |
+
"epoch": 0.26,
|
| 30 |
+
"step": 5
|
| 31 |
+
},
|
| 32 |
+
{
|
| 33 |
+
"loss": 0.7507,
|
| 34 |
+
"learning_rate": 0.0002,
|
| 35 |
+
"epoch": 0.32,
|
| 36 |
+
"step": 6
|
| 37 |
+
},
|
| 38 |
+
{
|
| 39 |
+
"loss": 0.6607,
|
| 40 |
+
"learning_rate": 0.0002,
|
| 41 |
+
"epoch": 0.37,
|
| 42 |
+
"step": 7
|
| 43 |
+
},
|
| 44 |
+
{
|
| 45 |
+
"loss": 0.6974,
|
| 46 |
+
"learning_rate": 0.0002,
|
| 47 |
+
"epoch": 0.42,
|
| 48 |
+
"step": 8
|
| 49 |
+
},
|
| 50 |
+
{
|
| 51 |
+
"loss": 0.727,
|
| 52 |
+
"learning_rate": 0.0002,
|
| 53 |
+
"epoch": 0.47,
|
| 54 |
+
"step": 9
|
| 55 |
+
},
|
| 56 |
+
{
|
| 57 |
+
"loss": 0.6979,
|
| 58 |
+
"learning_rate": 0.0002,
|
| 59 |
+
"epoch": 0.53,
|
| 60 |
+
"step": 10
|
| 61 |
+
},
|
| 62 |
+
{
|
| 63 |
+
"loss": 0.7196,
|
| 64 |
+
"learning_rate": 0.0002,
|
| 65 |
+
"epoch": 0.58,
|
| 66 |
+
"step": 11
|
| 67 |
+
},
|
| 68 |
+
{
|
| 69 |
+
"loss": 0.6835,
|
| 70 |
+
"learning_rate": 0.0002,
|
| 71 |
+
"epoch": 0.63,
|
| 72 |
+
"step": 12
|
| 73 |
+
},
|
| 74 |
+
{
|
| 75 |
+
"loss": 0.689,
|
| 76 |
+
"learning_rate": 0.0002,
|
| 77 |
+
"epoch": 0.68,
|
| 78 |
+
"step": 13
|
| 79 |
+
},
|
| 80 |
+
{
|
| 81 |
+
"loss": 0.6883,
|
| 82 |
+
"learning_rate": 0.0002,
|
| 83 |
+
"epoch": 0.74,
|
| 84 |
+
"step": 14
|
| 85 |
+
},
|
| 86 |
+
{
|
| 87 |
+
"loss": 0.6923,
|
| 88 |
+
"learning_rate": 0.0002,
|
| 89 |
+
"epoch": 0.79,
|
| 90 |
+
"step": 15
|
| 91 |
+
},
|
| 92 |
+
{
|
| 93 |
+
"loss": 0.6878,
|
| 94 |
+
"learning_rate": 0.0002,
|
| 95 |
+
"epoch": 0.84,
|
| 96 |
+
"step": 16
|
| 97 |
+
},
|
| 98 |
+
{
|
| 99 |
+
"loss": 0.6622,
|
| 100 |
+
"learning_rate": 0.0002,
|
| 101 |
+
"epoch": 0.89,
|
| 102 |
+
"step": 17
|
| 103 |
+
},
|
| 104 |
+
{
|
| 105 |
+
"loss": 0.6656,
|
| 106 |
+
"learning_rate": 0.0002,
|
| 107 |
+
"epoch": 0.95,
|
| 108 |
+
"step": 18
|
| 109 |
+
},
|
| 110 |
+
{
|
| 111 |
+
"loss": 0.6045,
|
| 112 |
+
"learning_rate": 0.0002,
|
| 113 |
+
"epoch": 1.0,
|
| 114 |
+
"step": 19
|
| 115 |
+
},
|
| 116 |
+
{
|
| 117 |
+
"loss": 0.7383,
|
| 118 |
+
"learning_rate": 0.0002,
|
| 119 |
+
"epoch": 1.05,
|
| 120 |
+
"step": 20
|
| 121 |
+
},
|
| 122 |
+
{
|
| 123 |
+
"loss": 0.6277,
|
| 124 |
+
"learning_rate": 0.0002,
|
| 125 |
+
"epoch": 1.11,
|
| 126 |
+
"step": 21
|
| 127 |
+
},
|
| 128 |
+
{
|
| 129 |
+
"loss": 0.7156,
|
| 130 |
+
"learning_rate": 0.0002,
|
| 131 |
+
"epoch": 1.16,
|
| 132 |
+
"step": 22
|
| 133 |
+
},
|
| 134 |
+
{
|
| 135 |
+
"loss": 0.6615,
|
| 136 |
+
"learning_rate": 0.0002,
|
| 137 |
+
"epoch": 1.21,
|
| 138 |
+
"step": 23
|
| 139 |
+
},
|
| 140 |
+
{
|
| 141 |
+
"loss": 0.681,
|
| 142 |
+
"learning_rate": 0.0002,
|
| 143 |
+
"epoch": 1.26,
|
| 144 |
+
"step": 24
|
| 145 |
+
},
|
| 146 |
+
{
|
| 147 |
+
"loss": 0.6368,
|
| 148 |
+
"learning_rate": 0.0002,
|
| 149 |
+
"epoch": 1.32,
|
| 150 |
+
"step": 25
|
| 151 |
+
},
|
| 152 |
+
{
|
| 153 |
+
"eval_code_is_correct_loss": 0.6748137474060059,
|
| 154 |
+
"eval_code_is_correct_score": -0.24037876725196838,
|
| 155 |
+
"eval_code_is_correct_brier_score": 0.24037876725196838,
|
| 156 |
+
"eval_code_is_correct_average_probability": 0.5268679857254028,
|
| 157 |
+
"eval_code_is_correct_accuracy": 0.56,
|
| 158 |
+
"eval_code_is_correct_probabilities": [
|
| 159 |
+
0.43019697070121765,
|
| 160 |
+
0.6227403879165649,
|
| 161 |
+
0.6141183376312256,
|
| 162 |
+
0.3344857096672058,
|
| 163 |
+
0.6293616890907288,
|
| 164 |
+
0.4664916694164276,
|
| 165 |
+
0.6512068510055542,
|
| 166 |
+
0.44083818793296814,
|
| 167 |
+
0.417765736579895,
|
| 168 |
+
0.3489360511302948,
|
| 169 |
+
0.6306682825088501,
|
| 170 |
+
0.40677180886268616,
|
| 171 |
+
0.36523571610450745,
|
| 172 |
+
