update files
Browse files- README.md +59 -0
- chat_demo.png +0 -0
- model-00001-of-00002.safetensors +3 -0
- model-00002-of-00002.safetensors +3 -0
- tokenizer.model +3 -0
- trainer_log.jsonl +0 -0
- trainer_state.json +0 -2830
- training_loss.png +0 -0
README.md
ADDED
|
@@ -0,0 +1,59 @@
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
| 1 |
+
---
|
| 2 |
+
license: other
|
| 3 |
+
base_model: /apdcephfs_cq10/share_919031/larsonwang/LLaMA-Factory/save_model/train_lora_1709305042/
|
| 4 |
+
tags:
|
| 5 |
+
- llama-factory
|
| 6 |
+
- full
|
| 7 |
+
- generated_from_trainer
|
| 8 |
+
model-index:
|
| 9 |
+
- name: train_lora_1709346779
|
| 10 |
+
results: []
|
| 11 |
+
---
|
| 12 |
+
|
| 13 |
+
<!-- This model card has been generated automatically according to the information the Trainer had access to. You
|
| 14 |
+
should probably proofread and complete it, then remove this comment. -->
|
| 15 |
+
|
| 16 |
+
# train_lora_1709346779
|
| 17 |
+
|
| 18 |
+
This model is a fine-tuned version of [/apdcephfs_cq10/share_919031/larsonwang/LLaMA-Factory/save_model/train_lora_1709305042/](https://huggingface.co//apdcephfs_cq10/share_919031/larsonwang/LLaMA-Factory/save_model/train_lora_1709305042/) on the sft_sample dataset.
|
| 19 |
+
|
| 20 |
+
## Model description
|
| 21 |
+
|
| 22 |
+
More information needed
|
| 23 |
+
|
| 24 |
+
## Intended uses & limitations
|
| 25 |
+
|
| 26 |
+
More information needed
|
| 27 |
+
|
| 28 |
+
## Training and evaluation data
|
| 29 |
+
|
| 30 |
+
More information needed
|
| 31 |
+
|
| 32 |
+
## Training procedure
|
| 33 |
+
|
| 34 |
+
### Training hyperparameters
|
| 35 |
+
|
| 36 |
+
The following hyperparameters were used during training:
|
| 37 |
+
- learning_rate: 5e-05
|
| 38 |
+
- train_batch_size: 8
|
| 39 |
+
- eval_batch_size: 8
|
| 40 |
+
- seed: 42
|
| 41 |
+
- distributed_type: multi-GPU
|
| 42 |
+
- num_devices: 8
|
| 43 |
+
- total_train_batch_size: 64
|
| 44 |
+
- total_eval_batch_size: 64
|
| 45 |
+
- optimizer: Adam with betas=(0.9,0.999) and epsilon=1e-08
|
| 46 |
+
- lr_scheduler_type: cosine
|
| 47 |
+
- num_epochs: 5.0
|
| 48 |
+
- mixed_precision_training: Native AMP
|
| 49 |
+
|
| 50 |
+
### Training results
|
| 51 |
+
|
| 52 |
+
|
| 53 |
+
|
| 54 |
+
### Framework versions
|
| 55 |
+
|
| 56 |
+
- Transformers 4.38.0
|
| 57 |
+
- Pytorch 2.0.1+cu118
|
| 58 |
+
- Datasets 2.17.1
|
| 59 |
+
- Tokenizers 0.15.2
|
chat_demo.png
ADDED
|
model-00001-of-00002.safetensors
ADDED
|
@@ -0,0 +1,3 @@
|
|
|
|
|
|
|
|
|
|
|
|
|
| 1 |
+
version https://git-lfs.github.com/spec/v1
|
| 2 |
+
oid sha256:929ea3b387ccae95ab652c8d5b6a92f8ce7e5411d74ed7934736dabe7662d189
|
| 3 |
+
size 4968048632
|
model-00002-of-00002.safetensors
ADDED
|
@@ -0,0 +1,3 @@
|
|
|
|
|
|
|
|
|
|
|
|
|
| 1 |
+
version https://git-lfs.github.com/spec/v1
|
| 2 |
+
oid sha256:54011c6a38ca1442415497c0737da7929ea97c3c2708e66c081a2501abf701da
|
| 3 |
+
size 67121600
|
tokenizer.model
ADDED
|
@@ -0,0 +1,3 @@
|
|
|
|
|
|
|
|
|
|
|
|
|
| 1 |
+
version https://git-lfs.github.com/spec/v1
|
| 2 |
+
oid sha256:885336494eed74f9943b15ec83781e8214fcb3bf3105aa51f18108dbbddf8cf0
|
| 3 |
+
size 4328881
|
trainer_log.jsonl
ADDED
|
The diff for this file is too large to render.
See raw diff
|
|
|
trainer_state.json
DELETED
|
@@ -1,2830 +0,0 @@
|
|
| 1 |
-
{
|
| 2 |
-
"best_metric": null,
|
| 3 |
-
"best_model_checkpoint": null,
|
| 4 |
-
"epoch": 5.0,
|
| 5 |
-
"eval_steps": 500,
|
| 6 |
-
"global_step": 4000,
|
| 7 |
-
"is_hyper_param_search": false,
|
| 8 |
-
"is_local_process_zero": true,
|
| 9 |
-
"is_world_process_zero": true,
|
| 10 |
-
"log_history": [
|
| 11 |
-
{
|
| 12 |
-
"epoch": 0.01,
|
| 13 |
-
"grad_norm": 18.572529250843502,
|
| 14 |
-
"learning_rate": 4.998749374687344e-05,
|
| 15 |
-
"loss": 0.6938,
|
| 16 |
-
"step": 10
|
| 17 |
-
},
|
| 18 |
-
{
|
| 19 |
-
"epoch": 0.03,
|
| 20 |
-
"grad_norm": 4.9721361054552675,
|
| 21 |
-
"learning_rate": 4.9874937468734366e-05,
|
| 22 |
-
"loss": 0.4097,
|
| 23 |
-
"step": 20
|
| 24 |
-
},
|
| 25 |
-
{
|
| 26 |
-
"epoch": 0.04,
|
| 27 |
-
"grad_norm": 4.416736002443796,
|
| 28 |
-
"learning_rate": 4.9749874937468736e-05,
|
| 29 |
-
"loss": 0.3894,
|
| 30 |
-
"step": 30
|
| 31 |
-
},
|
| 32 |
-
{
|
| 33 |
-
"epoch": 0.05,
|
| 34 |
-
"grad_norm": 5.588328371365274,
|
| 35 |
-
"learning_rate": 4.9624812406203106e-05,
|
| 36 |
-
"loss": 0.4047,
|
| 37 |
-
"step": 40
|
| 38 |
-
},
|
| 39 |
-
{
|
| 40 |
-
"epoch": 0.06,
|
| 41 |
-
"grad_norm": 4.53230725812265,
|
| 42 |
-
"learning_rate": 4.9499749874937476e-05,
|
| 43 |
-
"loss": 0.4542,
|
| 44 |
-
"step": 50
|
| 45 |
-
},
|
| 46 |
-
{
|
| 47 |
-
"epoch": 0.07,
|
| 48 |
-
"grad_norm": 4.826575762136805,
|
| 49 |
-
"learning_rate": 4.937468734367184e-05,
|
| 50 |
-
"loss": 0.4522,
|
| 51 |
-
"step": 60
|
| 52 |
-
},
|
| 53 |
-
{
|
| 54 |
-
"epoch": 0.09,
|
| 55 |
-
"grad_norm": 4.525790589117328,
|
| 56 |
-
"learning_rate": 4.92496248124062e-05,
|
| 57 |
-
"loss": 0.4822,
|
| 58 |
-
"step": 70
|
| 59 |
-
},
|
| 60 |
-
{
|
| 61 |
-
"epoch": 0.1,
|
| 62 |
-
"grad_norm": 5.296156572891334,
|
| 63 |
-
"learning_rate": 4.912456228114057e-05,
|
| 64 |
-
"loss": 0.5055,
|
| 65 |
-
"step": 80
|
| 66 |
-
},
|
| 67 |
-
{
|
| 68 |
-
"epoch": 0.11,
|
| 69 |
-
"grad_norm": 5.819088718868824,
|
| 70 |
-
"learning_rate": 4.8999499749874936e-05,
|
| 71 |
-
"loss": 0.473,
|
| 72 |
-
"step": 90
|
| 73 |
-
},
|
| 74 |
-
{
|
| 75 |
-
"epoch": 0.12,
|
| 76 |
-
"grad_norm": 5.433765575947364,
|
| 77 |
-
"learning_rate": 4.887443721860931e-05,
|
| 78 |
-
"loss": 0.5087,
|
| 79 |
-
"step": 100
|
| 80 |
-
},
|
| 81 |
-
{
|
| 82 |
-
"epoch": 0.14,
|
| 83 |
-
"grad_norm": 5.0965740589803925,
|
| 84 |
-
"learning_rate": 4.8749374687343676e-05,
|
| 85 |
-
"loss": 0.5234,
|
| 86 |
-
"step": 110
|
| 87 |
-
},
|
| 88 |
-
{
|
| 89 |
-
"epoch": 0.15,
|
| 90 |
-
"grad_norm": 5.850601504449537,
|
| 91 |
-
"learning_rate": 4.862431215607804e-05,
|
| 92 |
-
"loss": 0.5141,
|
| 93 |
-
"step": 120
|
| 94 |
-
},
|
| 95 |
-
{
|
| 96 |
-
"epoch": 0.16,
|
| 97 |
-
"grad_norm": 4.719305144742639,
|
| 98 |
-
"learning_rate": 4.849924962481241e-05,
|
| 99 |
-
"loss": 0.5434,
|
| 100 |
-
"step": 130
|
| 101 |
-
},
|
| 102 |
-
{
|
| 103 |
-
"epoch": 0.17,
|
| 104 |
-
"grad_norm": 5.005327628861193,
|
| 105 |
-
"learning_rate": 4.837418709354677e-05,
|
| 106 |
-
"loss": 0.5293,
|
| 107 |
-
"step": 140
|
| 108 |
-
},
|
| 109 |
-
{
|
| 110 |
-
"epoch": 0.19,
|
| 111 |
-
"grad_norm": 5.146773444206217,
|
| 112 |
-
"learning_rate": 4.824912456228114e-05,
|
| 113 |
-
"loss": 0.511,
|
| 114 |
-
"step": 150
|
| 115 |
-
},
|
| 116 |
-
{
|
| 117 |
-
"epoch": 0.2,
|
| 118 |
-
"grad_norm": 4.612446424385032,
|
| 119 |
-
"learning_rate": 4.812406203101551e-05,
|
| 120 |
-
"loss": 0.5401,
|
| 121 |
-
"step": 160
|
| 122 |
-
},
|
| 123 |
-
{
|
| 124 |
-
"epoch": 0.21,
|
| 125 |
-
"grad_norm": 5.293068567439793,
|
| 126 |
-
"learning_rate": 4.7998999499749876e-05,
|
| 127 |
-
"loss": 0.6019,
|
| 128 |
-
"step": 170
|
| 129 |
-
},
|
| 130 |
-
{
|
| 131 |
-
"epoch": 0.23,
|
| 132 |
-
"grad_norm": 5.1015740901237026,
|
| 133 |
-
