lch-qiwu-checkpoint-5000
Browse files- lch-qiwu-checkpoint-5000/config.json +187 -0
- lch-qiwu-checkpoint-5000/dataset_statistics.json +290 -0
- lch-qiwu-checkpoint-5000/model.safetensors +3 -0
- lch-qiwu-checkpoint-5000/optimizer.pt +3 -0
- lch-qiwu-checkpoint-5000/rng_state.pth +3 -0
- lch-qiwu-checkpoint-5000/scheduler.pt +3 -0
- lch-qiwu-checkpoint-5000/trainer_state.json +784 -0
- lch-qiwu-checkpoint-5000/training_args.bin +3 -0
lch-qiwu-checkpoint-5000/config.json
ADDED
|
@@ -0,0 +1,187 @@
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
| 1 |
+
{
|
| 2 |
+
"batch_size": 8,
|
| 3 |
+
"buffer_size": 64,
|
| 4 |
+
"eval_mix": "qiwu_test",
|
| 5 |
+
"frozen_keys": [
|
| 6 |
+
"*hf_model*"
|
| 7 |
+
],
|
| 8 |
+
"lora": false,
|
| 9 |
+
"lora_config": {
|
| 10 |
+
"bias": "none",
|
| 11 |
+
"lora_alpha": 16,
|
| 12 |
+
"lora_dropout": 0.05,
|
| 13 |
+
"r": 8
|
| 14 |
+
},
|
| 15 |
+
"model": {
|
| 16 |
+
"heads": {
|
| 17 |
+
"action": {
|
| 18 |
+
"args": [],
|
| 19 |
+
"kwargs": {
|
| 20 |
+
"action_dim": 7,
|
| 21 |
+
"action_horizon": 4,
|
| 22 |
+
"dropout_rate": 0.0,
|
| 23 |
+
"n_diffusion_samples": 1,
|
| 24 |
+
"readout_key": "readout_action",
|
| 25 |
+
"token_embedding_size": 768,
|
| 26 |
+
"use_map": false
|
| 27 |
+
},
|
| 28 |
+
"module": "octo.model.components.action_heads",
|
| 29 |
+
"name": "DiffusionActionHead"
|
| 30 |
+
}
|
| 31 |
+
},
|
| 32 |
+
"max_horizon": 10,
|
| 33 |
+
"observation_tokenizers": {
|
| 34 |
+
"primary": {
|
| 35 |
+
"args": [],
|
| 36 |
+
"kwargs": {
|
| 37 |
+
"encoder": {
|
| 38 |
+
"args": [],
|
| 39 |
+
"kwargs": {
|
| 40 |
+
"in_features": 6
|
| 41 |
+
},
|
| 42 |
+
"module": "octo.model.components.vit_encoders",
|
| 43 |
+
"name": "SmallStem16"
|
| 44 |
+
},
|
| 45 |
+
"obs_stack_keys": [
|
| 46 |
+
"image_primary"
|
| 47 |
+
],
|
| 48 |
+
"task_stack_keys": [
|
| 49 |
+
"image_primary"
|
| 50 |
+
]
|
| 51 |
+
},
|
| 52 |
+
"module": "octo.model.components.tokenizers",
|
| 53 |
+
"name": "ImageTokenizer"
|
| 54 |
+
},
|
| 55 |
+
"state": {
|
| 56 |
+
"args": [],
|
| 57 |
+
"kwargs": {
|
| 58 |
+
"bin_type": "normal",
|
| 59 |
+
"high": 2.0,
|
| 60 |
+
"low": -2.0,
|
| 61 |
+
"n_bins": 256,
|
| 62 |
+
"obs_keys": [
|
| 63 |
+
"state"
|
| 64 |
+
]
|
| 65 |
+
},
|
| 66 |
+
"module": "model.octo.components.tokenizers",
|
| 67 |
+
"name": "LowdimObsTokenizer"
|
| 68 |
+
},
|
| 69 |
+
"wrist": {
|
| 70 |
+
"args": [],
|
| 71 |
+
"kwargs": {
|
| 72 |
+
"encoder": {
|
| 73 |
+
"args": [],
|
| 74 |
+
"kwargs": {
|
| 75 |
+
"in_features": 6
|
| 76 |
+
},
|
| 77 |
+
"module": "octo.model.components.vit_encoders",
|
| 78 |
+
"name": "SmallStem16"
|
| 79 |
+
},
|
| 80 |
+
"obs_stack_keys": [
|
| 81 |
+
"image_wrist"
|
| 82 |
+
],
|
| 83 |
+
"task_stack_keys": [
|
| 84 |
+
"image_wrist"
|
| 85 |
+
]
|
| 86 |
+
},
|
| 87 |
+
"module": "octo.model.components.tokenizers",
|
| 88 |
+
"name": "ImageTokenizer"
|
| 89 |
+
}
|
| 90 |
+
},
|
| 91 |
+
"readouts": {
|
| 92 |
+
"action": 1
|
| 93 |
+
},
|
| 94 |
+
"repeat_task_tokens": true,
|
| 95 |
+
"task_tokenizers": {
|
| 96 |
+
"language": {
|
| 97 |
+
"args": [],
|
| 98 |
+
"kwargs": {
|
| 99 |
+
"encoder": "t5-base",
|
| 100 |
+
"finetune_encoder": false
|
| 101 |
+
},
|
| 102 |
+
"module": "octo.model.components.tokenizers",
|
| 103 |
+
"name": "LanguageTokenizer"
|
| 104 |
+
}
|
| 105 |
+
},
|
| 106 |
+
"token_embedding_size": 768,
|
| 107 |
+
"transformer_kwargs": {
|
| 108 |
+
"add_position_embedding": false,
|
| 109 |
+
"attention_dropout_rate": 0.0,
|
| 110 |
+
"dropout_rate": 0.0,
|
| 111 |
+
"mlp_dim": 3072,
|
| 112 |
+
"num_attention_heads": 12,
|
| 113 |
+
"num_layers": 12
|
| 114 |
+
},
|
| 115 |
+
"use_correct_attention": true
|
| 116 |
+
},
|
| 117 |
+
"no_state": false,
|
| 118 |
+
"obs_token_nums": {
|
| 119 |
+
"primary": 256,
|
| 120 |
+
"state": 7,
|
| 121 |
+
"wrist": 64
|
| 122 |
+
},
|
| 123 |
+
"output_dir": "qiwu_checkpoints_all_data",
|
| 124 |
+
"pretrained": {
|
| 125 |
+
"kwargs": {
|
| 126 |
+
"subpath": "oxe-g2-checkpoint-300000"
|
| 127 |
+
},
|
| 128 |
+
"model": "hf://chuanmew/octo_torch"
|
| 129 |
+
},
|
| 130 |
+
"run_name": "qiwu",
|
| 131 |
+
"sampler_num_samples": 2560000,
|
| 132 |
+
"seed": 2,
|
| 133 |
+
"subsample_length": 99999,
|
| 134 |
+
"text_processor": {
|
| 135 |
+
"args": [],
|
| 136 |
+
"kwargs": {
|
| 137 |
+
"encode_with_model": false,
|
| 138 |
+
"tokenizer_kwargs": {
|
| 139 |
+
"max_length": 16,
|
| 140 |
+
"padding": "max_length",
|
| 141 |
+
"return_tensors": "np",
|
| 142 |
+
"truncation": true
|
| 143 |
+
},
|
| 144 |
+
"tokenizer_name": "t5-base"
|
| 145 |
+
},
|
| 146 |
+
"module": "octo.components.text_processing",
|
| 147 |
+
"name": "HFTokenizer"
|
| 148 |
+
},
|
| 149 |
+
"train_mix": "all_qiwu_merges",
|