0.46857500076293945,
|
| 173 |
+
0.6413342356681824,
|
| 174 |
+
0.6176690459251404,
|
| 175 |
+
0.6807307004928589,
|
| 176 |
+
0.31330692768096924,
|
| 177 |
+
0.6547641754150391,
|
| 178 |
+
0.6119447350502014,
|
| 179 |
+
0.4726662337779999,
|
| 180 |
+
0.6131753325462341,
|
| 181 |
+
0.600269079208374,
|
| 182 |
+
0.33192408084869385,
|
| 183 |
+
0.3614324629306793,
|
| 184 |
+
0.34707942605018616,
|
| 185 |
+
0.6705795526504517,
|
| 186 |
+
0.6425331234931946,
|
| 187 |
+
0.5840022563934326,
|
| 188 |
+
0.3998101055622101,
|
| 189 |
+
0.6653164625167847,
|
| 190 |
+
0.574996292591095,
|
| 191 |
+
0.37777000665664673,
|
| 192 |
+
0.6291854977607727,
|
| 193 |
+
0.6616199612617493,
|
| 194 |
+
0.4081592559814453,
|
| 195 |
+
0.6753640174865723,
|
| 196 |
+
0.6286922693252563,
|
| 197 |
+
0.6147665977478027,
|
| 198 |
+
0.4005477726459503,
|
| 199 |
+
0.6473195552825928,
|
| 200 |
+
0.35249093174934387,
|
| 201 |
+
0.540384829044342,
|
| 202 |
+
0.6459094285964966,
|
| 203 |
+
0.6578614115715027,
|
| 204 |
+
0.683413028717041,
|
| 205 |
+
0.31949999928474426,
|
| 206 |
+
0.38618671894073486,
|
| 207 |
+
0.3694519102573395,
|
| 208 |
+
0.3965170979499817,
|
| 209 |
+
0.3777723014354706,
|
| 210 |
+
0.553892195224762,
|
| 211 |
+
0.35859158635139465,
|
| 212 |
+
0.6390789747238159,
|
| 213 |
+
0.6438376307487488,
|
| 214 |
+
0.6507594585418701,
|
| 215 |
+
0.3823850154876709,
|
| 216 |
+
0.6562716364860535,
|
| 217 |
+
0.6082704663276672,
|
| 218 |
+
0.6338680386543274,
|
| 219 |
+
0.678227961063385,
|
| 220 |
+
0.6367619037628174,
|
| 221 |
+
0.5889072418212891,
|
| 222 |
+
0.4969732165336609,
|
| 223 |
+
0.6729732751846313,
|
| 224 |
+
0.6613313555717468,
|
| 225 |
+
0.6332442164421082,
|
| 226 |
+
0.4045434892177582,
|
| 227 |
+
0.38479217886924744,
|
| 228 |
+
0.3474903404712677,
|
| 229 |
+
0.6648860573768616,
|
| 230 |
+
0.46690836548805237,
|
| 231 |
+
0.5336642861366272,
|
| 232 |
+
0.633358895778656,
|
| 233 |
+
0.4208388030529022,
|
| 234 |
+
0.630777895450592,
|
| 235 |
+
0.625720202922821,
|
| 236 |
+
0.6393349766731262,
|
| 237 |
+
0.6842279434204102,
|
| 238 |
+
0.34046560525894165,
|
| 239 |
+
0.6573684811592102,
|
| 240 |
+
0.623447597026825,
|
| 241 |
+
0.3347625434398651,
|
| 242 |
+
0.439355731010437,
|
| 243 |
+
0.6188381314277649,
|
| 244 |
+
0.6108013987541199,
|
| 245 |
+
0.673854649066925,
|
| 246 |
+
0.4441554844379425,
|
| 247 |
+
0.40074998140335083,
|
| 248 |
+
0.6538906097412109,
|
| 249 |
+
0.4779479205608368,
|
| 250 |
+
0.3645903170108795,
|
| 251 |
+
0.3905934989452362,
|
| 252 |
+
0.3223285377025604,
|
| 253 |
+
0.6476134061813354,
|
| 254 |
+
0.3171766400337219,
|
| 255 |
+
0.36856701970100403,
|
| 256 |
+
0.3680305778980255,
|
| 257 |
+
0.6812744140625,
|
| 258 |
+
0.6352286338806152
|
| 259 |
+
],
|
| 260 |
+
"eval_code_is_correct_runtime": 88.8545,
|
| 261 |
+
"eval_code_is_correct_samples_per_second": 1.125,
|
| 262 |
+
"eval_code_is_correct_steps_per_second": 0.045,
|
| 263 |
+
"epoch": 1.32,
|
| 264 |
+
"step": 25
|
| 265 |
+
},
|
| 266 |
+
{
|
| 267 |
+
"loss": 0.6594,
|
| 268 |
+
"learning_rate": 0.0002,
|
| 269 |
+
"epoch": 1.37,
|
| 270 |
+
"step": 26
|
| 271 |
+
},
|
| 272 |
+
{
|
| 273 |
+
"loss": 0.6907,
|
| 274 |
+
"learning_rate": 0.0002,
|
| 275 |
+
"epoch": 1.42,
|
| 276 |
+
"step": 27
|
| 277 |
+
},
|
| 278 |
+
{
|
| 279 |
+
"loss": 0.5943,
|
| 280 |
+
"learning_rate": 0.0002,
|
| 281 |
+
"epoch": 1.47,
|
| 282 |
+
"step": 28
|
| 283 |
+
},
|
| 284 |
+
{
|
| 285 |
+
"loss": 0.6849,
|
| 286 |
+
"learning_rate": 0.0002,
|
| 287 |
+
"epoch": 1.53,
|
| 288 |
+
"step": 29
|
| 289 |
+
},
|
| 290 |
+
{
|
| 291 |
+
"loss": 0.5921,
|
| 292 |
+
"learning_rate": 0.0002,