"learning_rate": 4.7873936968484246e-05,
|
| 134 |
-
"loss": 0.5633,
|
| 135 |
-
"step": 180
|
| 136 |
-
},
|
| 137 |
-
{
|
| 138 |
-
"epoch": 0.24,
|
| 139 |
-
"grad_norm": 4.824851299623291,
|
| 140 |
-
"learning_rate": 4.774887443721861e-05,
|
| 141 |
-
"loss": 0.583,
|
| 142 |
-
"step": 190
|
| 143 |
-
},
|
| 144 |
-
{
|
| 145 |
-
"epoch": 0.25,
|
| 146 |
-
"grad_norm": 5.062655599581182,
|
| 147 |
-
"learning_rate": 4.762381190595298e-05,
|
| 148 |
-
"loss": 0.6201,
|
| 149 |
-
"step": 200
|
| 150 |
-
},
|
| 151 |
-
{
|
| 152 |
-
"epoch": 0.26,
|
| 153 |
-
"grad_norm": 5.326127827444376,
|
| 154 |
-
"learning_rate": 4.749874937468735e-05,
|
| 155 |
-
"loss": 0.6125,
|
| 156 |
-
"step": 210
|
| 157 |
-
},
|
| 158 |
-
{
|
| 159 |
-
"epoch": 0.28,
|
| 160 |
-
"grad_norm": 4.9596551153810475,
|
| 161 |
-
"learning_rate": 4.737368684342171e-05,
|
| 162 |
-
"loss": 0.5686,
|
| 163 |
-
"step": 220
|
| 164 |
-
},
|
| 165 |
-
{
|
| 166 |
-
"epoch": 0.29,
|
| 167 |
-
"grad_norm": 4.658617889684648,
|
| 168 |
-
"learning_rate": 4.724862431215608e-05,
|
| 169 |
-
"loss": 0.6173,
|
| 170 |
-
"step": 230
|
| 171 |
-
},
|
| 172 |
-
{
|
| 173 |
-
"epoch": 0.3,
|
| 174 |
-
"grad_norm": 4.7094963235295015,
|
| 175 |
-
"learning_rate": 4.7123561780890446e-05,
|
| 176 |
-
"loss": 0.6373,
|
| 177 |
-
"step": 240
|
| 178 |
-
},
|
| 179 |
-
{
|
| 180 |
-
"epoch": 0.31,
|
| 181 |
-
"grad_norm": 5.322370626831193,
|
| 182 |
-
"learning_rate": 4.6998499249624816e-05,
|
| 183 |
-
"loss": 0.5606,
|
| 184 |
-
"step": 250
|
| 185 |
-
},
|
| 186 |
-
{
|
| 187 |
-
"epoch": 0.33,
|
| 188 |
-
"grad_norm": 4.959609351002434,
|
| 189 |
-
"learning_rate": 4.687343671835918e-05,
|
| 190 |
-
"loss": 0.5903,
|
| 191 |
-
"step": 260
|
| 192 |
-
},
|
| 193 |
-
{
|
| 194 |
-
"epoch": 0.34,
|
| 195 |
-
"grad_norm": 5.7940212741786485,
|
| 196 |
-
"learning_rate": 4.674837418709355e-05,
|
| 197 |
-
"loss": 0.6201,
|
| 198 |
-
"step": 270
|
| 199 |
-
},
|
| 200 |
-
{
|
| 201 |
-
"epoch": 0.35,
|
| 202 |
-
"grad_norm": 6.241362745205275,
|
| 203 |
-
"learning_rate": 4.662331165582792e-05,
|
| 204 |
-
"loss": 0.6295,
|
| 205 |
-
"step": 280
|
| 206 |
-
},
|
| 207 |
-
{
|
| 208 |
-
"epoch": 0.36,
|
| 209 |
-
"grad_norm": 4.421199895323022,
|
| 210 |
-
"learning_rate": 4.649824912456228e-05,
|
| 211 |
-
"loss": 0.6242,
|
| 212 |
-
"step": 290
|
| 213 |
-
},
|
| 214 |
-
{
|
| 215 |
-
"epoch": 0.38,
|
| 216 |
-
"grad_norm": 5.693381150794548,
|
| 217 |
-
"learning_rate": 4.637318659329665e-05,
|
| 218 |
-
"loss": 0.6343,
|
| 219 |
-
"step": 300
|
| 220 |
-
},
|
| 221 |
-
{
|
| 222 |
-
"epoch": 0.39,
|
| 223 |
-
"grad_norm": 5.84570997692131,
|
| 224 |
-
"learning_rate": 4.6248124062031015e-05,
|
| 225 |
-
"loss": 0.6293,
|
| 226 |
-
"step": 310
|
| 227 |
-
},
|
| 228 |
-
{
|
| 229 |
-
"epoch": 0.4,
|
| 230 |
-
"grad_norm": 4.891858807526382,
|
| 231 |
-
"learning_rate": 4.6123061530765386e-05,
|
| 232 |
-
"loss": 0.6699,
|
| 233 |
-
"step": 320
|
| 234 |
-
},
|
| 235 |
-
{
|
| 236 |
-
"epoch": 0.41,
|
| 237 |
-
"grad_norm": 5.476686832687266,
|
| 238 |
-
"learning_rate": 4.5997998999499756e-05,
|
| 239 |
-
"loss": 0.6448,
|
| 240 |
-
"step": 330
|
| 241 |
-
},
|
| 242 |
-
{
|
| 243 |
-
"epoch": 0.42,
|
| 244 |
-
"grad_norm": 5.629121076096647,
|
| 245 |
-
"learning_rate": 4.587293646823412e-05,
|
| 246 |
-
"loss": 0.6776,
|
| 247 |
-
"step": 340
|
| 248 |
-
},
|
| 249 |
-
{
|
| 250 |
-
"epoch": 0.44,
|
| 251 |
-
"grad_norm": 5.6676894368204485,
|
| 252 |
-
"learning_rate": 4.574787393696849e-05,
|
| 253 |
-
"loss": 0.6558,
|
| 254 |
-
"step": 350
|
| 255 |
-
},
|
| 256 |
-
{
|
| 257 |
-
"epoch": 0.45,
|
| 258 |
-
"grad_norm": 5.28952155100302,
|
| 259 |
-
"learning_rate": 4.562281140570285e-05,
|
| 260 |
-
"loss": 0.705,
|
| 261 |
-
"step": 360
|
| 262 |
-
},
|
| 263 |
-
{
|
| 264 |
-
"epoch": 0.46,
|
| 265 |
-
"grad_norm": 4.89682598466814,
|
| 266 |
-
"learning_rate": 4.5497748874437215e-05,
|
| 267 |
-
"loss": 0.6564,
|
| 268 |
-
"step": 370
|
| 269 |
-
},
|
| 270 |
-
{
|
| 271 |
-
"epoch": 0.47,
|
| 272 |
-
"grad_norm": 4.959965455950147,
|
| 273 |
-
"learning_rate": 4.5372686343171585e-05,
|
| 274 |
-
"loss": 0.6734,
|
| 275 |
-
"step": 380
|
| 276 |
-
},
|
| 277 |
-
{
|
| 278 |
-
"epoch": 0.49,
|
| 279 |
-
"grad_norm": 5.938681073424107,
|
| 280 |
-
"learning_rate": 4.5247623811905955e-05,
|
| 281 |
-
"loss": 0.7523,
|
| 282 |
-
"step": 390
|
| 283 |
-
},
|
| 284 |
-
{
|
| 285 |
-
"epoch": 0.5,
|
| 286 |
-
"grad_norm": 5.35665922615818,
|
| 287 |
-
"learning_rate": 4.5122561280640325e-05,
|
| 288 |
-
"loss": 0.7124,
|
| 289 |
-
"step": 400
|
| 290 |
-
},
|
| 291 |
-
{
|
| 292 |
-
"epoch": 0.51,
|
| 293 |
-
"grad_norm": 4.8123046042668625,
|
| 294 |
-
"learning_rate": 4.499749874937469e-05,
|
| 295 |
-
"loss": 0.702,
|
| 296 |
-
"step": 410
|
| 297 |
-
},
|
| 298 |
-
{
|
| 299 |
-
"epoch": 0.53,
|
| 300 |
-
"grad_norm": 5.411392852808034,
|
| 301 |
-
"learning_rate": 4.487243621810905e-05,
|
| 302 |
-
"loss": 0.6917,
|
| 303 |
-
"step": 420
|
| 304 |
-
},
|
| 305 |
-
{
|
| 306 |
-
"epoch": 0.54,
|
| 307 |
-
"grad_norm": 5.009053235776962,
|
| 308 |
-
"learning_rate": 4.474737368684342e-05,
|
| 309 |
-
"loss": 0.675,
|
| 310 |
-
"step": 430
|
| 311 |
-
},
|
| 312 |
-
{
|
| 313 |
-
"epoch": 0.55,
|
| 314 |
-
"grad_norm": 5.399248296904937,
|
| 315 |
-
"learning_rate": 4.462231115557779e-05,
|
| 316 |
-
"loss": 0.6916,
|
| 317 |
-
"step": 440
|
| 318 |
-
},
|
| 319 |
-
{
|
| 320 |
-
"epoch": 0.56,
|
| 321 |
-
"grad_norm": 5.195245498569367,
|
| 322 |
-
"learning_rate": 4.449724862431216e-05,
|
| 323 |
-
"loss": 0.6664,
|
| 324 |
-
"step": 450
|
| 325 |
-
},
|
| 326 |
-
{
|
| 327 |
-
"epoch": 0.57,
|
| 328 |
-
"grad_norm": 6.064114070634714,
|
| 329 |
-
"learning_rate": 4.4372186093046525e-05,
|
| 330 |
-
"loss": 0.7281,
|
| 331 |
-
"step": 460
|
| 332 |
-
},
|
| 333 |
-
{
|
| 334 |
-
"epoch": 0.59,
|
| 335 |
-
"grad_norm": 4.9950124660638044,
|
| 336 |
-
"learning_rate": 4.424712356178089e-05,
|
| 337 |
-
"loss": 0.6826,
|
| 338 |
-
"step": 470
|
| 339 |
-
},
|
| 340 |
-
{
|
| 341 |
-
"epoch": 0.6,
|
| 342 |
-
"grad_norm": 4.4353585247328,
|
| 343 |
-
"learning_rate": 4.412206103051526e-05,
|
| 344 |
-
"loss": 0.6919,
|
| 345 |
-
"step": 480
|
| 346 |
-
},
|
| 347 |
-
{
|
| 348 |
-
"epoch": 0.61,
|
| 349 |
-
"grad_norm": 5.289699318891135,
|
| 350 |
-
"learning_rate": 4.399699849924963e-05,
|
| 351 |
-
"loss": 0.689,
|
| 352 |
-
"step": 490
|
| 353 |
-
},
|
| 354 |
-
{
|
| 355 |
-
"epoch": 0.62,
|
| 356 |
-
"grad_norm": 6.126514792457231,
|
| 357 |
-
"learning_rate": 4.3871935967984e-05,
|
| 358 |
-
"loss": 0.7207,
|
| 359 |
-
"step": 500
|
| 360 |
-
},
|
| 361 |
-
{
|
| 362 |
-
"epoch": 0.64,
|
| 363 |
-
"grad_norm": 5.181000464013246,
|
| 364 |
-
"learning_rate": 4.374687343671836e-05,
|
| 365 |
-
"loss": 0.7047,
|
| 366 |
-
"step": 510
|
| 367 |
-
},
|
| 368 |
-
{
|
| 369 |
-
"epoch": 0.65,
|
| 370 |
-
"grad_norm": 6.228936674875526,
|
| 371 |
-
"learning_rate": 4.3621810905452725e-05,
|
| 372 |
-
"loss": 0.6799,
|
| 373 |
-
"step": 520
|
| 374 |
-
},
|
| 375 |
-
{
|
| 376 |
-
"epoch": 0.66,
|
| 377 |
-