| 150 |
+
"training_arguments": {
|
| 151 |
+
"bf16": true,
|
| 152 |
+
"dataloader_pin_memory": true,
|
| 153 |
+
"ddp_find_unused_parameters": true,
|
| 154 |
+
"eval_steps": 100,
|
| 155 |
+
"gradient_accumulation_steps": 1,
|
| 156 |
+
"learning_rate": 3e-05,
|
| 157 |
+
"logging_nan_inf_filter": false,
|
| 158 |
+
"logging_steps": 100,
|
| 159 |
+
"lr_scheduler_kwargs": {
|
| 160 |
+
"decay_type": "cosine",
|
| 161 |
+
"min_lr_ratio": 0.05,
|
| 162 |
+
"num_decay_steps": 0
|
| 163 |
+
},
|
| 164 |
+
"lr_scheduler_type": "warmup_stable_decay",
|
| 165 |
+
"max_grad_norm": 1.0,
|
| 166 |
+
"max_steps": 5000,
|
| 167 |
+
"optim": "adamw_torch_fused",
|
| 168 |
+
"per_device_eval_batch_size": 1,
|
| 169 |
+
"per_device_train_batch_size": 1,
|
| 170 |
+
"report_to": "wandb",
|
| 171 |
+
"save_safetensors": true,
|
| 172 |
+
"save_steps": 500,
|
| 173 |
+
"torch_compile": true,
|
| 174 |
+
"warmup_steps": 100,
|
| 175 |
+
"weight_decay": 0.1
|
| 176 |
+
},
|
| 177 |
+
"training_keys": [
|
| 178 |
+
"*action*",
|
| 179 |
+
"*observation_tokenizers*",
|
| 180 |
+
"*transformer.transformer*",
|
| 181 |
+
"*task_projections*",
|
| 182 |
+
"*obs_projections*",
|
| 183 |
+
"*readout_embeddings*",
|
| 184 |
+
"*task_pos_embeddings*",
|
| 185 |
+
"*obs_pos_embeddings*"
|
| 186 |
+
]
|
| 187 |
+
}
|
lch-qiwu-checkpoint-5000/dataset_statistics.json
ADDED
|
@@ -0,0 +1,290 @@
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
| 1 |
+
[
|
| 2 |
+
{
|
| 3 |
+
"action": {
|
| 4 |
+
"mean": [
|
| 5 |
+
-0.0021207339596003294,
|
| 6 |
+
0.00183492514770478,
|
| 7 |
+
0.0035495238844305277,
|
| 8 |
+
-4.220907612761948e-06,
|
| 9 |
+
-0.0055978186428546906,
|
| 10 |
+
-0.006060932297259569,
|
| 11 |
+
0.5278085470199585
|
| 12 |
+
],
|
| 13 |
+
"std": [
|
| 14 |
+
0.03802286833524704,
|
| 15 |
+
0.021030722185969353,
|
| 16 |
+
0.040165603160858154,
|
| 17 |
+
0.08549340814352036,
|
| 18 |
+
0.0911325067281723,
|
| 19 |
+
0.1323363482952118,
|
| 20 |
+
0.4878678619861603
|
| 21 |
+
],
|
| 22 |
+
"min": [
|
| 23 |
+
-0.264305979013443,
|
| 24 |
+
-0.11698240041732788,
|
| 25 |
+
-0.12788072228431702,
|
| 26 |
+
-0.7576491832733154,
|
| 27 |
+
-0.708380937576294,
|
| 28 |
+
-1.3020894527435303,
|
| 29 |
+
0.0
|
| 30 |
+
],
|
| 31 |
+
"max": [
|
| 32 |
+
0.17467179894447327,
|
| 33 |
+
0.1660223752260208,
|
| 34 |
+
0.18162232637405396,
|
| 35 |
+
0.8727858066558838,
|
| 36 |
+
0.8210165500640869,
|
| 37 |
+
1.276663064956665,
|
| 38 |
+
1.0
|
| 39 |
+
],
|
| 40 |
+
"mask": [
|
| 41 |
+
true,
|
| 42 |
+
true,
|
| 43 |
+
true,
|
| 44 |
+
true,
|
| 45 |
+
true,
|
| 46 |
+
true,
|
| 47 |
+
false
|
| 48 |
+
]
|
| 49 |
+
},
|
| 50 |
+
"state": {
|
| 51 |
+
"mean": [
|
| 52 |
+
-0.8593263626098633,
|
| 53 |
+
-1.336264729499817,
|
| 54 |
+
-1.0609148740768433,
|
| 55 |
+
-0.697239875793457,
|
| 56 |
+
-1.224779486656189,
|
| 57 |
+
0.2995929419994354,
|
| 58 |
+
0.5275121331214905
|
| 59 |
+
],
|
| 60 |
+
"std": [
|
| 61 |
+
0.20902122557163239,
|
| 62 |
+
0.35590317845344543,
|
| 63 |
+
0.4969709813594818,
|
| 64 |
+
0.21033141016960144,
|
| 65 |
+
0.2267223745584488,
|
| 66 |
+
0.4203825294971466,
|
| 67 |
+
0.48849955201148987
|
| 68 |
+
],
|
| 69 |
+
"min": [
|
| 70 |
+
-1.649301290512085,
|
| 71 |
+
-2.269015312194824,
|
| 72 |
+
-2.3562119007110596,
|
| 73 |
+
-1.5550708770751953,
|
| 74 |
+
-2.184192180633545,
|
| 75 |
+
-1.577620506286621,
|
| 76 |
+
0.0
|
| 77 |
+
],
|
| 78 |
+
"max": [
|
| 79 |
+
-0.257505863904953,
|
| 80 |
+
-0.13100440800189972,
|
| 81 |
+
0.5860641002655029,
|
| 82 |
+
-0.1613033264875412,
|
| 83 |
+
-0.5851041674613953,
|
| 84 |
+
1.552086353302002,
|
| 85 |
+
1.0
|
| 86 |
+
],
|
| 87 |
+
"mask": [
|
| 88 |
+
true,
|
| 89 |
+
true,
|
| 90 |
+
true,
|
| 91 |
+
true,
|
| 92 |
+
true,
|
| 93 |
+
true,
|
| 94 |
+
false
|
| 95 |
+
]
|
| 96 |
+
}
|
| 97 |
+
},
|
| 98 |
+
{
|
| 99 |
+
"action": {
|
| 100 |
+
"mean": [
|
| 101 |
+
0.002793338615447283,
|
| 102 |
+
0.0017030449816957116,
|
| 103 |
+
-0.0009504178306087852,
|
| 104 |
+
0.0014839255018159747,
|
| 105 |
+
0.0010163659462705255,
|
| 106 |
+
0.01281216274946928,
|
| 107 |
+
0.4567633867263794
|
| 108 |
+
],
|
| 109 |
+
"std": [
|
| 110 |
+
0.03352493420243263,
|
| 111 |
+
0.01704518496990204,
|
| 112 |
+
0.03547774627804756,
|
| 113 |
+
0.06987040489912033,
|
| 114 |
+
0.06841710209846497,
|
| 115 |
+
0.0847349762916565,
|
| 116 |
+
0.4794786274433136
|
| 117 |
+
],
|
| 118 |
+
"min": [
|
| 119 |
+
-0.24331116676330566,
|
| 120 |
+
-0.1301804780960083,
|
| 121 |
+
-0.15468895435333252,
|
| 122 |
+
-0.5852863788604736,
|
| 123 |
+
-0.5852053165435791,
|
| 124 |
+
-0.9242353439331055,
|
| 125 |
+
0.0
|
| 126 |
+
],
|
| 127 |
+
"max": [
|
| 128 |
+
0.1788746863603592,
|
| 129 |
+
0.13782057166099548,
|