|
| 293 |
+
"epoch": 1.58,
|
| 294 |
+
"step": 30
|
| 295 |
+
},
|
| 296 |
+
{
|
| 297 |
+
"loss": 0.7788,
|
| 298 |
+
"learning_rate": 0.0002,
|
| 299 |
+
"epoch": 1.63,
|
| 300 |
+
"step": 31
|
| 301 |
+
},
|
| 302 |
+
{
|
| 303 |
+
"loss": 0.64,
|
| 304 |
+
"learning_rate": 0.0002,
|
| 305 |
+
"epoch": 1.68,
|
| 306 |
+
"step": 32
|
| 307 |
+
},
|
| 308 |
+
{
|
| 309 |
+
"loss": 0.5442,
|
| 310 |
+
"learning_rate": 0.0002,
|
| 311 |
+
"epoch": 1.74,
|
| 312 |
+
"step": 33
|
| 313 |
+
},
|
| 314 |
+
{
|
| 315 |
+
"loss": 0.5134,
|
| 316 |
+
"learning_rate": 0.0002,
|
| 317 |
+
"epoch": 1.79,
|
| 318 |
+
"step": 34
|
| 319 |
+
},
|
| 320 |
+
{
|
| 321 |
+
"loss": 0.5845,
|
| 322 |
+
"learning_rate": 0.0002,
|
| 323 |
+
"epoch": 1.84,
|
| 324 |
+
"step": 35
|
| 325 |
+
},
|
| 326 |
+
{
|
| 327 |
+
"loss": 0.6618,
|
| 328 |
+
"learning_rate": 0.0002,
|
| 329 |
+
"epoch": 1.89,
|
| 330 |
+
"step": 36
|
| 331 |
+
},
|
| 332 |
+
{
|
| 333 |
+
"loss": 0.5564,
|
| 334 |
+
"learning_rate": 0.0002,
|
| 335 |
+
"epoch": 1.95,
|
| 336 |
+
"step": 37
|
| 337 |
+
},
|
| 338 |
+
{
|
| 339 |
+
"loss": 0.5295,
|
| 340 |
+
"learning_rate": 0.0002,
|
| 341 |
+
"epoch": 2.0,
|
| 342 |
+
"step": 38
|
| 343 |
+
},
|
| 344 |
+
{
|
| 345 |
+
"loss": 0.3214,
|
| 346 |
+
"learning_rate": 0.0002,
|
| 347 |
+
"epoch": 2.05,
|
| 348 |
+
"step": 39
|
| 349 |
+
},
|
| 350 |
+
{
|
| 351 |
+
"loss": 0.541,
|
| 352 |
+
"learning_rate": 0.0002,
|
| 353 |
+
"epoch": 2.11,
|
| 354 |
+
"step": 40
|
| 355 |
+
},
|
| 356 |
+
{
|
| 357 |
+
"loss": 0.4606,
|
| 358 |
+
"learning_rate": 0.0002,
|
| 359 |
+
"epoch": 2.16,
|
| 360 |
+
"step": 41
|
| 361 |
+
},
|
| 362 |
+
{
|
| 363 |
+
"loss": 0.3594,
|
| 364 |
+
"learning_rate": 0.0002,
|
| 365 |
+
"epoch": 2.21,
|
| 366 |
+
"step": 42
|
| 367 |
+
},
|
| 368 |
+
{
|
| 369 |
+
"loss": 0.4218,
|
| 370 |
+
"learning_rate": 0.0002,
|
| 371 |
+
"epoch": 2.26,
|
| 372 |
+
"step": 43
|
| 373 |
+
},
|
| 374 |
+
{
|
| 375 |
+
"loss": 0.7866,
|
| 376 |
+
"learning_rate": 0.0002,
|
| 377 |
+
"epoch": 2.32,
|
| 378 |
+
"step": 44
|
| 379 |
+
},
|
| 380 |
+
{
|
| 381 |
+
"loss": 0.4861,
|
| 382 |
+
"learning_rate": 0.0002,
|
| 383 |
+
"epoch": 2.37,
|
| 384 |
+
"step": 45
|
| 385 |
+
},
|
| 386 |
+
{
|
| 387 |
+
"loss": 0.715,
|
| 388 |
+
"learning_rate": 0.0002,
|
| 389 |
+
"epoch": 2.42,
|
| 390 |
+
"step": 46
|
| 391 |
+
},
|
| 392 |
+
{
|
| 393 |
+
"loss": 0.3984,
|
| 394 |
+
"learning_rate": 0.0002,
|
| 395 |
+
"epoch": 2.47,
|
| 396 |
+
"step": 47
|
| 397 |
+
},
|
| 398 |
+
{
|
| 399 |
+
"loss": 0.4485,
|
| 400 |
+
"learning_rate": 0.0002,
|
| 401 |
+
"epoch": 2.53,
|
| 402 |
+
"step": 48
|
| 403 |
+
},
|
| 404 |
+
{
|
| 405 |
+
"loss": 0.414,
|
| 406 |
+
"learning_rate": 0.0002,
|
| 407 |
+
"epoch": 2.58,
|
| 408 |
+
"step": 49
|
| 409 |
+
},
|
| 410 |
+
{
|
| 411 |
+
"loss": 0.5285,
|
| 412 |
+
"learning_rate": 0.0002,
|
| 413 |
+
"epoch": 2.63,
|
| 414 |
+
"step": 50
|
| 415 |
+
},
|
| 416 |
+
{
|
| 417 |
+
"eval_code_is_correct_loss": 0.4825914800167084,
|
| 418 |
+
"eval_code_is_correct_score": -0.15681606531143188,
|
| 419 |
+
"eval_code_is_correct_brier_score": 0.15681606531143188,
|
| 420 |
+
"eval_code_is_correct_average_probability": 0.6661742329597473,
|
| 421 |
+
"eval_code_is_correct_accuracy": 0.78,
|
| 422 |
+
"eval_code_is_correct_probabilities": [
|
| 423 |
+
0.9890788197517395,
|
| 424 |
+
0.5256270170211792,
|
| 425 |
+
0.31149938702583313,
|
| 426 |
+
0.4484615623950958,
|
| 427 |
+
0.6472581028938293,
|
| 428 |
+
0.990279495716095,
|
| 429 |
+
0.6154259443283081,
|
| 430 |
+
0.8025317192077637,
|
| 431 |
+
0.9357706308364868,
|
| 432 |
+
0.5503436923027039,
|
| 433 |
+
0.7849963903427124,