"grad_norm": 4.499920526438516,
|
| 378 |
-
"learning_rate": 4.3496748374187095e-05,
|
| 379 |
-
"loss": 0.7035,
|
| 380 |
-
"step": 530
|
| 381 |
-
},
|
| 382 |
-
{
|
| 383 |
-
"epoch": 0.68,
|
| 384 |
-
"grad_norm": 5.258199737028568,
|
| 385 |
-
"learning_rate": 4.337168584292146e-05,
|
| 386 |
-
"loss": 0.7166,
|
| 387 |
-
"step": 540
|
| 388 |
-
},
|
| 389 |
-
{
|
| 390 |
-
"epoch": 0.69,
|
| 391 |
-
"grad_norm": 5.078292703060163,
|
| 392 |
-
"learning_rate": 4.324662331165583e-05,
|
| 393 |
-
"loss": 0.7128,
|
| 394 |
-
"step": 550
|
| 395 |
-
},
|
| 396 |
-
{
|
| 397 |
-
"epoch": 0.7,
|
| 398 |
-
"grad_norm": 5.150500612093525,
|
| 399 |
-
"learning_rate": 4.31215607803902e-05,
|
| 400 |
-
"loss": 0.7314,
|
| 401 |
-
"step": 560
|
| 402 |
-
},
|
| 403 |
-
{
|
| 404 |
-
"epoch": 0.71,
|
| 405 |
-
"grad_norm": 4.789457967587879,
|
| 406 |
-
"learning_rate": 4.299649824912456e-05,
|
| 407 |
-
"loss": 0.7351,
|
| 408 |
-
"step": 570
|
| 409 |
-
},
|
| 410 |
-
{
|
| 411 |
-
"epoch": 0.72,
|
| 412 |
-
"grad_norm": 5.462318518655166,
|
| 413 |
-
"learning_rate": 4.287143571785893e-05,
|
| 414 |
-
"loss": 0.7117,
|
| 415 |
-
"step": 580
|
| 416 |
-
},
|
| 417 |
-
{
|
| 418 |
-
"epoch": 0.74,
|
| 419 |
-
"grad_norm": 6.448282045561541,
|
| 420 |
-
"learning_rate": 4.2746373186593295e-05,
|
| 421 |
-
"loss": 0.7075,
|
| 422 |
-
"step": 590
|
| 423 |
-
},
|
| 424 |
-
{
|
| 425 |
-
"epoch": 0.75,
|
| 426 |
-
"grad_norm": 6.337874374016145,
|
| 427 |
-
"learning_rate": 4.2621310655327665e-05,
|
| 428 |
-
"loss": 0.7332,
|
| 429 |
-
"step": 600
|
| 430 |
-
},
|
| 431 |
-
{
|
| 432 |
-
"epoch": 0.76,
|
| 433 |
-
"grad_norm": 4.904355017265367,
|
| 434 |
-
"learning_rate": 4.2496248124062035e-05,
|
| 435 |
-
"loss": 0.6525,
|
| 436 |
-
"step": 610
|
| 437 |
-
},
|
| 438 |
-
{
|
| 439 |
-
"epoch": 0.78,
|
| 440 |
-
"grad_norm": 5.428105097653496,
|
| 441 |
-
"learning_rate": 4.2371185592796405e-05,
|
| 442 |
-
"loss": 0.734,
|
| 443 |
-
"step": 620
|
| 444 |
-
},
|
| 445 |
-
{
|
| 446 |
-
"epoch": 0.79,
|
| 447 |
-
"grad_norm": 5.489902764520662,
|
| 448 |
-
"learning_rate": 4.224612306153077e-05,
|
| 449 |
-
"loss": 0.665,
|
| 450 |
-
"step": 630
|
| 451 |
-
},
|
| 452 |
-
{
|
| 453 |
-
"epoch": 0.8,
|
| 454 |
-
"grad_norm": 4.81269221727027,
|
| 455 |
-
"learning_rate": 4.212106053026513e-05,
|
| 456 |
-
"loss": 0.7485,
|
| 457 |
-
"step": 640
|
| 458 |
-
},
|
| 459 |
-
{
|
| 460 |
-
"epoch": 0.81,
|
| 461 |
-
"grad_norm": 5.90221965868893,
|
| 462 |
-
"learning_rate": 4.19959979989995e-05,
|
| 463 |
-
"loss": 0.7342,
|
| 464 |
-
"step": 650
|
| 465 |
-
},
|
| 466 |
-
{
|
| 467 |
-
"epoch": 0.82,
|
| 468 |
-
"grad_norm": 5.7730450116082315,
|
| 469 |
-
"learning_rate": 4.1870935467733865e-05,
|
| 470 |
-
"loss": 0.7082,
|
| 471 |
-
"step": 660
|
| 472 |
-
},
|
| 473 |
-
{
|
| 474 |
-
"epoch": 0.84,
|
| 475 |
-
"grad_norm": 4.274199534648249,
|
| 476 |
-
"learning_rate": 4.174587293646824e-05,
|
| 477 |
-
"loss": 0.7262,
|
| 478 |
-
"step": 670
|
| 479 |
-
},
|
| 480 |
-
{
|
| 481 |
-
"epoch": 0.85,
|
| 482 |
-
"grad_norm": 5.348282169245115,
|
| 483 |
-
"learning_rate": 4.1620810405202605e-05,
|
| 484 |
-
"loss": 0.6719,
|
| 485 |
-
"step": 680
|
| 486 |
-
},
|
| 487 |
-
{
|
| 488 |
-
"epoch": 0.86,
|
| 489 |
-
"grad_norm": 5.920417706986164,
|
| 490 |
-
"learning_rate": 4.149574787393697e-05,
|
| 491 |
-
"loss": 0.7369,
|
| 492 |
-
"step": 690
|
| 493 |
-
},
|
| 494 |
-
{
|
| 495 |
-
"epoch": 0.88,
|
| 496 |
-
"grad_norm": 5.466281181017966,
|
| 497 |
-
"learning_rate": 4.137068534267134e-05,
|
| 498 |
-
"loss": 0.7526,
|
| 499 |
-
"step": 700
|
| 500 |
-
},
|
| 501 |
-
{
|
| 502 |
-
"epoch": 0.89,
|
| 503 |
-
"grad_norm": 5.421656535471705,
|
| 504 |
-
"learning_rate": 4.12456228114057e-05,
|
| 505 |
-
"loss": 0.6593,
|
| 506 |
-
"step": 710
|
| 507 |
-
},
|
| 508 |
-
{
|
| 509 |
-
"epoch": 0.9,
|
| 510 |
-
"grad_norm": 4.971718914811608,
|
| 511 |
-
"learning_rate": 4.112056028014007e-05,
|
| 512 |
-
"loss": 0.7003,
|
| 513 |
-
"step": 720
|
| 514 |
-
},
|
| 515 |
-
{
|
| 516 |
-
"epoch": 0.91,
|
| 517 |
-
"grad_norm": 5.770908323092229,
|
| 518 |
-
"learning_rate": 4.099549774887444e-05,
|
| 519 |
-
"loss": 0.7414,
|
| 520 |
-
"step": 730
|
| 521 |
-
},
|
| 522 |
-
{
|
| 523 |
-
"epoch": 0.93,
|
| 524 |
-
"grad_norm": 6.190279249754522,
|
| 525 |
-
"learning_rate": 4.0870435217608805e-05,
|
| 526 |
-
"loss": 0.7034,
|
| 527 |
-
"step": 740
|
| 528 |
-
},
|
| 529 |
-
{
|
| 530 |
-
"epoch": 0.94,
|
| 531 |
-
"grad_norm": 5.246490395350578,
|
| 532 |
-
"learning_rate": 4.0745372686343175e-05,
|
| 533 |
-
"loss": 0.6962,
|
| 534 |
-
"step": 750
|
| 535 |
-
},
|
| 536 |
-
{
|
| 537 |
-
"epoch": 0.95,
|
| 538 |
-
"grad_norm": 5.581112870841728,
|
| 539 |
-
"learning_rate": 4.062031015507754e-05,
|
| 540 |
-
"loss": 0.7137,
|
| 541 |
-
"step": 760
|
| 542 |
-
},
|
| 543 |
-
{
|
| 544 |
-
"epoch": 0.96,
|
| 545 |
-
"grad_norm": 5.4016319352096325,
|
| 546 |
-
"learning_rate": 4.049524762381191e-05,
|
| 547 |
-
"loss": 0.7035,
|
| 548 |
-
"step": 770
|
| 549 |
-
},
|
| 550 |
-
{
|
| 551 |
-
"epoch": 0.97,
|
| 552 |
-
"grad_norm": 4.461968885599671,
|
| 553 |
-
"learning_rate": 4.037018509254628e-05,
|
| 554 |
-
"loss": 0.7133,
|
| 555 |
-
"step": 780
|
| 556 |
-
},
|
| 557 |
-
{
|
| 558 |
-
"epoch": 0.99,
|
| 559 |
-
"grad_norm": 4.644183814012947,
|
| 560 |
-
"learning_rate": 4.024512256128064e-05,
|
| 561 |
-
"loss": 0.7161,
|
| 562 |
-
"step": 790
|
| 563 |
-
},
|
| 564 |
-
{
|
| 565 |
-
"epoch": 1.0,
|
| 566 |
-
"grad_norm": 4.826017148028126,
|
| 567 |
-
"learning_rate": 4.012006003001501e-05,
|
| 568 |
-
"loss": 0.722,
|
| 569 |
-
"step": 800
|
| 570 |
-
},
|
| 571 |
-
{
|
| 572 |
-
"epoch": 1.01,
|
| 573 |
-
"grad_norm": 4.000342354428771,
|
| 574 |
-
"learning_rate": 3.9994997498749375e-05,
|
| 575 |
-
"loss": 0.3054,
|
| 576 |
-
"step": 810
|
| 577 |
-
},
|
| 578 |
-
{
|
| 579 |
-
"epoch": 1.02,
|
| 580 |
-
"grad_norm": 3.8374628270240616,
|
| 581 |
-
"learning_rate": 3.9869934967483745e-05,
|
| 582 |
-
"loss": 0.3016,
|
| 583 |
-
"step": 820
|
| 584 |
-
},
|
| 585 |
-
{
|
| 586 |
-
"epoch": 1.04,
|
| 587 |
-
"grad_norm": 5.107383783318757,
|
| 588 |
-
"learning_rate": 3.974487243621811e-05,
|
| 589 |
-
"loss": 0.3077,
|
| 590 |
-
"step": 830
|
| 591 |
-
},
|
| 592 |
-
{
|
| 593 |
-
"epoch": 1.05,
|
| 594 |
-
"grad_norm": 3.4427943293478123,
|
| 595 |
-
"learning_rate": 3.961980990495248e-05,
|
| 596 |
-
"loss": 0.2966,
|
| 597 |
-
"step": 840
|
| 598 |
-
},
|
| 599 |
-
{
|
| 600 |
-
"epoch": 1.06,
|
| 601 |
-
"grad_norm": 3.8325782253297174,
|
| 602 |
-
"learning_rate": 3.949474737368685e-05,
|
| 603 |
-
"loss": 0.3203,
|
| 604 |
-
"step": 850
|
| 605 |
-
},
|
| 606 |
-
{
|
| 607 |
-
"epoch": 1.07,
|
| 608 |
-
"grad_norm": 4.236035011038753,
|
| 609 |
-
"learning_rate": 3.936968484242121e-05,
|
| 610 |
-
"loss": 0.3019,
|
| 611 |
-
"step": 860
|
| 612 |
-
},
|
| 613 |
-
{
|
| 614 |
-
"epoch": 1.09,
|
| 615 |
-
"grad_norm": 4.491576469992581,
|
| 616 |
-
"learning_rate": 3.924462231115558e-05,
|
| 617 |
-
"loss": 0.325,
|
| 618 |
-
"step": 870
|
| 619 |
-
},
|
| 620 |