| 130 |
+
0.203914076089859,
|
| 131 |
+
0.8304381370544434,
|
| 132 |
+
0.595897912979126,
|
| 133 |
+
1.1058757305145264,
|
| 134 |
+
1.0
|
| 135 |
+
],
|
| 136 |
+
"mask": [
|
| 137 |
+
true,
|
| 138 |
+
true,
|
| 139 |
+
true,
|
| 140 |
+
true,
|
| 141 |
+
true,
|
| 142 |
+
true,
|
| 143 |
+
false
|
| 144 |
+
]
|
| 145 |
+
},
|
| 146 |
+
"state": {
|
| 147 |
+
"mean": [
|
| 148 |
+
-1.1439815759658813,
|
| 149 |
+
-1.0255025625228882,
|
| 150 |
+
-1.3527690172195435,
|
| 151 |
+
-0.519981324672699,
|
| 152 |
+
-1.0002315044403076,
|
| 153 |
+
0.1325121521949768,
|
| 154 |
+
0.45643800497055054
|
| 155 |
+
],
|
| 156 |
+
"std": [
|
| 157 |
+
0.20416870713233948,
|
| 158 |
+
0.28054729104042053,
|
| 159 |
+
0.24587351083755493,
|
| 160 |
+
0.25086745619773865,
|
| 161 |
+
0.23796093463897705,
|
| 162 |
+
0.41082870960235596,
|
| 163 |
+
0.4796474575996399
|
| 164 |
+
],
|
| 165 |
+
"min": [
|
| 166 |
+
-1.6860579252243042,
|
| 167 |
+
-1.752921462059021,
|
| 168 |
+
-2.0715138912200928,
|
| 169 |
+
-1.61619234085083,
|
| 170 |
+
-1.7697288990020752,
|
| 171 |
+
-1.2800593376159668,
|
| 172 |
+
0.0
|
| 173 |
+
],
|
| 174 |
+
"max": [
|
| 175 |
+
-0.478167861700058,
|
| 176 |
+
-0.07618362456560135,
|
| 177 |
+
-0.47656214237213135,
|
| 178 |
+
0.3251723051071167,
|
| 179 |
+
-0.23657937347888947,
|
| 180 |
+
1.8649541139602661,
|
| 181 |
+
1.0
|
| 182 |
+
],
|
| 183 |
+
"mask": [
|
| 184 |
+
true,
|
| 185 |
+
true,
|
| 186 |
+
true,
|
| 187 |
+
true,
|
| 188 |
+
true,
|
| 189 |
+
true,
|
| 190 |
+
false
|
| 191 |
+
]
|
| 192 |
+
}
|
| 193 |
+
},
|
| 194 |
+
{
|
| 195 |
+
"action": {
|
| 196 |
+
"mean": [
|
| 197 |
+
-0.001587849808856845,
|
| 198 |
+
0.0010738142300397158,
|
| 199 |
+
-0.000697988027241081,
|
| 200 |
+
-0.006544447969645262,
|
| 201 |
+
0.009556379169225693,
|
| 202 |
+
0.020673424005508423,
|
| 203 |
+
0.45607081055641174
|
| 204 |
+
],
|
| 205 |
+
"std": [
|
| 206 |
+
0.03518082574009895,
|
| 207 |
+
0.02438654750585556,
|
| 208 |
+
0.0410635769367218,
|
| 209 |
+
0.11703108996152878,
|
| 210 |
+
0.10797987133264542,
|
| 211 |
+
0.17201891541481018,
|
| 212 |
+
0.46286532282829285
|
| 213 |
+
],
|
| 214 |
+
"min": [
|
| 215 |
+
-0.2540467381477356,
|
| 216 |
+
-0.16329476237297058,
|
| 217 |
+
-0.19035658240318298,
|
| 218 |
+
-1.4179747104644775,
|
| 219 |
+
-0.8620505332946777,
|
| 220 |
+
-2.1886229515075684,
|
| 221 |
+
0.0
|
| 222 |
+
],
|
| 223 |
+
"max": [
|
| 224 |
+
0.18435201048851013,
|
| 225 |
+
0.15960073471069336,
|
| 226 |
+
0.2124226987361908,
|
| 227 |
+
1.345801591873169,
|
| 228 |
+
1.2207748889923096,
|
| 229 |
+
2.1311380863189697,
|
| 230 |
+
1.0
|
| 231 |
+
],
|
| 232 |
+
"mask": [
|
| 233 |
+
true,
|
| 234 |
+
true,
|
| 235 |
+
true,
|
| 236 |
+
true,
|
| 237 |
+
true,
|
| 238 |
+
true,
|
| 239 |
+
false
|
| 240 |
+
]
|
| 241 |
+
},
|
| 242 |
+
"state": {
|
| 243 |
+
"mean": [
|
| 244 |
+
-1.0741499662399292,
|
| 245 |
+
-1.0714620351791382,
|
| 246 |
+
-1.357372760772705,
|
| 247 |
+
-0.6174096465110779,
|
| 248 |
+
-0.880132257938385,
|
| 249 |
+
0.6550877094268799,
|
| 250 |
+
0.4558027982711792
|
| 251 |
+
],
|
| 252 |
+
"std": [
|
| 253 |
+
0.2520765960216522,
|
| 254 |
+
0.2545098066329956,
|
| 255 |
+
0.1924552172422409,
|
| 256 |
+
0.559248685836792,
|
| 257 |
+
0.41956013441085815,
|
| 258 |
+
1.6006327867507935,
|
| 259 |
+
0.46324989199638367
|
| 260 |
+
],
|
| 261 |
+
"min": [
|
| 262 |
+
-2.1934425830841064,
|
| 263 |
+
-1.8970158100128174,
|
| 264 |
+
-2.2100930213928223,
|
| 265 |
+
-3.1213817596435547,
|
| 266 |
+
-2.0163440704345703,
|
| 267 |
+
-6.283377170562744,
|
| 268 |
+
0.0
|
| 269 |
+
],
|
| 270 |
+
"max": [
|
| 271 |
+
-0.30637508630752563,
|
| 272 |
+
-0.13648474216461182,
|
| 273 |
+
-0.47022661566734314,
|
| 274 |
+
2.9531495571136475,
|
| 275 |
+
1.5494160652160645,
|
| 276 |
+
3.9177753925323486,
|
| 277 |
+
1.0
|
| 278 |
+
],
|
| 279 |
+
"mask": [
|
| 280 |
+
true,
|
| 281 |
+
true,
|
| 282 |
+
true,
|
| 283 |
+
true,
|
| 284 |
+
true,
|
| 285 |
+
true,
|
| 286 |
+
false
|
| 287 |
+
]
|
| 288 |
+
}
|
| 289 |
+
}
|
| 290 |
+
]
|
lch-qiwu-checkpoint-5000/model.safetensors
ADDED
|
@@ -0,0 +1,3 @@
|
|
|
|
|
|
|
|
|
|
|
|
|
| 1 |
+
version https://git-lfs.github.com/spec/v1
|
| 2 |
+
oid sha256:1c300a3aa6f6b5c4f25b1952bf98c71bd2e99c1dd5cca2ae2fa8ffa379e16fd9
|
| 3 |
+
size 405079824
|
lch-qiwu-checkpoint-5000/optimizer.pt
ADDED
|
@@ -0,0 +1,3 @@
|
|
|
|
|
|
|
|
|
|
|
|
|
| 1 |
+
version https://git-lfs.github.com/spec/v1
|
| 2 |
+
oid sha256:064523c065c942e6826f67aa812af8414af995e095b5a42bdd6b7ea5a4fb4f1b
|
| 3 |
+
size 371730170
|
lch-qiwu-checkpoint-5000/rng_state.pth
ADDED
|
@@ -0,0 +1,3 @@
|
|
|
|
|
|
|
|
|
|
|
|
|
| 1 |
+
version https://git-lfs.github.com/spec/v1
|
| 2 |
+