|
| 434 |
+
0.35905611515045166,
|
| 435 |
+
0.7320591807365417,
|
| 436 |
+
0.9779102206230164,
|
| 437 |
+
0.6362199187278748,
|
| 438 |
+
0.4611852467060089,
|
| 439 |
+
0.6441273093223572,
|
| 440 |
+
0.28485170006752014,
|
| 441 |
+
0.6162811517715454,
|
| 442 |
+
0.6565234661102295,
|
| 443 |
+
0.9911870360374451,
|
| 444 |
+
0.47038084268569946,
|
| 445 |
+
0.36016836762428284,
|
| 446 |
+
0.27217498421669006,
|
| 447 |
+
0.8941972851753235,
|
| 448 |
+
0.8256573677062988,
|
| 449 |
+
0.7179439663887024,
|
| 450 |
+
0.5055176019668579,
|
| 451 |
+
0.5404551029205322,
|
| 452 |
+
0.691025972366333,
|
| 453 |
+
0.48174765706062317,
|
| 454 |
+
0.3958812355995178,
|
| 455 |
+
0.9058158993721008,
|
| 456 |
+
0.3519555628299713,
|
| 457 |
+
0.7665671706199646,
|
| 458 |
+
0.9405277967453003,
|
| 459 |
+
0.5666556358337402,
|
| 460 |
+
0.6163994669914246,
|
| 461 |
+
0.7029475569725037,
|
| 462 |
+
0.4140672981739044,
|
| 463 |
+
0.6872601509094238,
|
| 464 |
+
0.47729596495628357,
|
| 465 |
+
0.9915251135826111,
|
| 466 |
+
0.6698000431060791,
|
| 467 |
+
0.5583893060684204,
|
| 468 |
+
0.5625171661376953,
|
| 469 |
+
0.3677026927471161,
|
| 470 |
+
0.9410431385040283,
|
| 471 |
+
0.8823411464691162,
|
| 472 |
+
0.7719405889511108,
|
| 473 |
+
0.8952547907829285,
|
| 474 |
+
0.9944015145301819,
|
| 475 |
+
0.6242450475692749,
|
| 476 |
+
0.7421666383743286,
|
| 477 |
+
0.6514956951141357,
|
| 478 |
+
0.5674803256988525,
|
| 479 |
+
0.888701319694519,
|
| 480 |
+
0.7480524778366089,
|
| 481 |
+
0.7139648199081421,
|
| 482 |
+
0.5017758011817932,
|
| 483 |
+
0.1396978348493576,
|
| 484 |
+
0.7807031273841858,
|
| 485 |
+
0.20994068682193756,
|
| 486 |
+
0.9944880604743958,
|
| 487 |
+
0.6406017541885376,
|
| 488 |
+
0.6515274047851562,
|
| 489 |
+
0.5916007161140442,
|
| 490 |
+
0.5801588296890259,
|
| 491 |
+
0.7971768975257874,
|
| 492 |
+
0.4683827757835388,
|
| 493 |
+
0.6826773285865784,
|
| 494 |
+
0.9399536848068237,
|
| 495 |
+
0.9587288498878479,
|
| 496 |
+
0.7912093997001648,
|
| 497 |
+
0.9399692416191101,
|
| 498 |
+
0.7925052642822266,
|
| 499 |
+
0.4648332893848419,
|
| 500 |
+
0.7946687936782837,
|
| 501 |
+
0.7067309021949768,
|
| 502 |
+
0.7728155851364136,
|
| 503 |
+
0.7523179054260254,
|
| 504 |
+
0.23098425567150116,
|
| 505 |
+
0.3329189121723175,
|
| 506 |
+
0.6994566321372986,
|
| 507 |
+
0.44642090797424316,
|
| 508 |
+
0.5221709609031677,
|
| 509 |
+
0.7311367988586426,
|
| 510 |
+
0.9462592601776123,
|
| 511 |
+
0.2204018384218216,
|
| 512 |
+
0.7329748868942261,
|
| 513 |
+
0.9443881511688232,
|
| 514 |
+
0.7437092661857605,
|
| 515 |
+
0.5737127065658569,
|
| 516 |
+
0.6452605724334717,
|
| 517 |
+
0.9964486360549927,
|
| 518 |
+
0.5256209373474121,
|
| 519 |
+
0.9053956270217896,
|
| 520 |
+
0.9367812275886536,
|
| 521 |
+
0.7426385879516602,
|
| 522 |
+
0.6719332337379456
|
| 523 |
+
],
|
| 524 |
+
"eval_code_is_correct_runtime": 89.0987,
|
| 525 |
+
"eval_code_is_correct_samples_per_second": 1.122,
|
| 526 |
+
"eval_code_is_correct_steps_per_second": 0.045,
|
| 527 |
+
"epoch": 2.63,
|
| 528 |
+
"step": 50
|
| 529 |
+
},
|
| 530 |
+
{
|
| 531 |
+
"loss": 0.3911,
|
| 532 |
+
"learning_rate": 0.0002,
|
| 533 |
+
"epoch": 2.68,
|
| 534 |
+
"step": 51
|
| 535 |
+
},
|
| 536 |
+
{
|
| 537 |
+
"loss": 0.4484,
|
| 538 |
+
"learning_rate": 0.0002,
|
| 539 |
+
"epoch": 2.74,
|
| 540 |
+
"step": 52
|
| 541 |
+
},
|
| 542 |
+
{
|
| 543 |
+
"loss": 0.4836,
|
| 544 |
+
"learning_rate": 0.0002,
|
| 545 |
+
"epoch": 2.79,
|
| 546 |
+
"step": 53
|
| 547 |
+
},
|
| 548 |
+
{
|
| 549 |
+
"loss": 0.2963,
|
| 550 |
+
"learning_rate": 0.0002,
|
| 551 |
+
"epoch": 2.84,
|
| 552 |
+
"step": 54