-
{
|
| 621 |
-
"epoch": 1.1,
|
| 622 |
-
"grad_norm": 3.961749774177233,
|
| 623 |
-
"learning_rate": 3.9119559779889945e-05,
|
| 624 |
-
"loss": 0.3371,
|
| 625 |
-
"step": 880
|
| 626 |
-
},
|
| 627 |
-
{
|
| 628 |
-
"epoch": 1.11,
|
| 629 |
-
"grad_norm": 4.672476739310163,
|
| 630 |
-
"learning_rate": 3.8994497248624315e-05,
|
| 631 |
-
"loss": 0.3306,
|
| 632 |
-
"step": 890
|
| 633 |
-
},
|
| 634 |
-
{
|
| 635 |
-
"epoch": 1.12,
|
| 636 |
-
"grad_norm": 3.810555384248572,
|
| 637 |
-
"learning_rate": 3.8869434717358685e-05,
|
| 638 |
-
"loss": 0.2961,
|
| 639 |
-
"step": 900
|
| 640 |
-
},
|
| 641 |
-
{
|
| 642 |
-
"epoch": 1.14,
|
| 643 |
-
"grad_norm": 3.662542292119634,
|
| 644 |
-
"learning_rate": 3.874437218609305e-05,
|
| 645 |
-
"loss": 0.3159,
|
| 646 |
-
"step": 910
|
| 647 |
-
},
|
| 648 |
-
{
|
| 649 |
-
"epoch": 1.15,
|
| 650 |
-
"grad_norm": 4.615080015205904,
|
| 651 |
-
"learning_rate": 3.861930965482742e-05,
|
| 652 |
-
"loss": 0.3321,
|
| 653 |
-
"step": 920
|
| 654 |
-
},
|
| 655 |
-
{
|
| 656 |
-
"epoch": 1.16,
|
| 657 |
-
"grad_norm": 3.5514929197773317,
|
| 658 |
-
"learning_rate": 3.849424712356178e-05,
|
| 659 |
-
"loss": 0.316,
|
| 660 |
-
"step": 930
|
| 661 |
-
},
|
| 662 |
-
{
|
| 663 |
-
"epoch": 1.18,
|
| 664 |
-
"grad_norm": 3.925218593074527,
|
| 665 |
-
"learning_rate": 3.8369184592296145e-05,
|
| 666 |
-
"loss": 0.3539,
|
| 667 |
-
"step": 940
|
| 668 |
-
},
|
| 669 |
-
{
|
| 670 |
-
"epoch": 1.19,
|
| 671 |
-
"grad_norm": 3.9002231338443205,
|
| 672 |
-
"learning_rate": 3.824412206103052e-05,
|
| 673 |
-
"loss": 0.3293,
|
| 674 |
-
"step": 950
|
| 675 |
-
},
|
| 676 |
-
{
|
| 677 |
-
"epoch": 1.2,
|
| 678 |
-
"grad_norm": 3.8671256012249478,
|
| 679 |
-
"learning_rate": 3.8119059529764885e-05,
|
| 680 |
-
"loss": 0.3169,
|
| 681 |
-
"step": 960
|
| 682 |
-
},
|
| 683 |
-
{
|
| 684 |
-
"epoch": 1.21,
|
| 685 |
-
"grad_norm": 3.61716640787587,
|
| 686 |
-
"learning_rate": 3.7993996998499255e-05,
|
| 687 |
-
"loss": 0.3321,
|
| 688 |
-
"step": 970
|
| 689 |
-
},
|
| 690 |
-
{
|
| 691 |
-
"epoch": 1.23,
|
| 692 |
-
"grad_norm": 4.137379652042079,
|
| 693 |
-
"learning_rate": 3.786893446723362e-05,
|
| 694 |
-
"loss": 0.3418,
|
| 695 |
-
"step": 980
|
| 696 |
-
},
|
| 697 |
-
{
|
| 698 |
-
"epoch": 1.24,
|
| 699 |
-
"grad_norm": 4.191073131566738,
|
| 700 |
-
"learning_rate": 3.774387193596798e-05,
|
| 701 |
-
"loss": 0.3277,
|
| 702 |
-
"step": 990
|
| 703 |
-
},
|
| 704 |
-
{
|
| 705 |
-
"epoch": 1.25,
|
| 706 |
-
"grad_norm": 4.654388465252297,
|
| 707 |
-
"learning_rate": 3.761880940470235e-05,
|
| 708 |
-
"loss": 0.3371,
|
| 709 |
-
"step": 1000
|
| 710 |
-
},
|
| 711 |
-
{
|
| 712 |
-
"epoch": 1.26,
|
| 713 |
-
"grad_norm": 4.900788395574059,
|
| 714 |
-
"learning_rate": 3.749374687343672e-05,
|
| 715 |
-
"loss": 0.347,
|
| 716 |
-
"step": 1010
|
| 717 |
-
},
|
| 718 |
-
{
|
| 719 |
-
"epoch": 1.27,
|
| 720 |
-
"grad_norm": 3.990193385932706,
|
| 721 |
-
"learning_rate": 3.736868434217109e-05,
|
| 722 |
-
"loss": 0.3358,
|
| 723 |
-
"step": 1020
|
| 724 |
-
},
|
| 725 |
-
{
|
| 726 |
-
"epoch": 1.29,
|
| 727 |
-
"grad_norm": 4.270444278519215,
|
| 728 |
-
"learning_rate": 3.7243621810905455e-05,
|
| 729 |
-
"loss": 0.3461,
|
| 730 |
-
"step": 1030
|
| 731 |
-
},
|
| 732 |
-
{
|
| 733 |
-
"epoch": 1.3,
|
| 734 |
-
"grad_norm": 4.286422166946011,
|
| 735 |
-
"learning_rate": 3.711855927963982e-05,
|
| 736 |
-
"loss": 0.3362,
|
| 737 |
-
"step": 1040
|
| 738 |
-
},
|
| 739 |
-
{
|
| 740 |
-
"epoch": 1.31,
|
| 741 |
-
"grad_norm": 4.78341924787361,
|
| 742 |
-
"learning_rate": 3.699349674837419e-05,
|
| 743 |
-
"loss": 0.3418,
|
| 744 |
-
"step": 1050
|
| 745 |
-
},
|
| 746 |
-
{
|
| 747 |
-
"epoch": 1.32,
|
| 748 |
-
"grad_norm": 4.193755303051207,
|
| 749 |
-
"learning_rate": 3.686843421710856e-05,
|
| 750 |
-
"loss": 0.3372,
|
| 751 |
-
"step": 1060
|
| 752 |
-
},
|
| 753 |
-
{
|
| 754 |
-
"epoch": 1.34,
|
| 755 |
-
"grad_norm": 3.989814545760004,
|
| 756 |
-
"learning_rate": 3.674337168584293e-05,
|
| 757 |
-
"loss": 0.3254,
|
| 758 |
-
"step": 1070
|
| 759 |
-
},
|
| 760 |
-
{
|
| 761 |
-
"epoch": 1.35,
|
| 762 |
-
"grad_norm": 3.823337597588243,
|
| 763 |
-
"learning_rate": 3.661830915457729e-05,
|
| 764 |
-
"loss": 0.3471,
|
| 765 |
-
"step": 1080
|
| 766 |
-
},
|
| 767 |
-
{
|
| 768 |
-
"epoch": 1.36,
|
| 769 |
-
"grad_norm": 3.793526762626974,
|
| 770 |
-
"learning_rate": 3.6493246623311655e-05,
|
| 771 |
-
"loss": 0.3432,
|
| 772 |
-
"step": 1090
|
| 773 |
-
},
|
| 774 |
-
{
|
| 775 |
-
"epoch": 1.38,
|
| 776 |
-
"grad_norm": 3.872109227027026,
|
| 777 |
-
"learning_rate": 3.6368184092046025e-05,
|
| 778 |
-
"loss": 0.3533,
|
| 779 |
-
"step": 1100
|
| 780 |
-
},
|
| 781 |
-
{
|
| 782 |
-
"epoch": 1.39,
|
| 783 |
-
"grad_norm": 4.437307219280221,
|
| 784 |
-
"learning_rate": 3.624312156078039e-05,
|
| 785 |
-
"loss": 0.3551,
|
| 786 |
-
"step": 1110
|
| 787 |
-
},
|
| 788 |
-
{
|
| 789 |
-
"epoch": 1.4,
|
| 790 |
-
"grad_norm": 4.298491572681636,
|
| 791 |
-
"learning_rate": 3.611805902951476e-05,
|
| 792 |
-
"loss": 0.3321,
|
| 793 |
-
"step": 1120
|
| 794 |
-
},
|
| 795 |
-
{
|
| 796 |
-
"epoch": 1.41,
|
| 797 |
-
"grad_norm": 3.9890804256002084,
|
| 798 |
-
"learning_rate": 3.599299649824913e-05,
|
| 799 |
-
"loss": 0.3339,
|
| 800 |
-
"step": 1130
|
| 801 |
-
},
|
| 802 |
-
{
|
| 803 |
-
"epoch": 1.43,
|
| 804 |
-
"grad_norm": 4.437500429825024,
|
| 805 |
-
"learning_rate": 3.586793396698349e-05,
|
| 806 |
-
"loss": 0.3272,
|
| 807 |
-
"step": 1140
|
| 808 |
-
},
|
| 809 |
-
{
|
| 810 |
-
"epoch": 1.44,
|
| 811 |
-
"grad_norm": 4.218831379423431,
|
| 812 |
-
"learning_rate": 3.574287143571786e-05,
|
| 813 |
-
"loss": 0.3527,
|
| 814 |
-
"step": 1150
|
| 815 |
-
},
|
| 816 |
-
{
|
| 817 |
-
"epoch": 1.45,
|
| 818 |
-
"grad_norm": 3.747375332703581,
|
| 819 |
-
"learning_rate": 3.5617808904452225e-05,
|
| 820 |
-
"loss": 0.3513,
|
| 821 |
-
"step": 1160
|
| 822 |
-
},
|
| 823 |
-
{
|
| 824 |
-
"epoch": 1.46,
|
| 825 |
-
"grad_norm": 4.0875968991614355,
|
| 826 |
-
"learning_rate": 3.5492746373186595e-05,
|
| 827 |
-
"loss": 0.339,
|
| 828 |
-
"step": 1170
|
| 829 |
-
},
|
| 830 |
-
{
|
| 831 |
-
"epoch": 1.48,
|
| 832 |
-
"grad_norm": 4.039233442042736,
|
| 833 |
-
"learning_rate": 3.5367683841920965e-05,
|
| 834 |
-
"loss": 0.3511,
|
| 835 |
-
"step": 1180
|
| 836 |
-
},
|
| 837 |
-
{
|
| 838 |
-
"epoch": 1.49,
|
| 839 |
-
"grad_norm": 3.5825847169026135,
|
| 840 |
-
"learning_rate": 3.524262131065533e-05,
|
| 841 |
-
"loss": 0.3349,
|
| 842 |
-
"step": 1190
|
| 843 |
-
},
|
| 844 |
-
{
|
| 845 |
-
"epoch": 1.5,
|
| 846 |
-
"grad_norm": 4.175820759942294,
|
| 847 |
-
"learning_rate": 3.51175587793897e-05,
|
| 848 |
-
"loss": 0.3619,
|
| 849 |
-
"step": 1200
|
| 850 |
-
},
|
| 851 |
-
{
|
| 852 |
-
"epoch": 1.51,
|
| 853 |
-
"grad_norm": 4.714008298365686,
|
| 854 |
-
"learning_rate": 3.499249624812406e-05,
|
| 855 |
-
"loss": 0.3487,
|
| 856 |
-
"step": 1210
|
| 857 |
-
},
|
| 858 |
-
{
|
| 859 |
-
"epoch": 1.52,
|
| 860 |
-
"grad_norm": 3.9156939062827925,
|
| 861 |
-
"learning_rate": 3.486743371685843e-05,