oid sha256:890f790833c5a8adaec27758006635632a6c5dd3693ac7bb0e8ecdeb28180b86
|
| 3 |
+
size 15958
|
lch-qiwu-checkpoint-5000/scheduler.pt
ADDED
|
@@ -0,0 +1,3 @@
|
|
|
|
|
|
|
|
|
|
|
|
|
| 1 |
+
version https://git-lfs.github.com/spec/v1
|
| 2 |
+
oid sha256:83ec18b452c1b862765711b3e16c26b4ab0e9350a942b8e4738c8eee7b92ecdf
|
| 3 |
+
size 1064
|
lch-qiwu-checkpoint-5000/trainer_state.json
ADDED
|
@@ -0,0 +1,784 @@
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
| 1 |
+
{
|
| 2 |
+
"best_global_step": null,
|
| 3 |
+
"best_metric": null,
|
| 4 |
+
"best_model_checkpoint": null,
|
| 5 |
+
"epoch": 0.015625,
|
| 6 |
+
"eval_steps": 100,
|
| 7 |
+
"global_step": 5000,
|
| 8 |
+
"is_hyper_param_search": false,
|
| 9 |
+
"is_local_process_zero": true,
|
| 10 |
+
"is_world_process_zero": true,
|
| 11 |
+
"log_history": [
|
| 12 |
+
{
|
| 13 |
+
"epoch": 0.0003125,
|
| 14 |
+
"grad_norm": 37.75,
|
| 15 |
+
"learning_rate": 2.9715000000000003e-05,
|
| 16 |
+
"loss": 1.4494,
|
| 17 |
+
"step": 100
|
| 18 |
+
},
|
| 19 |
+
{
|
| 20 |
+
"epoch": 0.0003125,
|
| 21 |
+
"eval_loss": 0.9893627762794495,
|
| 22 |
+
"eval_runtime": 83.6129,
|
| 23 |
+
"eval_samples_per_second": 241.793,
|
| 24 |
+
"eval_steps_per_second": 241.793,
|
| 25 |
+
"step": 100
|
| 26 |
+
},
|
| 27 |
+
{
|
| 28 |
+
"epoch": 0.000625,
|
| 29 |
+
"grad_norm": 22.125,
|
| 30 |
+
"learning_rate": 3e-05,
|
| 31 |
+
"loss": 1.1374,
|
| 32 |
+
"step": 200
|
| 33 |
+
},
|
| 34 |
+
{
|
| 35 |
+
"epoch": 0.000625,
|
| 36 |
+
"eval_loss": 0.9329086542129517,
|
| 37 |
+
"eval_runtime": 93.6096,
|
| 38 |
+
"eval_samples_per_second": 215.971,
|
| 39 |
+
"eval_steps_per_second": 215.971,
|
| 40 |
+
"step": 200
|
| 41 |
+
},
|
| 42 |
+
{
|
| 43 |
+
"epoch": 0.0009375,
|
| 44 |
+
"grad_norm": 23.5,
|
| 45 |
+
"learning_rate": 3e-05,
|
| 46 |
+
"loss": 1.0519,
|
| 47 |
+
"step": 300
|
| 48 |
+
},
|
| 49 |
+
{
|
| 50 |
+
"epoch": 0.0009375,
|
| 51 |
+
"eval_loss": 0.9117364883422852,
|
| 52 |
+
"eval_runtime": 89.2029,
|
| 53 |
+
"eval_samples_per_second": 226.641,
|
| 54 |
+
"eval_steps_per_second": 226.641,
|
| 55 |
+
"step": 300
|
| 56 |
+
},
|
| 57 |
+
{
|
| 58 |
+
"epoch": 0.00125,
|
| 59 |
+
"grad_norm": 22.25,
|
| 60 |
+
"learning_rate": 3e-05,
|
| 61 |
+
"loss": 1.0369,
|
| 62 |
+
"step": 400
|
| 63 |
+
},
|
| 64 |
+
{
|
| 65 |
+
"epoch": 0.00125,
|
| 66 |
+
"eval_loss": 0.8862078785896301,
|
| 67 |
+
"eval_runtime": 80.6248,
|
| 68 |
+
"eval_samples_per_second": 250.754,
|
| 69 |
+
"eval_steps_per_second": 250.754,
|
| 70 |
+
"step": 400
|
| 71 |
+
},
|
| 72 |
+
{
|
| 73 |
+
"epoch": 0.0015625,
|
| 74 |
+
"grad_norm": 22.875,
|
| 75 |
+
"learning_rate": 3e-05,
|
| 76 |
+
"loss": 1.0349,
|
| 77 |
+
"step": 500
|
| 78 |
+
},
|
| 79 |
+
{
|
| 80 |
+
"epoch": 0.0015625,
|
| 81 |
+
"eval_loss": 0.8915422558784485,
|
| 82 |
+
"eval_runtime": 93.3883,
|
| 83 |
+
"eval_samples_per_second": 216.483,
|
| 84 |
+
"eval_steps_per_second": 216.483,
|
| 85 |
+
"step": 500
|
| 86 |
+
},
|
| 87 |
+
{
|
| 88 |
+
"epoch": 0.001875,
|
| 89 |
+
"grad_norm": 25.25,
|
| 90 |
+
"learning_rate": 3e-05,
|
| 91 |
+
"loss": 1.0373,
|
| 92 |
+
"step": 600
|
| 93 |
+
},
|
| 94 |
+
{
|
| 95 |
+
"epoch": 0.001875,
|
| 96 |
+
"eval_loss": 0.8709920644760132,
|
| 97 |
+
"eval_runtime": 88.6181,
|
| 98 |
+
"eval_samples_per_second": 228.136,
|
| 99 |
+
"eval_steps_per_second": 228.136,
|
| 100 |
+
"step": 600
|
| 101 |
+
},
|
| 102 |
+
{
|
| 103 |
+
"epoch": 0.0021875,
|
| 104 |
+
"grad_norm": 29.125,
|
| 105 |
+
"learning_rate": 3e-05,
|
| 106 |
+
"loss": 0.9768,
|
| 107 |
+
"step": 700
|
| 108 |
+
},
|
| 109 |
+
{
|
| 110 |
+
"epoch": 0.0021875,
|
| 111 |
+
"eval_loss": 0.8621766567230225,
|
| 112 |
+
"eval_runtime": 87.3819,
|
| 113 |
+
"eval_samples_per_second": 231.364,
|
| 114 |
+
"eval_steps_per_second": 231.364,
|
| 115 |
+
"step": 700
|
| 116 |
+
},
|
| 117 |
+
{
|
| 118 |
+
"epoch": 0.0025,
|
| 119 |
+
"grad_norm": 31.625,
|
| 120 |
+
"learning_rate": 3e-05,
|
| 121 |
+
"loss": 0.9864,
|
| 122 |
+
"step": 800
|
| 123 |
+
},
|
| 124 |
+
{
|
| 125 |
+
"epoch": 0.0025,
|
| 126 |
+
"eval_loss": 0.8656457662582397,
|
| 127 |
+
"eval_runtime": 87.0258,
|
| 128 |
+
"eval_samples_per_second": 232.31,
|
| 129 |
+
"eval_steps_per_second": 232.31,
|
| 130 |
+
"step": 800
|
| 131 |
+
},
|
| 132 |
+
{
|
| 133 |
+
"epoch": 0.0028125,
|
| 134 |
+
"grad_norm": 24.25,
|
| 135 |
+
"learning_rate": 3e-05,
|
| 136 |
+
"loss": 0.971,
|
| 137 |
+
"step": 900
|
| 138 |
+
},
|
| 139 |
+
{
|
| 140 |
+
"epoch": 0.0028125,
|
| 141 |
+
"eval_loss": 0.8618795871734619,
|
| 142 |
+
"eval_runtime": 86.5884,
|
| 143 |
+
"eval_samples_per_second": 233.484,
|
| 144 |
+
"eval_steps_per_second": 233.484,
|
| 145 |
+
"step": 900
|
| 146 |
+
},
|
| 147 |
+
{
|
| 148 |
+
"epoch": 0.003125,