|
| 553 |
+
},
|
| 554 |
+
{
|
| 555 |
+
"loss": 0.3431,
|
| 556 |
+
"learning_rate": 0.0002,
|
| 557 |
+
"epoch": 2.89,
|
| 558 |
+
"step": 55
|
| 559 |
+
},
|
| 560 |
+
{
|
| 561 |
+
"loss": 0.4689,
|
| 562 |
+
"learning_rate": 0.0002,
|
| 563 |
+
"epoch": 2.95,
|
| 564 |
+
"step": 56
|
| 565 |
+
},
|
| 566 |
+
{
|
| 567 |
+
"loss": 0.3658,
|
| 568 |
+
"learning_rate": 0.0002,
|
| 569 |
+
"epoch": 3.0,
|
| 570 |
+
"step": 57
|
| 571 |
+
},
|
| 572 |
+
{
|
| 573 |
+
"loss": 0.2631,
|
| 574 |
+
"learning_rate": 0.0002,
|
| 575 |
+
"epoch": 3.05,
|
| 576 |
+
"step": 58
|
| 577 |
+
},
|
| 578 |
+
{
|
| 579 |
+
"loss": 0.219,
|
| 580 |
+
"learning_rate": 0.0002,
|
| 581 |
+
"epoch": 3.11,
|
| 582 |
+
"step": 59
|
| 583 |
+
},
|
| 584 |
+
{
|
| 585 |
+
"loss": 0.3664,
|
| 586 |
+
"learning_rate": 0.0002,
|
| 587 |
+
"epoch": 3.16,
|
| 588 |
+
"step": 60
|
| 589 |
+
},
|
| 590 |
+
{
|
| 591 |
+
"loss": 0.0843,
|
| 592 |
+
"learning_rate": 0.0002,
|
| 593 |
+
"epoch": 3.21,
|
| 594 |
+
"step": 61
|
| 595 |
+
},
|
| 596 |
+
{
|
| 597 |
+
"loss": 0.1934,
|
| 598 |
+
"learning_rate": 0.0002,
|
| 599 |
+
"epoch": 3.26,
|
| 600 |
+
"step": 62
|
| 601 |
+
},
|
| 602 |
+
{
|
| 603 |
+
"loss": 0.5592,
|
| 604 |
+
"learning_rate": 0.0002,
|
| 605 |
+
"epoch": 3.32,
|
| 606 |
+
"step": 63
|
| 607 |
+
},
|
| 608 |
+
{
|
| 609 |
+
"loss": 0.2187,
|
| 610 |
+
"learning_rate": 0.0002,
|
| 611 |
+
"epoch": 3.37,
|
| 612 |
+
"step": 64
|
| 613 |
+
},
|
| 614 |
+
{
|
| 615 |
+
"loss": 0.5812,
|
| 616 |
+
"learning_rate": 0.0002,
|
| 617 |
+
"epoch": 3.42,
|
| 618 |
+
"step": 65
|
| 619 |
+
},
|
| 620 |
+
{
|
| 621 |
+
"loss": 0.1651,
|
| 622 |
+
"learning_rate": 0.0002,
|
| 623 |
+
"epoch": 3.47,
|
| 624 |
+
"step": 66
|
| 625 |
+
},
|
| 626 |
+
{
|
| 627 |
+
"loss": 0.2347,
|
| 628 |
+
"learning_rate": 0.0002,
|
| 629 |
+
"epoch": 3.53,
|
| 630 |
+
"step": 67
|
| 631 |
+
},
|
| 632 |
+
{
|
| 633 |
+
"loss": 0.2858,
|
| 634 |
+
"learning_rate": 0.0002,
|
| 635 |
+
"epoch": 3.58,
|
| 636 |
+
"step": 68
|
| 637 |
+
},
|
| 638 |
+
{
|
| 639 |
+
"loss": 0.398,
|
| 640 |
+
"learning_rate": 0.0002,
|
| 641 |
+
"epoch": 3.63,
|
| 642 |
+
"step": 69
|
| 643 |
+
},
|
| 644 |
+
{
|
| 645 |
+
"loss": 0.2263,
|
| 646 |
+
"learning_rate": 0.0002,
|
| 647 |
+
"epoch": 3.68,
|
| 648 |
+
"step": 70
|
| 649 |
+
},
|
| 650 |
+
{
|
| 651 |
+
"loss": 0.1997,
|
| 652 |
+
"learning_rate": 0.0002,
|
| 653 |
+
"epoch": 3.74,
|
| 654 |
+
"step": 71
|
| 655 |
+
},
|
| 656 |
+
{
|
| 657 |
+
"loss": 0.1962,
|
| 658 |
+
"learning_rate": 0.0002,
|
| 659 |
+
"epoch": 3.79,
|
| 660 |
+
"step": 72
|
| 661 |
+
},
|
| 662 |
+
{
|
| 663 |
+
"loss": 0.3922,
|
| 664 |
+
"learning_rate": 0.0002,
|
| 665 |
+
"epoch": 3.84,
|
| 666 |
+
"step": 73
|
| 667 |
+
},
|
| 668 |
+
{
|
| 669 |
+
"loss": 0.2901,
|
| 670 |
+
"learning_rate": 0.0002,
|
| 671 |
+
"epoch": 3.89,
|
| 672 |
+
"step": 74
|
| 673 |
+
},
|
| 674 |
+
{
|
| 675 |
+
"loss": 0.354,
|
| 676 |
+
"learning_rate": 0.0002,
|
| 677 |
+
"epoch": 3.95,
|
| 678 |
+
"step": 75
|
| 679 |
+
},
|
| 680 |
+
{
|
| 681 |
+
"eval_code_is_correct_loss": 0.4227831959724426,
|
| 682 |
+
"eval_code_is_correct_score": -0.12974894046783447,
|
| 683 |
+
"eval_code_is_correct_brier_score": 0.12974894046783447,
|
| 684 |
+
"eval_code_is_correct_average_probability": 0.7597215175628662,
|
| 685 |
+
"eval_code_is_correct_accuracy": 0.81,
|
| 686 |
+
"eval_code_is_correct_probabilities": [
|
| 687 |
+
0.9988728165626526,
|
| 688 |
+
0.48829779028892517,
|
| 689 |
+
0.5296067595481873,
|
| 690 |
+
0.2780781388282776,
|
| 691 |
+
0.7249220609664917,
|
| 692 |
+
0.9996078610420227,
|
| 693 |
+
0.8340080976486206,
|
| 694 |
+
0.861427366733551,