|
| 862 |
-
"loss": 0.3293,
|
| 863 |
-
"step": 1220
|
| 864 |
-
},
|
| 865 |
-
{
|
| 866 |
-
"epoch": 1.54,
|
| 867 |
-
"grad_norm": 3.922432544485633,
|
| 868 |
-
"learning_rate": 3.4742371185592795e-05,
|
| 869 |
-
"loss": 0.3424,
|
| 870 |
-
"step": 1230
|
| 871 |
-
},
|
| 872 |
-
{
|
| 873 |
-
"epoch": 1.55,
|
| 874 |
-
"grad_norm": 3.871337267070979,
|
| 875 |
-
"learning_rate": 3.4617308654327165e-05,
|
| 876 |
-
"loss": 0.3453,
|
| 877 |
-
"step": 1240
|
| 878 |
-
},
|
| 879 |
-
{
|
| 880 |
-
"epoch": 1.56,
|
| 881 |
-
"grad_norm": 4.563498191855707,
|
| 882 |
-
"learning_rate": 3.4492246123061535e-05,
|
| 883 |
-
"loss": 0.3576,
|
| 884 |
-
"step": 1250
|
| 885 |
-
},
|
| 886 |
-
{
|
| 887 |
-
"epoch": 1.57,
|
| 888 |
-
"grad_norm": 4.11645130618609,
|
| 889 |
-
"learning_rate": 3.43671835917959e-05,
|
| 890 |
-
"loss": 0.3597,
|
| 891 |
-
"step": 1260
|
| 892 |
-
},
|
| 893 |
-
{
|
| 894 |
-
"epoch": 1.59,
|
| 895 |
-
"grad_norm": 4.38717548441809,
|
| 896 |
-
"learning_rate": 3.424212106053027e-05,
|
| 897 |
-
"loss": 0.3274,
|
| 898 |
-
"step": 1270
|
| 899 |
-
},
|
| 900 |
-
{
|
| 901 |
-
"epoch": 1.6,
|
| 902 |
-
"grad_norm": 4.400823403388009,
|
| 903 |
-
"learning_rate": 3.411705852926463e-05,
|
| 904 |
-
"loss": 0.345,
|
| 905 |
-
"step": 1280
|
| 906 |
-
},
|
| 907 |
-
{
|
| 908 |
-
"epoch": 1.61,
|
| 909 |
-
"grad_norm": 4.767739149075579,
|
| 910 |
-
"learning_rate": 3.3991995997999e-05,
|
| 911 |
-
"loss": 0.3432,
|
| 912 |
-
"step": 1290
|
| 913 |
-
},
|
| 914 |
-
{
|
| 915 |
-
"epoch": 1.62,
|
| 916 |
-
"grad_norm": 3.879637250755376,
|
| 917 |
-
"learning_rate": 3.386693346673337e-05,
|
| 918 |
-
"loss": 0.3643,
|
| 919 |
-
"step": 1300
|
| 920 |
-
},
|
| 921 |
-
{
|
| 922 |
-
"epoch": 1.64,
|
| 923 |
-
"grad_norm": 3.7442488756684345,
|
| 924 |
-
"learning_rate": 3.3741870935467735e-05,
|
| 925 |
-
"loss": 0.3488,
|
| 926 |
-
"step": 1310
|
| 927 |
-
},
|
| 928 |
-
{
|
| 929 |
-
"epoch": 1.65,
|
| 930 |
-
"grad_norm": 3.8345905122140884,
|
| 931 |
-
"learning_rate": 3.3616808404202105e-05,
|
| 932 |
-
"loss": 0.3505,
|
| 933 |
-
"step": 1320
|
| 934 |
-
},
|
| 935 |
-
{
|
| 936 |
-
"epoch": 1.66,
|
| 937 |
-
"grad_norm": 4.9315147338510945,
|
| 938 |
-
"learning_rate": 3.349174587293647e-05,
|
| 939 |
-
"loss": 0.3577,
|
| 940 |
-
"step": 1330
|
| 941 |
-
},
|
| 942 |
-
{
|
| 943 |
-
"epoch": 1.68,
|
| 944 |
-
"grad_norm": 4.151768383452871,
|
| 945 |
-
"learning_rate": 3.336668334167083e-05,
|
| 946 |
-
"loss": 0.3471,
|
| 947 |
-
"step": 1340
|
| 948 |
-
},
|
| 949 |
-
{
|
| 950 |
-
"epoch": 1.69,
|
| 951 |
-
"grad_norm": 3.751703511185265,
|
| 952 |
-
"learning_rate": 3.324162081040521e-05,
|
| 953 |
-
"loss": 0.3489,
|
| 954 |
-
"step": 1350
|
| 955 |
-
},
|
| 956 |
-
{
|
| 957 |
-
"epoch": 1.7,
|
| 958 |
-
"grad_norm": 4.52533794005262,
|
| 959 |
-
"learning_rate": 3.311655827913957e-05,
|
| 960 |
-
"loss": 0.3362,
|
| 961 |
-
"step": 1360
|
| 962 |
-
},
|
| 963 |
-
{
|
| 964 |
-
"epoch": 1.71,
|
| 965 |
-
"grad_norm": 3.938063323883952,
|
| 966 |
-
"learning_rate": 3.299149574787394e-05,
|
| 967 |
-
"loss": 0.3496,
|
| 968 |
-
"step": 1370
|
| 969 |
-
},
|
| 970 |
-
{
|
| 971 |
-
"epoch": 1.73,
|
| 972 |
-
"grad_norm": 4.321107980845758,
|
| 973 |
-
"learning_rate": 3.2866433216608305e-05,
|
| 974 |
-
"loss": 0.3565,
|
| 975 |
-
"step": 1380
|
| 976 |
-
},
|
| 977 |
-
{
|
| 978 |
-
"epoch": 1.74,
|
| 979 |
-
"grad_norm": 4.180361684455809,
|
| 980 |
-
"learning_rate": 3.274137068534267e-05,
|
| 981 |
-
"loss": 0.3419,
|
| 982 |
-
"step": 1390
|
| 983 |
-
},
|
| 984 |
-
{
|
| 985 |
-
"epoch": 1.75,
|
| 986 |
-
"grad_norm": 4.073762043329108,
|
| 987 |
-
"learning_rate": 3.261630815407704e-05,
|
| 988 |
-
"loss": 0.3368,
|
| 989 |
-
"step": 1400
|
| 990 |
-
},
|
| 991 |
-
{
|
| 992 |
-
"epoch": 1.76,
|
| 993 |
-
"grad_norm": 3.7744796293221623,
|
| 994 |
-
"learning_rate": 3.249124562281141e-05,
|
| 995 |
-
"loss": 0.34,
|
| 996 |
-
"step": 1410
|
| 997 |
-
},
|
| 998 |
-
{
|
| 999 |
-
"epoch": 1.77,
|
| 1000 |
-
"grad_norm": 4.272009909229001,
|
| 1001 |
-
"learning_rate": 3.236618309154578e-05,
|
| 1002 |
-
"loss": 0.346,
|
| 1003 |
-
"step": 1420
|
| 1004 |
-
},
|
| 1005 |
-
{
|
| 1006 |
-
"epoch": 1.79,
|
| 1007 |
-
"grad_norm": 4.400007005165766,
|
| 1008 |
-
"learning_rate": 3.224112056028014e-05,
|
| 1009 |
-
"loss": 0.3422,
|
| 1010 |
-
"step": 1430
|
| 1011 |
-
},
|
| 1012 |
-
{
|
| 1013 |
-
"epoch": 1.8,
|
| 1014 |
-
"grad_norm": 4.245867346769938,
|
| 1015 |
-
"learning_rate": 3.2116058029014505e-05,
|
| 1016 |
-
"loss": 0.3457,
|
| 1017 |
-
"step": 1440
|
| 1018 |
-
},
|
| 1019 |
-
{
|
| 1020 |
-
"epoch": 1.81,
|
| 1021 |
-
"grad_norm": 4.076869025792562,
|
| 1022 |
-
"learning_rate": 3.1990995497748875e-05,
|
| 1023 |
-
"loss": 0.3428,
|
| 1024 |
-
"step": 1450
|
| 1025 |
-
},
|
| 1026 |
-
{
|
| 1027 |
-
"epoch": 1.82,
|
| 1028 |
-
"grad_norm": 3.992331306271212,
|
| 1029 |
-
"learning_rate": 3.1865932966483245e-05,
|
| 1030 |
-
"loss": 0.3423,
|
| 1031 |
-
"step": 1460
|
| 1032 |
-
},
|
| 1033 |
-
{
|
| 1034 |
-
"epoch": 1.84,
|
| 1035 |
-
"grad_norm": 4.329708099040903,
|
| 1036 |
-
"learning_rate": 3.1740870435217615e-05,
|
| 1037 |
-
"loss": 0.3353,
|
| 1038 |
-
"step": 1470
|
| 1039 |
-
},
|
| 1040 |
-
{
|
| 1041 |
-
"epoch": 1.85,
|
| 1042 |
-
"grad_norm": 4.228767990087569,
|
| 1043 |
-
"learning_rate": 3.161580790395198e-05,
|
| 1044 |
-
"loss": 0.3313,
|
| 1045 |
-
"step": 1480
|
| 1046 |
-
},
|
| 1047 |
-
{
|
| 1048 |
-
"epoch": 1.86,
|
| 1049 |
-
"grad_norm": 4.100143216120651,
|
| 1050 |
-
"learning_rate": 3.149074537268634e-05,
|
| 1051 |
-
"loss": 0.3448,
|
| 1052 |
-
"step": 1490
|
| 1053 |
-
},
|
| 1054 |
-
{
|
| 1055 |
-
"epoch": 1.88,
|
| 1056 |
-
"grad_norm": 4.846420241173142,
|
| 1057 |
-
"learning_rate": 3.136568284142071e-05,
|
| 1058 |
-
"loss": 0.3515,
|
| 1059 |
-
"step": 1500
|
| 1060 |
-
},
|
| 1061 |
-
{
|
| 1062 |
-
"epoch": 1.89,
|
| 1063 |
-
"grad_norm": 4.287502873811787,
|
| 1064 |
-
"learning_rate": 3.1240620310155075e-05,
|
| 1065 |
-
"loss": 0.3462,
|
| 1066 |
-
"step": 1510
|
| 1067 |
-
},
|
| 1068 |
-
{
|
| 1069 |
-
"epoch": 1.9,
|
| 1070 |
-
"grad_norm": 3.769572994020805,
|
| 1071 |
-
"learning_rate": 3.111555777888945e-05,
|
| 1072 |
-
"loss": 0.3593,
|
| 1073 |
-
"step": 1520
|
| 1074 |
-
},
|
| 1075 |
-
{
|
| 1076 |
-
"epoch": 1.91,
|
| 1077 |
-
"grad_norm": 3.7731122178800645,
|
| 1078 |
-
"learning_rate": 3.0990495247623815e-05,
|
| 1079 |
-
"loss": 0.3467,
|
| 1080 |
-
"step": 1530
|
| 1081 |
-
},
|
| 1082 |
-
{
|
| 1083 |
-
"epoch": 1.93,
|
| 1084 |
-
"grad_norm": 4.257071726342832,
|
| 1085 |
-
"learning_rate": 3.086543271635818e-05,
|
| 1086 |
-
"loss": 0.3447,
|
| 1087 |
-
"step": 1540
|
| 1088 |
-
},
|
| 1089 |
-
{
|
| 1090 |
-
"epoch": 1.94,
|
| 1091 |
-
"grad_norm": 4.192530556773847,
|
| 1092 |
-
"learning_rate": 3.074037018509255e-05,
|
| 1093 |
-
"loss": 0.3575,
|
| 1094 |
-
"step": 1550
|
| 1095 |
-
},
|
| 1096 |
-
{
|
| 1097 |
-
"epoch": 1.95,
|
| 1098 |
-
"grad_norm": 4.488222817355165,
|
| 1099 |
-
"learning_rate": 3.061530765382691e-05,
|
| 1100 |
-