|
| 149 |
+
"grad_norm": 29.0,
|
| 150 |
+
"learning_rate": 3e-05,
|
| 151 |
+
"loss": 0.9625,
|
| 152 |
+
"step": 1000
|
| 153 |
+
},
|
| 154 |
+
{
|
| 155 |
+
"epoch": 0.003125,
|
| 156 |
+
"eval_loss": 0.8543567061424255,
|
| 157 |
+
"eval_runtime": 94.848,
|
| 158 |
+
"eval_samples_per_second": 213.152,
|
| 159 |
+
"eval_steps_per_second": 213.152,
|
| 160 |
+
"step": 1000
|
| 161 |
+
},
|
| 162 |
+
{
|
| 163 |
+
"epoch": 0.0034375,
|
| 164 |
+
"grad_norm": 26.625,
|
| 165 |
+
"learning_rate": 3e-05,
|
| 166 |
+
"loss": 1.0014,
|
| 167 |
+
"step": 1100
|
| 168 |
+
},
|
| 169 |
+
{
|
| 170 |
+
"epoch": 0.0034375,
|
| 171 |
+
"eval_loss": 0.8730055689811707,
|
| 172 |
+
"eval_runtime": 87.0017,
|
| 173 |
+
"eval_samples_per_second": 232.375,
|
| 174 |
+
"eval_steps_per_second": 232.375,
|
| 175 |
+
"step": 1100
|
| 176 |
+
},
|
| 177 |
+
{
|
| 178 |
+
"epoch": 0.00375,
|
| 179 |
+
"grad_norm": 16.0,
|
| 180 |
+
"learning_rate": 3e-05,
|
| 181 |
+
"loss": 0.9598,
|
| 182 |
+
"step": 1200
|
| 183 |
+
},
|
| 184 |
+
{
|
| 185 |
+
"epoch": 0.00375,
|
| 186 |
+
"eval_loss": 0.8562670350074768,
|
| 187 |
+
"eval_runtime": 89.5828,
|
| 188 |
+
"eval_samples_per_second": 225.68,
|
| 189 |
+
"eval_steps_per_second": 225.68,
|
| 190 |
+
"step": 1200
|
| 191 |
+
},
|
| 192 |
+
{
|
| 193 |
+
"epoch": 0.0040625,
|
| 194 |
+
"grad_norm": 19.25,
|
| 195 |
+
"learning_rate": 3e-05,
|
| 196 |
+
"loss": 0.9512,
|
| 197 |
+
"step": 1300
|
| 198 |
+
},
|
| 199 |
+
{
|
| 200 |
+
"epoch": 0.0040625,
|
| 201 |
+
"eval_loss": 0.865445077419281,
|
| 202 |
+
"eval_runtime": 90.7843,
|
| 203 |
+
"eval_samples_per_second": 222.693,
|
| 204 |
+
"eval_steps_per_second": 222.693,
|
| 205 |
+
"step": 1300
|
| 206 |
+
},
|
| 207 |
+
{
|
| 208 |
+
"epoch": 0.004375,
|
| 209 |
+
"grad_norm": 22.75,
|
| 210 |
+
"learning_rate": 3e-05,
|
| 211 |
+
"loss": 0.9378,
|
| 212 |
+
"step": 1400
|
| 213 |
+
},
|
| 214 |
+
{
|
| 215 |
+
"epoch": 0.004375,
|
| 216 |
+
"eval_loss": 0.8774417042732239,
|
| 217 |
+
"eval_runtime": 86.9963,
|
| 218 |
+
"eval_samples_per_second": 232.389,
|
| 219 |
+
"eval_steps_per_second": 232.389,
|
| 220 |
+
"step": 1400
|
| 221 |
+
},
|
| 222 |
+
{
|
| 223 |
+
"epoch": 0.0046875,
|
| 224 |
+
"grad_norm": 23.75,
|
| 225 |
+
"learning_rate": 3e-05,
|
| 226 |
+
"loss": 0.9397,
|
| 227 |
+
"step": 1500
|
| 228 |
+
},
|
| 229 |
+
{
|
| 230 |
+
"epoch": 0.0046875,
|
| 231 |
+
"eval_loss": 0.8403639197349548,
|
| 232 |
+
"eval_runtime": 92.1015,
|
| 233 |
+
"eval_samples_per_second": 219.508,
|
| 234 |
+
"eval_steps_per_second": 219.508,
|
| 235 |
+
"step": 1500
|
| 236 |
+
},
|
| 237 |
+
{
|
| 238 |
+
"epoch": 0.005,
|
| 239 |
+
"grad_norm": 20.75,
|
| 240 |
+
"learning_rate": 3e-05,
|
| 241 |
+
"loss": 0.9588,
|
| 242 |
+
"step": 1600
|
| 243 |
+
},
|
| 244 |
+
{
|
| 245 |
+
"epoch": 0.005,
|
| 246 |
+
"eval_loss": 0.8386977910995483,
|
| 247 |
+
"eval_runtime": 91.8663,
|
| 248 |
+
"eval_samples_per_second": 220.07,
|
| 249 |
+
"eval_steps_per_second": 220.07,
|
| 250 |
+
"step": 1600
|
| 251 |
+
},
|
| 252 |
+
{
|
| 253 |
+
"epoch": 0.0053125,
|
| 254 |
+
"grad_norm": 25.125,
|
| 255 |
+
"learning_rate": 3e-05,
|
| 256 |
+
"loss": 0.986,
|
| 257 |
+
"step": 1700
|
| 258 |
+
},
|
| 259 |
+
{
|
| 260 |
+
"epoch": 0.0053125,
|
| 261 |
+
"eval_loss": 0.8436814546585083,
|
| 262 |
+
"eval_runtime": 85.7235,
|
| 263 |
+
"eval_samples_per_second": 235.84,
|
| 264 |
+
"eval_steps_per_second": 235.84,
|
| 265 |
+
"step": 1700
|
| 266 |
+
},
|
| 267 |
+
{
|
| 268 |
+
"epoch": 0.005625,
|
| 269 |
+
"grad_norm": 22.5,
|
| 270 |
+
"learning_rate": 3e-05,
|
| 271 |
+
"loss": 0.9679,
|
| 272 |
+
"step": 1800
|
| 273 |
+
},
|
| 274 |
+
{
|
| 275 |
+
"epoch": 0.005625,
|
| 276 |
+
"eval_loss": 0.8496856689453125,
|
| 277 |
+
"eval_runtime": 87.5793,
|
| 278 |
+
"eval_samples_per_second": 230.842,
|
| 279 |
+
"eval_steps_per_second": 230.842,
|
| 280 |
+
"step": 1800
|
| 281 |
+
},
|
| 282 |
+
{
|
| 283 |
+
"epoch": 0.0059375,
|
| 284 |
+
"grad_norm": 14.75,
|
| 285 |
+
"learning_rate": 3e-05,
|
| 286 |
+
"loss": 0.9407,
|
| 287 |
+
"step": 1900
|
| 288 |
+
},
|
| 289 |
+
{
|
| 290 |
+
"epoch": 0.0059375,
|
| 291 |
+
"eval_loss": 0.8395476341247559,
|
| 292 |
+
"eval_runtime": 85.866,
|
| 293 |
+
"eval_samples_per_second": 235.448,
|
| 294 |
+
"eval_steps_per_second": 235.448,
|
| 295 |
+
"step": 1900
|
| 296 |
+
},
|
| 297 |
+
{
|
| 298 |
+
"epoch": 0.00625,
|
| 299 |
+
"grad_norm": 25.0,
|
| 300 |
+
"learning_rate": 3e-05,
|
| 301 |
+
"loss": 0.9344,
|
| 302 |
+
"step": 2000
|
| 303 |
+
},
|
| 304 |
+
{
|
| 305 |
+
"epoch": 0.00625,
|
| 306 |
+
"eval_loss": 0.8381168842315674,
|
| 307 |
+
"eval_runtime": 85.3278,
|
| 308 |
+
"eval_samples_per_second": 236.933,
|
| 309 |
+