|
| 695 |
+
0.9820781350135803,
|
| 696 |
+
0.785228431224823,
|
| 697 |
+
0.9509273171424866,
|
| 698 |
+
0.2166699916124344,
|
| 699 |
+
0.8465635180473328,
|
| 700 |
+
0.998379111289978,
|
| 701 |
+
0.5253419280052185,
|
| 702 |
+
0.34222710132598877,
|
| 703 |
+
0.9068326950073242,
|
| 704 |
+
0.1039603129029274,
|
| 705 |
+
0.953624963760376,
|
| 706 |
+
0.8484816551208496,
|
| 707 |
+
0.999100923538208,
|
| 708 |
+
0.7170456051826477,
|
| 709 |
+
0.4141126275062561,
|
| 710 |
+
0.08783219009637833,
|
| 711 |
+
0.9520941376686096,
|
| 712 |
+
0.9142940044403076,
|
| 713 |
+
0.922050416469574,
|
| 714 |
+
0.7979483604431152,
|
| 715 |
+
0.7149856686592102,
|
| 716 |
+
0.8099195957183838,
|
| 717 |
+
0.6641011238098145,
|
| 718 |
+
0.731330156326294,
|
| 719 |
+
0.987343430519104,
|
| 720 |
+
0.6779031157493591,
|
| 721 |
+
0.9263503551483154,
|
| 722 |
+
0.9907992482185364,
|
| 723 |
+
0.7801071405410767,
|
| 724 |
+
0.8316705226898193,
|
| 725 |
+
0.7261466979980469,
|
| 726 |
+
0.15835784375667572,
|
| 727 |
+
0.7666859030723572,
|
| 728 |
+
0.3682948648929596,
|
| 729 |
+
0.999472439289093,
|
| 730 |
+
0.8673612475395203,
|
| 731 |
+
0.7001268863677979,
|
| 732 |
+
0.7785540223121643,
|
| 733 |
+
0.3994494676589966,
|
| 734 |
+
0.9955601692199707,
|
| 735 |
+
0.9255762696266174,
|
| 736 |
+
0.9511540532112122,
|
| 737 |
+
0.9472480416297913,
|
| 738 |
+
0.9996776580810547,
|
| 739 |
+
0.7907305955886841,
|
| 740 |
+
0.9202771782875061,
|
| 741 |
+
0.7952592968940735,
|
| 742 |
+
0.9026097059249878,
|
| 743 |
+
0.971696138381958,
|
| 744 |
+
0.9698152542114258,
|
| 745 |
+
0.8726643919944763,
|
| 746 |
+
0.9103471636772156,
|
| 747 |
+
0.06682927906513214,
|
| 748 |
+
0.983897864818573,
|
| 749 |
+
0.4112030267715454,
|
| 750 |
+
0.9997578263282776,
|
| 751 |
+
0.8332673907279968,
|
| 752 |
+
0.9109005331993103,
|
| 753 |
+
0.8171624541282654,
|
| 754 |
+
0.6875969767570496,
|
| 755 |
+
0.9541371464729309,
|
| 756 |
+
0.8013278245925903,
|
| 757 |
+
0.8411256074905396,
|
| 758 |
+
0.9754989147186279,
|
| 759 |
+
0.9865782856941223,
|
| 760 |
+
0.9783108830451965,
|
| 761 |
+
0.9962783455848694,
|
| 762 |
+
0.9778187274932861,
|
| 763 |
+
0.43399563431739807,
|
| 764 |
+
0.9774395227432251,
|
| 765 |
+
0.9118335843086243,
|
| 766 |
+
0.4630809426307678,
|
| 767 |
+
0.9608380198478699,
|
| 768 |
+
0.13883967697620392,
|
| 769 |
+
0.10208643227815628,
|
| 770 |
+
0.5756421685218811,
|
| 771 |
+
0.5382805466651917,
|
| 772 |
+
0.6014205813407898,
|
| 773 |
+
0.9483709931373596,
|
| 774 |
+
0.997037410736084,
|
| 775 |
+
0.027208847925066948,
|
| 776 |
+
0.951516330242157,
|
| 777 |
+
0.9978541731834412,
|
| 778 |
+
0.3260488510131836,
|
| 779 |
+
0.40766820311546326,
|
| 780 |
+
0.8333786725997925,
|
| 781 |
+
0.9996563196182251,
|
| 782 |
+
0.9164862632751465,
|
| 783 |
+
0.9894892573356628,
|
| 784 |
+
0.9954167604446411,
|
| 785 |
+
0.9306528568267822,
|
| 786 |
+
0.9150286316871643
|
| 787 |
+
],
|
| 788 |
+
"eval_code_is_correct_runtime": 89.1528,
|
| 789 |
+
"eval_code_is_correct_samples_per_second": 1.122,
|
| 790 |
+
"eval_code_is_correct_steps_per_second": 0.045,
|
| 791 |
+
"epoch": 3.95,
|
| 792 |
+
"step": 75
|
| 793 |
+
},
|
| 794 |
+
{
|
| 795 |
+
"loss": 0.1587,
|
| 796 |
+
"learning_rate": 0.0002,
|
| 797 |
+
"epoch": 4.0,
|
| 798 |
+
"step": 76
|
| 799 |
+
},
|
| 800 |
+
{
|
| 801 |
+
"loss": 0.1926,
|
| 802 |
+
"learning_rate": 0.0002,
|
| 803 |
+
"epoch": 4.05,
|
| 804 |
+
"step": 77
|
| 805 |
+
},
|
| 806 |
+
{
|
| 807 |
+
"loss": 0.1286,
|
| 808 |
+
"learning_rate": 0.0002,
|
| 809 |
+
"epoch": 4.11,
|
| 810 |
+
"step": 78
|
| 811 |
+
},
|
| 812 |
+
{
|
| 813 |
+
"loss": 0.2267,