"loss": 0.3392,
|
| 1101 |
-
"step": 1560
|
| 1102 |
-
},
|
| 1103 |
-
{
|
| 1104 |
-
"epoch": 1.96,
|
| 1105 |
-
"grad_norm": 4.166714909592134,
|
| 1106 |
-
"learning_rate": 3.0490245122561285e-05,
|
| 1107 |
-
"loss": 0.341,
|
| 1108 |
-
"step": 1570
|
| 1109 |
-
},
|
| 1110 |
-
{
|
| 1111 |
-
"epoch": 1.98,
|
| 1112 |
-
"grad_norm": 4.2906704582812445,
|
| 1113 |
-
"learning_rate": 3.0365182591295648e-05,
|
| 1114 |
-
"loss": 0.3337,
|
| 1115 |
-
"step": 1580
|
| 1116 |
-
},
|
| 1117 |
-
{
|
| 1118 |
-
"epoch": 1.99,
|
| 1119 |
-
"grad_norm": 4.631668644360512,
|
| 1120 |
-
"learning_rate": 3.0240120060030014e-05,
|
| 1121 |
-
"loss": 0.3352,
|
| 1122 |
-
"step": 1590
|
| 1123 |
-
},
|
| 1124 |
-
{
|
| 1125 |
-
"epoch": 2.0,
|
| 1126 |
-
"grad_norm": 3.6817020158177343,
|
| 1127 |
-
"learning_rate": 3.0115057528764385e-05,
|
| 1128 |
-
"loss": 0.338,
|
| 1129 |
-
"step": 1600
|
| 1130 |
-
},
|
| 1131 |
-
{
|
| 1132 |
-
"epoch": 2.01,
|
| 1133 |
-
"grad_norm": 2.4770673844893736,
|
| 1134 |
-
"learning_rate": 2.9989994997498748e-05,
|
| 1135 |
-
"loss": 0.1203,
|
| 1136 |
-
"step": 1610
|
| 1137 |
-
},
|
| 1138 |
-
{
|
| 1139 |
-
"epoch": 2.02,
|
| 1140 |
-
"grad_norm": 2.8486557450546752,
|
| 1141 |
-
"learning_rate": 2.986493246623312e-05,
|
| 1142 |
-
"loss": 0.1226,
|
| 1143 |
-
"step": 1620
|
| 1144 |
-
},
|
| 1145 |
-
{
|
| 1146 |
-
"epoch": 2.04,
|
| 1147 |
-
"grad_norm": 2.094604545543018,
|
| 1148 |
-
"learning_rate": 2.9739869934967484e-05,
|
| 1149 |
-
"loss": 0.1236,
|
| 1150 |
-
"step": 1630
|
| 1151 |
-
},
|
| 1152 |
-
{
|
| 1153 |
-
"epoch": 2.05,
|
| 1154 |
-
"grad_norm": 2.892529659135666,
|
| 1155 |
-
"learning_rate": 2.961480740370185e-05,
|
| 1156 |
-
"loss": 0.1242,
|
| 1157 |
-
"step": 1640
|
| 1158 |
-
},
|
| 1159 |
-
{
|
| 1160 |
-
"epoch": 2.06,
|
| 1161 |
-
"grad_norm": 3.0607648332468993,
|
| 1162 |
-
"learning_rate": 2.948974487243622e-05,
|
| 1163 |
-
"loss": 0.1243,
|
| 1164 |
-
"step": 1650
|
| 1165 |
-
},
|
| 1166 |
-
{
|
| 1167 |
-
"epoch": 2.08,
|
| 1168 |
-
"grad_norm": 3.31204465219556,
|
| 1169 |
-
"learning_rate": 2.9364682341170584e-05,
|
| 1170 |
-
"loss": 0.1235,
|
| 1171 |
-
"step": 1660
|
| 1172 |
-
},
|
| 1173 |
-
{
|
| 1174 |
-
"epoch": 2.09,
|
| 1175 |
-
"grad_norm": 2.4729852201434306,
|
| 1176 |
-
"learning_rate": 2.9239619809904954e-05,
|
| 1177 |
-
"loss": 0.1226,
|
| 1178 |
-
"step": 1670
|
| 1179 |
-
},
|
| 1180 |
-
{
|
| 1181 |
-
"epoch": 2.1,
|
| 1182 |
-
"grad_norm": 2.5257406678082583,
|
| 1183 |
-
"learning_rate": 2.911455727863932e-05,
|
| 1184 |
-
"loss": 0.1207,
|
| 1185 |
-
"step": 1680
|
| 1186 |
-
},
|
| 1187 |
-
{
|
| 1188 |
-
"epoch": 2.11,
|
| 1189 |
-
"grad_norm": 3.103672799030984,
|
| 1190 |
-
"learning_rate": 2.8989494747373684e-05,
|
| 1191 |
-
"loss": 0.1324,
|
| 1192 |
-
"step": 1690
|
| 1193 |
-
},
|
| 1194 |
-
{
|
| 1195 |
-
"epoch": 2.12,
|
| 1196 |
-
"grad_norm": 2.3782595301910043,
|
| 1197 |
-
"learning_rate": 2.8864432216108054e-05,
|
| 1198 |
-
"loss": 0.132,
|
| 1199 |
-
"step": 1700
|
| 1200 |
-
},
|
| 1201 |
-
{
|
| 1202 |
-
"epoch": 2.14,
|
| 1203 |
-
"grad_norm": 2.6059740025591505,
|
| 1204 |
-
"learning_rate": 2.873936968484242e-05,
|
| 1205 |
-
"loss": 0.1264,
|
| 1206 |
-
"step": 1710
|
| 1207 |
-
},
|
| 1208 |
-
{
|
| 1209 |
-
"epoch": 2.15,
|
| 1210 |
-
"grad_norm": 2.538398820262338,
|
| 1211 |
-
"learning_rate": 2.861430715357679e-05,
|
| 1212 |
-
"loss": 0.1366,
|
| 1213 |
-
"step": 1720
|
| 1214 |
-
},
|
| 1215 |
-
{
|
| 1216 |
-
"epoch": 2.16,
|
| 1217 |
-
"grad_norm": 2.499812214474419,
|
| 1218 |
-
"learning_rate": 2.8489244622311158e-05,
|
| 1219 |
-
"loss": 0.1377,
|
| 1220 |
-
"step": 1730
|
| 1221 |
-
},
|
| 1222 |
-
{
|
| 1223 |
-
"epoch": 2.17,
|
| 1224 |
-
"grad_norm": 2.3941826372137305,
|
| 1225 |
-
"learning_rate": 2.836418209104552e-05,
|
| 1226 |
-
"loss": 0.1225,
|
| 1227 |
-
"step": 1740
|
| 1228 |
-
},
|
| 1229 |
-
{
|
| 1230 |
-
"epoch": 2.19,
|
| 1231 |
-
"grad_norm": 2.6244609824311014,
|
| 1232 |
-
"learning_rate": 2.823911955977989e-05,
|
| 1233 |
-
"loss": 0.1466,
|
| 1234 |
-
"step": 1750
|
| 1235 |
-
},
|
| 1236 |
-
{
|
| 1237 |
-
"epoch": 2.2,
|
| 1238 |
-
"grad_norm": 2.5276124979370915,
|
| 1239 |
-
"learning_rate": 2.8114057028514258e-05,
|
| 1240 |
-
"loss": 0.1415,
|
| 1241 |
-
"step": 1760
|
| 1242 |
-
},
|
| 1243 |
-
{
|
| 1244 |
-
"epoch": 2.21,
|
| 1245 |
-
"grad_norm": 2.916618637416117,
|
| 1246 |
-
"learning_rate": 2.7988994497248628e-05,
|
| 1247 |
-
"loss": 0.1309,
|
| 1248 |
-
"step": 1770
|
| 1249 |
-
},
|
| 1250 |
-
{
|
| 1251 |
-
"epoch": 2.23,
|
| 1252 |
-
"grad_norm": 2.753008410683784,
|
| 1253 |
-
"learning_rate": 2.786393196598299e-05,
|
| 1254 |
-
"loss": 0.1432,
|
| 1255 |
-
"step": 1780
|
| 1256 |
-
},
|
| 1257 |
-
{
|
| 1258 |
-
"epoch": 2.24,
|
| 1259 |
-
"grad_norm": 3.1016622762701935,
|
| 1260 |
-
"learning_rate": 2.7738869434717364e-05,
|
| 1261 |
-
"loss": 0.1366,
|
| 1262 |
-
"step": 1790
|
| 1263 |
-
},
|
| 1264 |
-
{
|
| 1265 |
-
"epoch": 2.25,
|
| 1266 |
-
"grad_norm": 2.7966281206508032,
|
| 1267 |
-
"learning_rate": 2.7613806903451728e-05,
|
| 1268 |
-
"loss": 0.1408,
|
| 1269 |
-
"step": 1800
|
| 1270 |
-
},
|
| 1271 |
-
{
|
| 1272 |
-
"epoch": 2.26,
|
| 1273 |
-
"grad_norm": 2.439577928684241,
|
| 1274 |
-
"learning_rate": 2.7488744372186094e-05,
|
| 1275 |
-
"loss": 0.1474,
|
| 1276 |
-
"step": 1810
|
| 1277 |
-
},
|
| 1278 |
-
{
|
| 1279 |
-
"epoch": 2.27,
|
| 1280 |
-
"grad_norm": 2.5528279574109884,
|
| 1281 |
-
"learning_rate": 2.7363681840920464e-05,
|
| 1282 |
-
"loss": 0.1393,
|
| 1283 |
-
"step": 1820
|
| 1284 |
-
},
|
| 1285 |
-
{
|
| 1286 |
-
"epoch": 2.29,
|
| 1287 |
-
"grad_norm": 2.882844754170427,
|
| 1288 |
-
"learning_rate": 2.7238619309654828e-05,
|
| 1289 |
-
"loss": 0.1271,
|
| 1290 |
-
"step": 1830
|
| 1291 |
-
},
|
| 1292 |
-
{
|
| 1293 |
-
"epoch": 2.3,
|
| 1294 |
-
"grad_norm": 3.1442381159468424,
|
| 1295 |
-
"learning_rate": 2.7113556778389198e-05,
|
| 1296 |
-
"loss": 0.1401,
|
| 1297 |
-
"step": 1840
|
| 1298 |
-
},
|
| 1299 |
-
{
|
| 1300 |
-
"epoch": 2.31,
|
| 1301 |
-
"grad_norm": 2.3171759061722663,
|
| 1302 |
-
"learning_rate": 2.6988494247123564e-05,
|
| 1303 |
-
"loss": 0.1363,
|
| 1304 |
-
"step": 1850
|
| 1305 |
-
},
|
| 1306 |
-
{
|
| 1307 |
-
"epoch": 2.33,
|
| 1308 |
-
"grad_norm": 2.5545785863215533,
|
| 1309 |
-
"learning_rate": 2.6863431715857928e-05,
|
| 1310 |
-
"loss": 0.1346,
|
| 1311 |
-
"step": 1860
|
| 1312 |
-
},
|
| 1313 |
-
{
|
| 1314 |
-
"epoch": 2.34,
|
| 1315 |
-
"grad_norm": 2.6824086777926412,
|
| 1316 |
-
"learning_rate": 2.6738369184592298e-05,
|
| 1317 |
-
"loss": 0.1341,
|
| 1318 |
-
"step": 1870
|
| 1319 |
-
},
|
| 1320 |
-
{
|
| 1321 |
-
"epoch": 2.35,
|
| 1322 |
-
"grad_norm": 2.935390607845893,
|
| 1323 |
-
"learning_rate": 2.6613306653326664e-05,
|
| 1324 |
-
"loss": 0.1365,
|
| 1325 |
-
"step": 1880
|
| 1326 |
-
},
|
| 1327 |
-
{
|
| 1328 |
-
"epoch": 2.36,
|
| 1329 |
-
"grad_norm": 2.6179278337223364,
|
| 1330 |
-
"learning_rate": 2.6488244122061034e-05,
|
| 1331 |
-
"loss": 0.1302,
|
| 1332 |
-
"step": 1890
|
| 1333 |
-
},
|
| 1334 |