"eval_steps_per_second": 236.933,
|
| 310 |
+
"step": 2000
|
| 311 |
+
},
|
| 312 |
+
{
|
| 313 |
+
"epoch": 0.0065625,
|
| 314 |
+
"grad_norm": 18.5,
|
| 315 |
+
"learning_rate": 3e-05,
|
| 316 |
+
"loss": 0.946,
|
| 317 |
+
"step": 2100
|
| 318 |
+
},
|
| 319 |
+
{
|
| 320 |
+
"epoch": 0.0065625,
|
| 321 |
+
"eval_loss": 0.832157552242279,
|
| 322 |
+
"eval_runtime": 89.6324,
|
| 323 |
+
"eval_samples_per_second": 225.555,
|
| 324 |
+
"eval_steps_per_second": 225.555,
|
| 325 |
+
"step": 2100
|
| 326 |
+
},
|
| 327 |
+
{
|
| 328 |
+
"epoch": 0.006875,
|
| 329 |
+
"grad_norm": 24.625,
|
| 330 |
+
"learning_rate": 3e-05,
|
| 331 |
+
"loss": 0.9094,
|
| 332 |
+
"step": 2200
|
| 333 |
+
},
|
| 334 |
+
{
|
| 335 |
+
"epoch": 0.006875,
|
| 336 |
+
"eval_loss": 0.8447126746177673,
|
| 337 |
+
"eval_runtime": 86.4865,
|
| 338 |
+
"eval_samples_per_second": 233.759,
|
| 339 |
+
"eval_steps_per_second": 233.759,
|
| 340 |
+
"step": 2200
|
| 341 |
+
},
|
| 342 |
+
{
|
| 343 |
+
"epoch": 0.0071875,
|
| 344 |
+
"grad_norm": 17.625,
|
| 345 |
+
"learning_rate": 3e-05,
|
| 346 |
+
"loss": 0.919,
|
| 347 |
+
"step": 2300
|
| 348 |
+
},
|
| 349 |
+
{
|
| 350 |
+
"epoch": 0.0071875,
|
| 351 |
+
"eval_loss": 0.8236312866210938,
|
| 352 |
+
"eval_runtime": 90.4874,
|
| 353 |
+
"eval_samples_per_second": 223.423,
|
| 354 |
+
"eval_steps_per_second": 223.423,
|
| 355 |
+
"step": 2300
|
| 356 |
+
},
|
| 357 |
+
{
|
| 358 |
+
"epoch": 0.0075,
|
| 359 |
+
"grad_norm": 34.5,
|
| 360 |
+
"learning_rate": 3e-05,
|
| 361 |
+
"loss": 0.91,
|
| 362 |
+
"step": 2400
|
| 363 |
+
},
|
| 364 |
+
{
|
| 365 |
+
"epoch": 0.0075,
|
| 366 |
+
"eval_loss": 0.8279756903648376,
|
| 367 |
+
"eval_runtime": 84.7034,
|
| 368 |
+
"eval_samples_per_second": 238.68,
|
| 369 |
+
"eval_steps_per_second": 238.68,
|
| 370 |
+
"step": 2400
|
| 371 |
+
},
|
| 372 |
+
{
|
| 373 |
+
"epoch": 0.0078125,
|
| 374 |
+
"grad_norm": 14.75,
|
| 375 |
+
"learning_rate": 3e-05,
|
| 376 |
+
"loss": 0.9367,
|
| 377 |
+
"step": 2500
|
| 378 |
+
},
|
| 379 |
+
{
|
| 380 |
+
"epoch": 0.0078125,
|
| 381 |
+
"eval_loss": 0.8254591822624207,
|
| 382 |
+
"eval_runtime": 88.7179,
|
| 383 |
+
"eval_samples_per_second": 227.88,
|
| 384 |
+
"eval_steps_per_second": 227.88,
|
| 385 |
+
"step": 2500
|
| 386 |
+
},
|
| 387 |
+
{
|
| 388 |
+
"epoch": 0.008125,
|
| 389 |
+
"grad_norm": 22.25,
|
| 390 |
+
"learning_rate": 3e-05,
|
| 391 |
+
"loss": 0.93,
|
| 392 |
+
"step": 2600
|
| 393 |
+
},
|
| 394 |
+
{
|
| 395 |
+
"epoch": 0.008125,
|
| 396 |
+
"eval_loss": 0.8389394283294678,
|
| 397 |
+
"eval_runtime": 92.6356,
|
| 398 |
+
"eval_samples_per_second": 218.242,
|
| 399 |
+
"eval_steps_per_second": 218.242,
|
| 400 |
+
"step": 2600
|
| 401 |
+
},
|
| 402 |
+
{
|
| 403 |
+
"epoch": 0.0084375,
|
| 404 |
+
"grad_norm": 30.625,
|
| 405 |
+
"learning_rate": 3e-05,
|
| 406 |
+
"loss": 0.9673,
|
| 407 |
+
"step": 2700
|
| 408 |
+
},
|
| 409 |
+
{
|
| 410 |
+
"epoch": 0.0084375,
|
| 411 |
+
"eval_loss": 0.8376287221908569,
|
| 412 |
+
"eval_runtime": 109.9646,
|
| 413 |
+
"eval_samples_per_second": 183.85,
|
| 414 |
+
"eval_steps_per_second": 183.85,
|
| 415 |
+
"step": 2700
|
| 416 |
+
},
|
| 417 |
+
{
|
| 418 |
+
"epoch": 0.00875,
|
| 419 |
+
"grad_norm": 21.5,
|
| 420 |
+
"learning_rate": 3e-05,
|
| 421 |
+
"loss": 0.9068,
|
| 422 |
+
"step": 2800
|
| 423 |
+
},
|
| 424 |
+
{
|
| 425 |
+
"epoch": 0.00875,
|
| 426 |
+
"eval_loss": 0.8400581479072571,
|
| 427 |
+
"eval_runtime": 84.9979,
|
| 428 |
+
"eval_samples_per_second": 237.853,
|
| 429 |
+
"eval_steps_per_second": 237.853,
|
| 430 |
+
"step": 2800
|
| 431 |
+
},
|
| 432 |
+
{
|
| 433 |
+
"epoch": 0.0090625,
|
| 434 |
+
"grad_norm": 25.5,
|
| 435 |
+
"learning_rate": 3e-05,
|
| 436 |
+
"loss": 0.8893,
|
| 437 |
+
"step": 2900
|
| 438 |
+
},
|
| 439 |
+
{
|
| 440 |
+
"epoch": 0.0090625,
|
| 441 |
+
"eval_loss": 0.8247655630111694,
|
| 442 |
+
"eval_runtime": 84.3633,
|
| 443 |
+
"eval_samples_per_second": 239.642,
|
| 444 |
+
"eval_steps_per_second": 239.642,
|
| 445 |
+
"step": 2900
|
| 446 |
+
},
|
| 447 |
+
{
|
| 448 |
+
"epoch": 0.009375,
|
| 449 |
+
"grad_norm": 25.75,
|
| 450 |
+
"learning_rate": 3e-05,
|
| 451 |
+
"loss": 0.9009,
|
| 452 |
+
"step": 3000
|
| 453 |
+
},
|
| 454 |
+
{
|
| 455 |
+
"epoch": 0.009375,
|
| 456 |
+
"eval_loss": 0.8231511116027832,
|
| 457 |
+
"eval_runtime": 105.5337,
|
| 458 |
+
"eval_samples_per_second": 191.569,
|
| 459 |
+
"eval_steps_per_second": 191.569,
|
| 460 |
+
"step": 3000
|
| 461 |
+
},
|
| 462 |
+
{
|
| 463 |
+
"epoch": 0.0096875,
|
| 464 |
+
"grad_norm": 22.875,
|
| 465 |
+
"learning_rate": 3e-05,
|
| 466 |
+
"loss": 0.9238,
|
| 467 |
+
"step": 3100
|
| 468 |
+
},
|
| 469 |
+
{
|
| 470 |
+
"epoch": 0.0096875,
|
| 471 |
+
"eval_loss": 0.8190457224845886,