|
| 814 |
+
"learning_rate": 0.0002,
|
| 815 |
+
"epoch": 4.16,
|
| 816 |
+
"step": 79
|
| 817 |
+
},
|
| 818 |
+
{
|
| 819 |
+
"loss": 0.1104,
|
| 820 |
+
"learning_rate": 0.0002,
|
| 821 |
+
"epoch": 4.21,
|
| 822 |
+
"step": 80
|
| 823 |
+
},
|
| 824 |
+
{
|
| 825 |
+
"loss": 0.1765,
|
| 826 |
+
"learning_rate": 0.0002,
|
| 827 |
+
"epoch": 4.26,
|
| 828 |
+
"step": 81
|
| 829 |
+
},
|
| 830 |
+
{
|
| 831 |
+
"loss": 0.3804,
|
| 832 |
+
"learning_rate": 0.0002,
|
| 833 |
+
"epoch": 4.32,
|
| 834 |
+
"step": 82
|
| 835 |
+
},
|
| 836 |
+
{
|
| 837 |
+
"loss": 0.1671,
|
| 838 |
+
"learning_rate": 0.0002,
|
| 839 |
+
"epoch": 4.37,
|
| 840 |
+
"step": 83
|
| 841 |
+
},
|
| 842 |
+
{
|
| 843 |
+
"loss": 0.1771,
|
| 844 |
+
"learning_rate": 0.0002,
|
| 845 |
+
"epoch": 4.42,
|
| 846 |
+
"step": 84
|
| 847 |
+
},
|
| 848 |
+
{
|
| 849 |
+
"loss": 0.1781,
|
| 850 |
+
"learning_rate": 0.0002,
|
| 851 |
+
"epoch": 4.47,
|
| 852 |
+
"step": 85
|
| 853 |
+
},
|
| 854 |
+
{
|
| 855 |
+
"loss": 0.0831,
|
| 856 |
+
"learning_rate": 0.0002,
|
| 857 |
+
"epoch": 4.53,
|
| 858 |
+
"step": 86
|
| 859 |
+
},
|
| 860 |
+
{
|
| 861 |
+
"loss": 0.1021,
|
| 862 |
+
"learning_rate": 0.0002,
|
| 863 |
+
"epoch": 4.58,
|
| 864 |
+
"step": 87
|
| 865 |
+
},
|
| 866 |
+
{
|
| 867 |
+
"loss": 0.3782,
|
| 868 |
+
"learning_rate": 0.0002,
|
| 869 |
+
"epoch": 4.63,
|
| 870 |
+
"step": 88
|
| 871 |
+
},
|
| 872 |
+
{
|
| 873 |
+
"loss": 0.582,
|
| 874 |
+
"learning_rate": 0.0002,
|
| 875 |
+
"epoch": 4.68,
|
| 876 |
+
"step": 89
|
| 877 |
+
},
|
| 878 |
+
{
|
| 879 |
+
"loss": 0.5176,
|
| 880 |
+
"learning_rate": 0.0002,
|
| 881 |
+
"epoch": 4.74,
|
| 882 |
+
"step": 90
|
| 883 |
+
},
|
| 884 |
+
{
|
| 885 |
+
"loss": 0.4676,
|
| 886 |
+
"learning_rate": 0.0002,
|
| 887 |
+
"epoch": 4.79,
|
| 888 |
+
"step": 91
|
| 889 |
+
},
|
| 890 |
+
{
|
| 891 |
+
"loss": 0.1139,
|
| 892 |
+
"learning_rate": 0.0002,
|
| 893 |
+
"epoch": 4.84,
|
| 894 |
+
"step": 92
|
| 895 |
+
},
|
| 896 |
+
{
|
| 897 |
+
"loss": 0.1754,
|
| 898 |
+
"learning_rate": 0.0002,
|
| 899 |
+
"epoch": 4.89,
|
| 900 |
+
"step": 93
|
| 901 |
+
},
|
| 902 |
+
{
|
| 903 |
+
"loss": 0.1578,
|
| 904 |
+
"learning_rate": 0.0002,
|
| 905 |
+
"epoch": 4.95,
|
| 906 |
+
"step": 94
|
| 907 |
+
},
|
| 908 |
+
{
|
| 909 |
+
"loss": 0.2121,
|
| 910 |
+
"learning_rate": 0.0002,
|
| 911 |
+
"epoch": 5.0,
|
| 912 |
+
"step": 95
|
| 913 |
+
},
|
| 914 |
+
{
|
| 915 |
+
"loss": 0.1099,
|
| 916 |
+
"learning_rate": 0.0002,
|
| 917 |
+
"epoch": 5.05,
|
| 918 |
+
"step": 96
|
| 919 |
+
},
|
| 920 |
+
{
|
| 921 |
+
"loss": 0.1318,
|
| 922 |
+
"learning_rate": 0.0002,
|
| 923 |
+
"epoch": 5.11,
|
| 924 |
+
"step": 97
|
| 925 |
+
},
|
| 926 |
+
{
|
| 927 |
+
"loss": 0.1261,
|
| 928 |
+
"learning_rate": 0.0002,
|
| 929 |
+
"epoch": 5.16,
|
| 930 |
+
"step": 98
|
| 931 |
+
},
|
| 932 |
+
{
|
| 933 |
+
"loss": 0.1414,
|
| 934 |
+
"learning_rate": 0.0002,
|
| 935 |
+
"epoch": 5.21,
|
| 936 |
+
"step": 99
|
| 937 |
+
},
|
| 938 |
+
{
|
| 939 |
+
"loss": 0.0838,
|
| 940 |
+
"learning_rate": 0.0002,
|
| 941 |
+
"epoch": 5.26,
|
| 942 |
+
"step": 100
|
| 943 |
+
},
|
| 944 |
+
{
|
| 945 |
+
"eval_code_is_correct_loss": 0.6910045146942139,
|
| 946 |
+
"eval_code_is_correct_score": -0.18250787258148193,
|
| 947 |
+
"eval_code_is_correct_brier_score": 0.18250787258148193,
|
| 948 |
+
"eval_code_is_correct_average_probability": 0.7548456788063049,
|
| 949 |
+
"eval_code_is_correct_accuracy": 0.77,
|
| 950 |
+
"eval_code_is_correct_probabilities": [
|
| 951 |
+
0.9995959401130676,
|
| 952 |
+
0.0028756344690918922,
|
| 953 |
+
0.5750472545623779,
|
| 954 |
+
0.3717687427997589,
|
| 955 |
+
0.9003951549530029,