-
{
|
| 1335 |
-
"epoch": 2.38,
|
| 1336 |
-
"grad_norm": 2.5835120129063265,
|
| 1337 |
-
"learning_rate": 2.63631815907954e-05,
|
| 1338 |
-
"loss": 0.1437,
|
| 1339 |
-
"step": 1900
|
| 1340 |
-
},
|
| 1341 |
-
{
|
| 1342 |
-
"epoch": 2.39,
|
| 1343 |
-
"grad_norm": 2.406526227676954,
|
| 1344 |
-
"learning_rate": 2.6238119059529764e-05,
|
| 1345 |
-
"loss": 0.1329,
|
| 1346 |
-
"step": 1910
|
| 1347 |
-
},
|
| 1348 |
-
{
|
| 1349 |
-
"epoch": 2.4,
|
| 1350 |
-
"grad_norm": 2.516912379701269,
|
| 1351 |
-
"learning_rate": 2.6113056528264134e-05,
|
| 1352 |
-
"loss": 0.1307,
|
| 1353 |
-
"step": 1920
|
| 1354 |
-
},
|
| 1355 |
-
{
|
| 1356 |
-
"epoch": 2.41,
|
| 1357 |
-
"grad_norm": 2.4927236524388294,
|
| 1358 |
-
"learning_rate": 2.59879939969985e-05,
|
| 1359 |
-
"loss": 0.1345,
|
| 1360 |
-
"step": 1930
|
| 1361 |
-
},
|
| 1362 |
-
{
|
| 1363 |
-
"epoch": 2.42,
|
| 1364 |
-
"grad_norm": 2.6724439762477457,
|
| 1365 |
-
"learning_rate": 2.586293146573287e-05,
|
| 1366 |
-
"loss": 0.134,
|
| 1367 |
-
"step": 1940
|
| 1368 |
-
},
|
| 1369 |
-
{
|
| 1370 |
-
"epoch": 2.44,
|
| 1371 |
-
"grad_norm": 2.637033763011149,
|
| 1372 |
-
"learning_rate": 2.5737868934467234e-05,
|
| 1373 |
-
"loss": 0.1287,
|
| 1374 |
-
"step": 1950
|
| 1375 |
-
},
|
| 1376 |
-
{
|
| 1377 |
-
"epoch": 2.45,
|
| 1378 |
-
"grad_norm": 2.7214532326123377,
|
| 1379 |
-
"learning_rate": 2.56128064032016e-05,
|
| 1380 |
-
"loss": 0.157,
|
| 1381 |
-
"step": 1960
|
| 1382 |
-
},
|
| 1383 |
-
{
|
| 1384 |
-
"epoch": 2.46,
|
| 1385 |
-
"grad_norm": 2.3869393054803205,
|
| 1386 |
-
"learning_rate": 2.548774387193597e-05,
|
| 1387 |
-
"loss": 0.124,
|
| 1388 |
-
"step": 1970
|
| 1389 |
-
},
|
| 1390 |
-
{
|
| 1391 |
-
"epoch": 2.48,
|
| 1392 |
-
"grad_norm": 3.097006730897294,
|
| 1393 |
-
"learning_rate": 2.5362681340670334e-05,
|
| 1394 |
-
"loss": 0.1403,
|
| 1395 |
-
"step": 1980
|
| 1396 |
-
},
|
| 1397 |
-
{
|
| 1398 |
-
"epoch": 2.49,
|
| 1399 |
-
"grad_norm": 2.703830428494632,
|
| 1400 |
-
"learning_rate": 2.5237618809404708e-05,
|
| 1401 |
-
"loss": 0.1422,
|
| 1402 |
-
"step": 1990
|
| 1403 |
-
},
|
| 1404 |
-
{
|
| 1405 |
-
"epoch": 2.5,
|
| 1406 |
-
"grad_norm": 2.6503565314509916,
|
| 1407 |
-
"learning_rate": 2.511255627813907e-05,
|
| 1408 |
-
"loss": 0.1326,
|
| 1409 |
-
"step": 2000
|
| 1410 |
-
},
|
| 1411 |
-
{
|
| 1412 |
-
"epoch": 2.51,
|
| 1413 |
-
"grad_norm": 2.8501605206174148,
|
| 1414 |
-
"learning_rate": 2.4987493746873438e-05,
|
| 1415 |
-
"loss": 0.1295,
|
| 1416 |
-
"step": 2010
|
| 1417 |
-
},
|
| 1418 |
-
{
|
| 1419 |
-
"epoch": 2.52,
|
| 1420 |
-
"grad_norm": 2.7904621961035954,
|
| 1421 |
-
"learning_rate": 2.4862431215607804e-05,
|
| 1422 |
-
"loss": 0.1314,
|
| 1423 |
-
"step": 2020
|
| 1424 |
-
},
|
| 1425 |
-
{
|
| 1426 |
-
"epoch": 2.54,
|
| 1427 |
-
"grad_norm": 2.555989811347152,
|
| 1428 |
-
"learning_rate": 2.473736868434217e-05,
|
| 1429 |
-
"loss": 0.1346,
|
| 1430 |
-
"step": 2030
|
| 1431 |
-
},
|
| 1432 |
-
{
|
| 1433 |
-
"epoch": 2.55,
|
| 1434 |
-
"grad_norm": 2.5326240483562694,
|
| 1435 |
-
"learning_rate": 2.461230615307654e-05,
|
| 1436 |
-
"loss": 0.134,
|
| 1437 |
-
"step": 2040
|
| 1438 |
-
},
|
| 1439 |
-
{
|
| 1440 |
-
"epoch": 2.56,
|
| 1441 |
-
"grad_norm": 2.878178705134795,
|
| 1442 |
-
"learning_rate": 2.4487243621810908e-05,
|
| 1443 |
-
"loss": 0.1237,
|
| 1444 |
-
"step": 2050
|
| 1445 |
-
},
|
| 1446 |
-
{
|
| 1447 |
-
"epoch": 2.58,
|
| 1448 |
-
"grad_norm": 2.4896292156848974,
|
| 1449 |
-
"learning_rate": 2.4362181090545274e-05,
|
| 1450 |
-
"loss": 0.1374,
|
| 1451 |
-
"step": 2060
|
| 1452 |
-
},
|
| 1453 |
-
{
|
| 1454 |
-
"epoch": 2.59,
|
| 1455 |
-
"grad_norm": 2.4175247992710087,
|
| 1456 |
-
"learning_rate": 2.423711855927964e-05,
|
| 1457 |
-
"loss": 0.1375,
|
| 1458 |
-
"step": 2070
|
| 1459 |
-
},
|
| 1460 |
-
{
|
| 1461 |
-
"epoch": 2.6,
|
| 1462 |
-
"grad_norm": 3.1027213365169675,
|
| 1463 |
-
"learning_rate": 2.4112056028014007e-05,
|
| 1464 |
-
"loss": 0.1392,
|
| 1465 |
-
"step": 2080
|
| 1466 |
-
},
|
| 1467 |
-
{
|
| 1468 |
-
"epoch": 2.61,
|
| 1469 |
-
"grad_norm": 2.6677840295965782,
|
| 1470 |
-
"learning_rate": 2.3986993496748374e-05,
|
| 1471 |
-
"loss": 0.1218,
|
| 1472 |
-
"step": 2090
|
| 1473 |
-
},
|
| 1474 |
-
{
|
| 1475 |
-
"epoch": 2.62,
|
| 1476 |
-
"grad_norm": 2.679408467615217,
|
| 1477 |
-
"learning_rate": 2.3861930965482744e-05,
|
| 1478 |
-
"loss": 0.1304,
|
| 1479 |
-
"step": 2100
|
| 1480 |
-
},
|
| 1481 |
-
{
|
| 1482 |
-
"epoch": 2.64,
|
| 1483 |
-
"grad_norm": 2.7859729569896468,
|
| 1484 |
-
"learning_rate": 2.373686843421711e-05,
|
| 1485 |
-
"loss": 0.1278,
|
| 1486 |
-
"step": 2110
|
| 1487 |
-
},
|
| 1488 |
-
{
|
| 1489 |
-
"epoch": 2.65,
|
| 1490 |
-
"grad_norm": 2.8594418710695764,
|
| 1491 |
-
"learning_rate": 2.3611805902951477e-05,
|
| 1492 |
-
"loss": 0.1297,
|
| 1493 |
-
"step": 2120
|
| 1494 |
-
},
|
| 1495 |
-
{
|
| 1496 |
-
"epoch": 2.66,
|
| 1497 |
-
"grad_norm": 2.684541004807738,
|
| 1498 |
-
"learning_rate": 2.3486743371685844e-05,
|
| 1499 |
-
"loss": 0.1203,
|
| 1500 |
-
"step": 2130
|
| 1501 |
-
},
|
| 1502 |
-
{
|
| 1503 |
-
"epoch": 2.67,
|
| 1504 |
-
"grad_norm": 2.751766677724115,
|
| 1505 |
-
"learning_rate": 2.336168084042021e-05,
|
| 1506 |
-
"loss": 0.118,
|
| 1507 |
-
"step": 2140
|
| 1508 |
-
},
|
| 1509 |
-
{
|
| 1510 |
-
"epoch": 2.69,
|
| 1511 |
-
"grad_norm": 3.110848849081005,
|
| 1512 |
-
"learning_rate": 2.3236618309154577e-05,
|
| 1513 |
-
"loss": 0.132,
|
| 1514 |
-
"step": 2150
|
| 1515 |
-
},
|
| 1516 |
-
{
|
| 1517 |
-
"epoch": 2.7,
|
| 1518 |
-
"grad_norm": 2.6525580385637593,
|
| 1519 |
-
"learning_rate": 2.3111555777888947e-05,
|
| 1520 |
-
"loss": 0.137,
|
| 1521 |
-
"step": 2160
|
| 1522 |
-
},
|
| 1523 |
-
{
|
| 1524 |
-
"epoch": 2.71,
|
| 1525 |
-
"grad_norm": 2.593636751575111,
|
| 1526 |
-
"learning_rate": 2.2986493246623314e-05,
|
| 1527 |
-
"loss": 0.1359,
|
| 1528 |
-
"step": 2170
|
| 1529 |
-
},
|
| 1530 |
-
{
|
| 1531 |
-
"epoch": 2.73,
|
| 1532 |
-
"grad_norm": 3.0950612073524977,
|
| 1533 |
-
"learning_rate": 2.2861430715357677e-05,
|
| 1534 |
-
"loss": 0.1264,
|
| 1535 |
-
"step": 2180
|
| 1536 |
-
},
|
| 1537 |
-
{
|
| 1538 |
-
"epoch": 2.74,
|
| 1539 |
-
"grad_norm": 2.661775116005948,
|
| 1540 |
-
"learning_rate": 2.2736368184092047e-05,
|
| 1541 |
-
"loss": 0.1251,
|
| 1542 |
-
"step": 2190
|
| 1543 |
-
},
|
| 1544 |
-
{
|
| 1545 |
-
"epoch": 2.75,
|
| 1546 |
-
"grad_norm": 3.230211019325226,
|
| 1547 |
-
"learning_rate": 2.2611305652826414e-05,
|
| 1548 |
-
"loss": 0.1296,
|
| 1549 |
-
"step": 2200
|
| 1550 |
-
},
|
| 1551 |
-
{
|
| 1552 |
-
"epoch": 2.76,
|
| 1553 |
-
"grad_norm": 3.081544477462097,
|
| 1554 |
-
"learning_rate": 2.248624312156078e-05,
|
| 1555 |
-
"loss": 0.1247,
|
| 1556 |
-
"step": 2210
|
| 1557 |
-
},
|
| 1558 |
-
{
|
| 1559 |
-
"epoch": 2.77,
|
| 1560 |
-
"grad_norm": 3.1293638659096885,
|
| 1561 |
-
"learning_rate": 2.236118059029515e-05,
|
| 1562 |
-
"loss": 0.1303,
|
| 1563 |
-
"step": 2220
|
| 1564 |
-
},
|
| 1565 |
-
{
|
| 1566 |
-
"epoch": 2.79,
|
| 1567 |
-
"grad_norm": 2.5946152863853404,