|
| 472 |
+
"eval_runtime": 82.5894,
|
| 473 |
+
"eval_samples_per_second": 244.789,
|
| 474 |
+
"eval_steps_per_second": 244.789,
|
| 475 |
+
"step": 3100
|
| 476 |
+
},
|
| 477 |
+
{
|
| 478 |
+
"epoch": 0.01,
|
| 479 |
+
"grad_norm": 23.875,
|
| 480 |
+
"learning_rate": 3e-05,
|
| 481 |
+
"loss": 0.9032,
|
| 482 |
+
"step": 3200
|
| 483 |
+
},
|
| 484 |
+
{
|
| 485 |
+
"epoch": 0.01,
|
| 486 |
+
"eval_loss": 0.8098758459091187,
|
| 487 |
+
"eval_runtime": 88.6206,
|
| 488 |
+
"eval_samples_per_second": 228.13,
|
| 489 |
+
"eval_steps_per_second": 228.13,
|
| 490 |
+
"step": 3200
|
| 491 |
+
},
|
| 492 |
+
{
|
| 493 |
+
"epoch": 0.0103125,
|
| 494 |
+
"grad_norm": 26.125,
|
| 495 |
+
"learning_rate": 3e-05,
|
| 496 |
+
"loss": 0.9544,
|
| 497 |
+
"step": 3300
|
| 498 |
+
},
|
| 499 |
+
{
|
| 500 |
+
"epoch": 0.0103125,
|
| 501 |
+
"eval_loss": 0.8281997442245483,
|
| 502 |
+
"eval_runtime": 88.8475,
|
| 503 |
+
"eval_samples_per_second": 227.547,
|
| 504 |
+
"eval_steps_per_second": 227.547,
|
| 505 |
+
"step": 3300
|
| 506 |
+
},
|
| 507 |
+
{
|
| 508 |
+
"epoch": 0.010625,
|
| 509 |
+
"grad_norm": 30.25,
|
| 510 |
+
"learning_rate": 3e-05,
|
| 511 |
+
"loss": 0.8872,
|
| 512 |
+
"step": 3400
|
| 513 |
+
},
|
| 514 |
+
{
|
| 515 |
+
"epoch": 0.010625,
|
| 516 |
+
"eval_loss": 0.8237181305885315,
|
| 517 |
+
"eval_runtime": 87.3723,
|
| 518 |
+
"eval_samples_per_second": 231.389,
|
| 519 |
+
"eval_steps_per_second": 231.389,
|
| 520 |
+
"step": 3400
|
| 521 |
+
},
|
| 522 |
+
{
|
| 523 |
+
"epoch": 0.0109375,
|
| 524 |
+
"grad_norm": 21.625,
|
| 525 |
+
"learning_rate": 3e-05,
|
| 526 |
+
"loss": 0.8919,
|
| 527 |
+
"step": 3500
|
| 528 |
+
},
|
| 529 |
+
{
|
| 530 |
+
"epoch": 0.0109375,
|
| 531 |
+
"eval_loss": 0.8282459378242493,
|
| 532 |
+
"eval_runtime": 93.0947,
|
| 533 |
+
"eval_samples_per_second": 217.166,
|
| 534 |
+
"eval_steps_per_second": 217.166,
|
| 535 |
+
"step": 3500
|
| 536 |
+
},
|
| 537 |
+
{
|
| 538 |
+
"epoch": 0.01125,
|
| 539 |
+
"grad_norm": 29.5,
|
| 540 |
+
"learning_rate": 3e-05,
|
| 541 |
+
"loss": 0.9237,
|
| 542 |
+
"step": 3600
|
| 543 |
+
},
|
| 544 |
+
{
|
| 545 |
+
"epoch": 0.01125,
|
| 546 |
+
"eval_loss": 0.8266128301620483,
|
| 547 |
+
"eval_runtime": 83.0185,
|
| 548 |
+
"eval_samples_per_second": 243.524,
|
| 549 |
+
"eval_steps_per_second": 243.524,
|
| 550 |
+
"step": 3600
|
| 551 |
+
},
|
| 552 |
+
{
|
| 553 |
+
"epoch": 0.0115625,
|
| 554 |
+
"grad_norm": 18.75,
|
| 555 |
+
"learning_rate": 3e-05,
|
| 556 |
+
"loss": 0.9235,
|
| 557 |
+
"step": 3700
|
| 558 |
+
},
|
| 559 |
+
{
|
| 560 |
+
"epoch": 0.0115625,
|
| 561 |
+
"eval_loss": 0.859565258026123,
|
| 562 |
+
"eval_runtime": 97.9546,
|
| 563 |
+
"eval_samples_per_second": 206.391,
|
| 564 |
+
"eval_steps_per_second": 206.391,
|
| 565 |
+
"step": 3700
|
| 566 |
+
},
|
| 567 |
+
{
|
| 568 |
+
"epoch": 0.011875,
|
| 569 |
+
"grad_norm": 22.0,
|
| 570 |
+
"learning_rate": 3e-05,
|
| 571 |
+
"loss": 0.9026,
|
| 572 |
+
"step": 3800
|
| 573 |
+
},
|
| 574 |
+
{
|
| 575 |
+
"epoch": 0.011875,
|
| 576 |
+
"eval_loss": 0.8544378280639648,
|
| 577 |
+
"eval_runtime": 82.9352,
|
| 578 |
+
"eval_samples_per_second": 243.769,
|
| 579 |
+
"eval_steps_per_second": 243.769,
|
| 580 |
+
"step": 3800
|
| 581 |
+
},
|
| 582 |
+
{
|
| 583 |
+
"epoch": 0.0121875,
|
| 584 |
+
"grad_norm": 20.5,
|
| 585 |
+
"learning_rate": 3e-05,
|
| 586 |
+
"loss": 0.9764,
|
| 587 |
+
"step": 3900
|
| 588 |
+
},
|
| 589 |
+
{
|
| 590 |
+
"epoch": 0.0121875,
|
| 591 |
+
"eval_loss": 0.8370941877365112,
|
| 592 |
+
"eval_runtime": 93.308,
|
| 593 |
+
"eval_samples_per_second": 216.67,
|
| 594 |
+
"eval_steps_per_second": 216.67,
|
| 595 |
+
"step": 3900
|
| 596 |
+
},
|
| 597 |
+
{
|
| 598 |
+
"epoch": 0.0125,
|
| 599 |
+
"grad_norm": 24.875,
|
| 600 |
+
"learning_rate": 3e-05,
|
| 601 |
+
"loss": 0.9143,
|
| 602 |
+
"step": 4000
|
| 603 |
+
},
|
| 604 |
+
{
|
| 605 |
+
"epoch": 0.0125,
|
| 606 |
+
"eval_loss": 0.8286308646202087,
|
| 607 |
+
"eval_runtime": 88.2515,
|
| 608 |
+
"eval_samples_per_second": 229.084,
|
| 609 |
+
"eval_steps_per_second": 229.084,
|
| 610 |
+
"step": 4000
|
| 611 |
+
},
|
| 612 |
+
{
|
| 613 |
+
"epoch": 0.0128125,
|
| 614 |
+
"grad_norm": 27.5,
|
| 615 |
+
"learning_rate": 3e-05,
|
| 616 |
+
"loss": 0.9089,
|
| 617 |
+
"step": 4100
|
| 618 |
+
},
|
| 619 |
+
{
|
| 620 |
+
"epoch": 0.0128125,
|
| 621 |
+
"eval_loss": 0.8309729695320129,
|
| 622 |
+
"eval_runtime": 95.4091,
|
| 623 |
+
"eval_samples_per_second": 211.898,
|
| 624 |
+
"eval_steps_per_second": 211.898,
|
| 625 |
+
"step": 4100
|
| 626 |
+
},
|
| 627 |
+
{
|
| 628 |
+
"epoch": 0.013125,
|
| 629 |
+
"grad_norm": 19.625,
|
| 630 |
+
"learning_rate": 3e-05,
|
| 631 |
+
"loss": 0.9725,
|
| 632 |
+
"step": 4200