|
| 956 |
+
0.9999061822891235,
|
| 957 |
+
0.34974727034568787,
|
| 958 |
+
0.9861345887184143,
|
| 959 |
+
0.997637152671814,
|
| 960 |
+
0.986896812915802,
|
| 961 |
+
0.997134804725647,
|
| 962 |
+
0.12967519462108612,
|
| 963 |
+
0.9531847238540649,
|
| 964 |
+
0.9992759823799133,
|
| 965 |
+
0.20013374090194702,
|
| 966 |
+
0.06892868876457214,
|
| 967 |
+
0.9374101758003235,
|
| 968 |
+
0.021766141057014465,
|
| 969 |
+
0.7759655117988586,
|
| 970 |
+
0.87876957654953,
|
| 971 |
+
0.9996174573898315,
|
| 972 |
+
0.4874546229839325,
|
| 973 |
+
0.05481430143117905,
|
| 974 |
+
0.07372044026851654,
|
| 975 |
+
0.9974690675735474,
|
| 976 |
+
0.9834303855895996,
|
| 977 |
+
0.9566057920455933,
|
| 978 |
+
0.9595711827278137,
|
| 979 |
+
0.8324235081672668,
|
| 980 |
+
0.9975442290306091,
|
| 981 |
+
0.6663374304771423,
|
| 982 |
+
0.9153403043746948,
|
| 983 |
+
0.9981799125671387,
|
| 984 |
+
0.21779882907867432,
|
| 985 |
+
0.9679686427116394,
|
| 986 |
+
0.9959613680839539,
|
| 987 |
+
0.6366267800331116,
|
| 988 |
+
0.9374737739562988,
|
| 989 |
+
0.5792506337165833,
|
| 990 |
+
0.071602463722229,
|
| 991 |
+
0.14550523459911346,
|
| 992 |
+
0.08400178700685501,
|
| 993 |
+
0.999697208404541,
|
| 994 |
+
0.9761826395988464,
|
| 995 |
+
0.7455314993858337,
|
| 996 |
+
0.23751415312290192,
|
| 997 |
+
0.13082778453826904,
|
| 998 |
+
0.9988062381744385,
|
| 999 |
+
0.9915608763694763,
|
| 1000 |
+
0.9885241389274597,
|
| 1001 |
+
0.9979588985443115,
|
| 1002 |
+
0.9998948574066162,
|
| 1003 |
+
0.980595588684082,
|
| 1004 |
+
0.9910929799079895,
|
| 1005 |
+
0.8016942739486694,
|
| 1006 |
+
0.9494630694389343,
|
| 1007 |
+
0.9971957206726074,
|
| 1008 |
+
0.9917238354682922,
|
| 1009 |
+
0.9877293109893799,
|
| 1010 |
+
0.7314670085906982,
|
| 1011 |
+
0.012357236817479134,
|
| 1012 |
+
0.9953846335411072,
|
| 1013 |
+
0.8007073998451233,
|
| 1014 |
+
0.9999083280563354,
|
| 1015 |
+
0.8738554120063782,
|
| 1016 |
+
0.9894973039627075,
|
| 1017 |
+
0.9153193235397339,
|
| 1018 |
+
0.9965702295303345,
|
| 1019 |
+
0.9988541603088379,
|
| 1020 |
+
0.9915642738342285,
|
| 1021 |
+
0.9359388947486877,
|
| 1022 |
+
0.9891656041145325,
|
| 1023 |
+
0.996810257434845,
|
| 1024 |
+
0.9968699812889099,
|
| 1025 |
+
0.9994422793388367,
|
| 1026 |
+
0.9953920841217041,
|
| 1027 |
+
0.015405597165226936,
|
| 1028 |
+
0.9934167861938477,
|
| 1029 |
+
0.966199517250061,
|
| 1030 |
+
0.8508797287940979,
|
| 1031 |
+
0.9916513562202454,
|
| 1032 |
+
0.8259217739105225,
|
| 1033 |
+
0.0839756578207016,
|
| 1034 |
+
0.72227942943573,
|
| 1035 |
+
0.4742227792739868,
|
| 1036 |
+
0.007106057368218899,
|
| 1037 |
+
0.9803650379180908,
|
| 1038 |
+
0.9996927976608276,
|
| 1039 |
+
0.0069855921901762486,
|
| 1040 |
+
0.9928151369094849,
|
| 1041 |
+
0.9996411800384521,
|
| 1042 |
+
0.1924436241388321,
|
| 1043 |
+
0.8279777765274048,
|
| 1044 |
+
0.9689691066741943,
|
| 1045 |
+
0.9998074173927307,
|
| 1046 |
+
0.994343638420105,
|
| 1047 |
+
0.9985198378562927,
|
| 1048 |
+
0.9983432292938232,
|
| 1049 |
+
0.9436729550361633,
|
| 1050 |
+
0.9738808274269104
|
| 1051 |
+
],
|
| 1052 |
+
"eval_code_is_correct_runtime": 89.0718,
|
| 1053 |
+
"eval_code_is_correct_samples_per_second": 1.123,
|
| 1054 |
+
"eval_code_is_correct_steps_per_second": 0.045,
|
| 1055 |
+
"epoch": 5.26,
|
| 1056 |
+
"step": 100
|
| 1057 |
+
},
|
| 1058 |
+
{
|
| 1059 |
+
"train_runtime": 6498.1873,
|
| 1060 |
+
"train_samples_per_second": 0.492,
|
| 1061 |
+
"train_steps_per_second": 0.015,
|
| 1062 |
+
"total_flos": 0.0,
|
| 1063 |
+
"train_loss": 0.443891556635499,
|
| 1064 |
+
"epoch": 5.26,
|
| 1065 |
+
"step": 100
|
| 1066 |
+
}
|
| 1067 |
+
]
|