|
| 633 |
+
},
|
| 634 |
+
{
|
| 635 |
+
"epoch": 0.013125,
|
| 636 |
+
"eval_loss": 0.8205706477165222,
|
| 637 |
+
"eval_runtime": 84.2153,
|
| 638 |
+
"eval_samples_per_second": 240.063,
|
| 639 |
+
"eval_steps_per_second": 240.063,
|
| 640 |
+
"step": 4200
|
| 641 |
+
},
|
| 642 |
+
{
|
| 643 |
+
"epoch": 0.0134375,
|
| 644 |
+
"grad_norm": 24.375,
|
| 645 |
+
"learning_rate": 3e-05,
|
| 646 |
+
"loss": 0.8795,
|
| 647 |
+
"step": 4300
|
| 648 |
+
},
|
| 649 |
+
{
|
| 650 |
+
"epoch": 0.0134375,
|
| 651 |
+
"eval_loss": 0.854993999004364,
|
| 652 |
+
"eval_runtime": 85.1335,
|
| 653 |
+
"eval_samples_per_second": 237.474,
|
| 654 |
+
"eval_steps_per_second": 237.474,
|
| 655 |
+
"step": 4300
|
| 656 |
+
},
|
| 657 |
+
{
|
| 658 |
+
"epoch": 0.01375,
|
| 659 |
+
"grad_norm": 34.25,
|
| 660 |
+
"learning_rate": 3e-05,
|
| 661 |
+
"loss": 0.9001,
|
| 662 |
+
"step": 4400
|
| 663 |
+
},
|
| 664 |
+
{
|
| 665 |
+
"epoch": 0.01375,
|
| 666 |
+
"eval_loss": 0.8291048407554626,
|
| 667 |
+
"eval_runtime": 85.3999,
|
| 668 |
+
"eval_samples_per_second": 236.733,
|
| 669 |
+
"eval_steps_per_second": 236.733,
|
| 670 |
+
"step": 4400
|
| 671 |
+
},
|
| 672 |
+
{
|
| 673 |
+
"epoch": 0.0140625,
|
| 674 |
+
"grad_norm": 27.75,
|
| 675 |
+
"learning_rate": 3e-05,
|
| 676 |
+
"loss": 0.9336,
|
| 677 |
+
"step": 4500
|
| 678 |
+
},
|
| 679 |
+
{
|
| 680 |
+
"epoch": 0.0140625,
|
| 681 |
+
"eval_loss": 0.8212500214576721,
|
| 682 |
+
"eval_runtime": 89.2017,
|
| 683 |
+
"eval_samples_per_second": 226.644,
|
| 684 |
+
"eval_steps_per_second": 226.644,
|
| 685 |
+
"step": 4500
|
| 686 |
+
},
|
| 687 |
+
{
|
| 688 |
+
"epoch": 0.014375,
|
| 689 |
+
"grad_norm": 23.25,
|
| 690 |
+
"learning_rate": 3e-05,
|
| 691 |
+
"loss": 0.8808,
|
| 692 |
+
"step": 4600
|
| 693 |
+
},
|
| 694 |
+
{
|
| 695 |
+
"epoch": 0.014375,
|
| 696 |
+
"eval_loss": 0.8099567890167236,
|
| 697 |
+
"eval_runtime": 90.1311,
|
| 698 |
+
"eval_samples_per_second": 224.307,
|
| 699 |
+
"eval_steps_per_second": 224.307,
|
| 700 |
+
"step": 4600
|
| 701 |
+
},
|
| 702 |
+
{
|
| 703 |
+
"epoch": 0.0146875,
|
| 704 |
+
"grad_norm": 24.75,
|
| 705 |
+
"learning_rate": 3e-05,
|
| 706 |
+
"loss": 0.9096,
|
| 707 |
+
"step": 4700
|
| 708 |
+
},
|
| 709 |
+
{
|
| 710 |
+
"epoch": 0.0146875,
|
| 711 |
+
"eval_loss": 0.8344265222549438,
|
| 712 |
+
"eval_runtime": 85.2507,
|
| 713 |
+
"eval_samples_per_second": 237.148,
|
| 714 |
+
"eval_steps_per_second": 237.148,
|
| 715 |
+
"step": 4700
|
| 716 |
+
},
|
| 717 |
+
{
|
| 718 |
+
"epoch": 0.015,
|
| 719 |
+
"grad_norm": 34.0,
|
| 720 |
+
"learning_rate": 3e-05,
|
| 721 |
+
"loss": 0.879,
|
| 722 |
+
"step": 4800
|
| 723 |
+
},
|
| 724 |
+
{
|
| 725 |
+
"epoch": 0.015,
|
| 726 |
+
"eval_loss": 0.8386716246604919,
|
| 727 |
+
"eval_runtime": 83.0978,
|
| 728 |
+
"eval_samples_per_second": 243.292,
|
| 729 |
+
"eval_steps_per_second": 243.292,
|
| 730 |
+
"step": 4800
|
| 731 |
+
},
|
| 732 |
+
{
|
| 733 |
+
"epoch": 0.0153125,
|
| 734 |
+
"grad_norm": 29.875,
|
| 735 |
+
"learning_rate": 3e-05,
|
| 736 |
+
"loss": 0.8607,
|
| 737 |
+
"step": 4900
|
| 738 |
+
},
|
| 739 |
+
{
|
| 740 |
+
"epoch": 0.0153125,
|
| 741 |
+
"eval_loss": 0.8188918828964233,
|
| 742 |
+
"eval_runtime": 83.1992,
|
| 743 |
+
"eval_samples_per_second": 242.995,
|
| 744 |
+
"eval_steps_per_second": 242.995,
|
| 745 |
+
"step": 4900
|
| 746 |
+
},
|
| 747 |
+
{
|
| 748 |
+
"epoch": 0.015625,
|
| 749 |
+
"grad_norm": 14.9375,
|
| 750 |
+
"learning_rate": 3e-05,
|
| 751 |
+
"loss": 0.9215,
|
| 752 |
+
"step": 5000
|
| 753 |
+
},
|
| 754 |
+
{
|
| 755 |
+
"epoch": 0.015625,
|
| 756 |
+
"eval_loss": 0.8134725093841553,
|
| 757 |
+
"eval_runtime": 82.3232,
|
| 758 |
+
"eval_samples_per_second": 245.581,
|
| 759 |
+
"eval_steps_per_second": 245.581,
|
| 760 |
+
"step": 5000
|
| 761 |
+
}
|
| 762 |
+
],
|
| 763 |
+
"logging_steps": 100,
|
| 764 |
+
"max_steps": 5000,
|
| 765 |
+
"num_input_tokens_seen": 0,
|
| 766 |
+
"num_train_epochs": 1,
|
| 767 |
+
"save_steps": 500,
|
| 768 |
+
"stateful_callbacks": {
|
| 769 |
+
"TrainerControl": {
|
| 770 |
+
"args": {
|
| 771 |
+
"should_epoch_stop": false,
|
| 772 |
+
"should_evaluate": false,
|
| 773 |
+
"should_log": false,
|
| 774 |
+
"should_save": true,
|
| 775 |
+
"should_training_stop": true
|
| 776 |
+
},
|
| 777 |
+
"attributes": {}
|
| 778 |
+
}
|
| 779 |
+
},
|
| 780 |
+
"total_flos": 0.0,
|
| 781 |
+
"train_batch_size": 1,
|
| 782 |
+
"trial_name": null,
|
| 783 |
+
"trial_params": null
|
| 784 |
+
}
|
lch-qiwu-checkpoint-5000/training_args.bin
ADDED
|
@@ -0,0 +1,3 @@
|
|
|
|
|
|
|
|
|
|
|
|
|
| 1 |
+
version https://git-lfs.github.com/spec/v1
|
| 2 |
+
oid sha256:b9dcbcecaa8883b9529a3d238487136d360bb4b37dc807e05c24a5c6cf8d75f1
|
| 3 |
+
size 5432
|