|
| 1568 |
-
"learning_rate": 2.2236118059029514e-05,
|
| 1569 |
-
"loss": 0.131,
|
| 1570 |
-
"step": 2230
|
| 1571 |
-
},
|
| 1572 |
-
{
|
| 1573 |
-
"epoch": 2.8,
|
| 1574 |
-
"grad_norm": 2.6311170965776722,
|
| 1575 |
-
"learning_rate": 2.2111055527763884e-05,
|
| 1576 |
-
"loss": 0.1427,
|
| 1577 |
-
"step": 2240
|
| 1578 |
-
},
|
| 1579 |
-
{
|
| 1580 |
-
"epoch": 2.81,
|
| 1581 |
-
"grad_norm": 2.6125582163771894,
|
| 1582 |
-
"learning_rate": 2.198599299649825e-05,
|
| 1583 |
-
"loss": 0.1313,
|
| 1584 |
-
"step": 2250
|
| 1585 |
-
},
|
| 1586 |
-
{
|
| 1587 |
-
"epoch": 2.83,
|
| 1588 |
-
"grad_norm": 2.79415838341794,
|
| 1589 |
-
"learning_rate": 2.1860930465232617e-05,
|
| 1590 |
-
"loss": 0.1306,
|
| 1591 |
-
"step": 2260
|
| 1592 |
-
},
|
| 1593 |
-
{
|
| 1594 |
-
"epoch": 2.84,
|
| 1595 |
-
"grad_norm": 2.4025875805185057,
|
| 1596 |
-
"learning_rate": 2.1735867933966987e-05,
|
| 1597 |
-
"loss": 0.127,
|
| 1598 |
-
"step": 2270
|
| 1599 |
-
},
|
| 1600 |
-
{
|
| 1601 |
-
"epoch": 2.85,
|
| 1602 |
-
"grad_norm": 2.6788154272935376,
|
| 1603 |
-
"learning_rate": 2.161080540270135e-05,
|
| 1604 |
-
"loss": 0.1191,
|
| 1605 |
-
"step": 2280
|
| 1606 |
-
},
|
| 1607 |
-
{
|
| 1608 |
-
"epoch": 2.86,
|
| 1609 |
-
"grad_norm": 2.6992815510003854,
|
| 1610 |
-
"learning_rate": 2.1485742871435717e-05,
|
| 1611 |
-
"loss": 0.1256,
|
| 1612 |
-
"step": 2290
|
| 1613 |
-
},
|
| 1614 |
-
{
|
| 1615 |
-
"epoch": 2.88,
|
| 1616 |
-
"grad_norm": 2.6704557341346087,
|
| 1617 |
-
"learning_rate": 2.1360680340170087e-05,
|
| 1618 |
-
"loss": 0.1238,
|
| 1619 |
-
"step": 2300
|
| 1620 |
-
},
|
| 1621 |
-
{
|
| 1622 |
-
"epoch": 2.89,
|
| 1623 |
-
"grad_norm": 2.7264637751590683,
|
| 1624 |
-
"learning_rate": 2.1235617808904454e-05,
|
| 1625 |
-
"loss": 0.1271,
|
| 1626 |
-
"step": 2310
|
| 1627 |
-
},
|
| 1628 |
-
{
|
| 1629 |
-
"epoch": 2.9,
|
| 1630 |
-
"grad_norm": 2.952720170302931,
|
| 1631 |
-
"learning_rate": 2.111055527763882e-05,
|
| 1632 |
-
"loss": 0.1255,
|
| 1633 |
-
"step": 2320
|
| 1634 |
-
},
|
| 1635 |
-
{
|
| 1636 |
-
"epoch": 2.91,
|
| 1637 |
-
"grad_norm": 3.0158536359854273,
|
| 1638 |
-
"learning_rate": 2.0985492746373187e-05,
|
| 1639 |
-
"loss": 0.1291,
|
| 1640 |
-
"step": 2330
|
| 1641 |
-
},
|
| 1642 |
-
{
|
| 1643 |
-
"epoch": 2.92,
|
| 1644 |
-
"grad_norm": 2.3829743877512035,
|
| 1645 |
-
"learning_rate": 2.0860430215107554e-05,
|
| 1646 |
-
"loss": 0.137,
|
| 1647 |
-
"step": 2340
|
| 1648 |
-
},
|
| 1649 |
-
{
|
| 1650 |
-
"epoch": 2.94,
|
| 1651 |
-
"grad_norm": 2.2748284034151083,
|
| 1652 |
-
"learning_rate": 2.073536768384192e-05,
|
| 1653 |
-
"loss": 0.1274,
|
| 1654 |
-
"step": 2350
|
| 1655 |
-
},
|
| 1656 |
-
{
|
| 1657 |
-
"epoch": 2.95,
|
| 1658 |
-
"grad_norm": 2.526854855449805,
|
| 1659 |
-
"learning_rate": 2.061030515257629e-05,
|
| 1660 |
-
"loss": 0.121,
|
| 1661 |
-
"step": 2360
|
| 1662 |
-
},
|
| 1663 |
-
{
|
| 1664 |
-
"epoch": 2.96,
|
| 1665 |
-
"grad_norm": 2.61083769775896,
|
| 1666 |
-
"learning_rate": 2.0485242621310657e-05,
|
| 1667 |
-
"loss": 0.1221,
|
| 1668 |
-
"step": 2370
|
| 1669 |
-
},
|
| 1670 |
-
{
|
| 1671 |
-
"epoch": 2.98,
|
| 1672 |
-
"grad_norm": 2.732571077670084,
|
| 1673 |
-
"learning_rate": 2.0360180090045024e-05,
|
| 1674 |
-
"loss": 0.1274,
|
| 1675 |
-
"step": 2380
|
| 1676 |
-
},
|
| 1677 |
-
{
|
| 1678 |
-
"epoch": 2.99,
|
| 1679 |
-
"grad_norm": 2.438170511993909,
|
| 1680 |
-
"learning_rate": 2.023511755877939e-05,
|
| 1681 |
-
"loss": 0.1161,
|
| 1682 |
-
"step": 2390
|
| 1683 |
-
},
|
| 1684 |
-
{
|
| 1685 |
-
"epoch": 3.0,
|
| 1686 |
-
"grad_norm": 2.150128950643317,
|
| 1687 |
-
"learning_rate": 2.0110055027513757e-05,
|
| 1688 |
-
"loss": 0.1261,
|
| 1689 |
-
"step": 2400
|
| 1690 |
-
},
|
| 1691 |
-
{
|
| 1692 |
-
"epoch": 3.01,
|
| 1693 |
-
"grad_norm": 1.143682327769968,
|
| 1694 |
-
"learning_rate": 1.9984992496248124e-05,
|
| 1695 |
-
"loss": 0.0474,
|
| 1696 |
-
"step": 2410
|
| 1697 |
-
},
|
| 1698 |
-
{
|
| 1699 |
-
"epoch": 3.02,
|
| 1700 |
-
"grad_norm": 1.6652489672492834,
|
| 1701 |
-
"learning_rate": 1.9859929964982494e-05,
|
| 1702 |
-
"loss": 0.0443,
|
| 1703 |
-
"step": 2420
|
| 1704 |
-
},
|
| 1705 |
-
{
|
| 1706 |
-
"epoch": 3.04,
|
| 1707 |
-
"grad_norm": 1.572882072960937,
|
| 1708 |
-
"learning_rate": 1.9734867433716857e-05,
|
| 1709 |
-
"loss": 0.0424,
|
| 1710 |
-
"step": 2430
|
| 1711 |
-
},
|
| 1712 |
-
{
|
| 1713 |
-
"epoch": 3.05,
|
| 1714 |
-
"grad_norm": 1.3741877497591568,
|
| 1715 |
-
"learning_rate": 1.9609804902451227e-05,
|
| 1716 |
-
"loss": 0.0545,
|
| 1717 |
-
"step": 2440
|
| 1718 |
-
},
|
| 1719 |
-
{
|
| 1720 |
-
"epoch": 3.06,
|
| 1721 |
-
"grad_norm": 1.7064284737834505,
|
| 1722 |
-
"learning_rate": 1.9484742371185594e-05,
|
| 1723 |
-
"loss": 0.0539,
|
| 1724 |
-
"step": 2450
|
| 1725 |
-
},
|
| 1726 |
-
{
|
| 1727 |
-
"epoch": 3.08,
|
| 1728 |
-
"grad_norm": 1.6462372473336482,
|
| 1729 |
-
"learning_rate": 1.935967983991996e-05,
|
| 1730 |
-
"loss": 0.0516,
|
| 1731 |
-
"step": 2460
|
| 1732 |
-
},
|
| 1733 |
-
{
|
| 1734 |
-
"epoch": 3.09,
|
| 1735 |
-
"grad_norm": 1.560256568155276,
|
| 1736 |
-
"learning_rate": 1.923461730865433e-05,
|
| 1737 |
-
"loss": 0.0496,
|
| 1738 |
-
"step": 2470
|
| 1739 |
-
},
|
| 1740 |
-
{
|
| 1741 |
-
"epoch": 3.1,
|
| 1742 |
-
"grad_norm": 1.381325865513328,
|
| 1743 |
-
"learning_rate": 1.9109554777388694e-05,
|
| 1744 |
-
"loss": 0.048,
|
| 1745 |
-
"step": 2480
|
| 1746 |
-
},
|
| 1747 |
-
{
|
| 1748 |
-
"epoch": 3.11,
|
| 1749 |
-
"grad_norm": 1.8124841492058452,
|
| 1750 |
-
"learning_rate": 1.898449224612306e-05,
|
| 1751 |
-
"loss": 0.054,
|
| 1752 |
-
"step": 2490
|
| 1753 |
-
},
|
| 1754 |
-
{
|
| 1755 |
-
"epoch": 3.12,
|
| 1756 |
-
"grad_norm": 1.6535986065065578,
|
| 1757 |
-
"learning_rate": 1.885942971485743e-05,
|
| 1758 |
-
"loss": 0.06,
|
| 1759 |
-
"step": 2500
|
| 1760 |
-
},
|
| 1761 |
-
{
|
| 1762 |
-
"epoch": 3.14,
|
| 1763 |
-
"grad_norm": 1.9819166319488526,
|
| 1764 |
-
"learning_rate": 1.8734367183591797e-05,
|
| 1765 |
-
"loss": 0.056,
|
| 1766 |
-
"step": 2510
|
| 1767 |
-
},
|
| 1768 |
-
{
|
| 1769 |
-
"epoch": 3.15,
|
| 1770 |
-
"grad_norm": 2.087812060027925,
|
| 1771 |
-
"learning_rate": 1.8609304652326164e-05,
|
| 1772 |
-
"loss": 0.0558,
|
| 1773 |
-
"step": 2520
|
| 1774 |
-
},
|
| 1775 |
-
{
|
| 1776 |
-
"epoch": 3.16,
|
| 1777 |
-
"grad_norm": 1.4283212272341697,
|
| 1778 |
-
"learning_rate": 1.848424212106053e-05,
|
| 1779 |
-
"loss": 0.0541,
|
| 1780 |
-
"step": 2530
|
| 1781 |
-
},
|
| 1782 |
-
{
|
| 1783 |
-
"epoch": 3.17,
|
| 1784 |
-
"grad_norm": 2.1967893363996667,
|
| 1785 |
-
"learning_rate": 1.8359179589794897e-05,
|
| 1786 |
-
"loss": 0.0586,
|
| 1787 |
-
"step": 2540
|
| 1788 |
-
},
|
| 1789 |
-
{
|
| 1790 |
-
"epoch": 3.19,
|
| 1791 |
-
"grad_norm": 1.6441632581708676,
|
| 1792 |
-
"learning_rate": 1.8234117058529264e-05,
|
| 1793 |
-
"loss": 0.05,
|
| 1794 |
-
"step": 2550
|
| 1795 |
-
},
|
| 1796 |
-
{
|
| 1797 |
-
"epoch": 3.2,
|
| 1798 |
-
"grad_norm": 1.7647257852834424,
|
| 1799 |
-
"learning_rate": 1.8109054527263634e-05,
|
| 1800 |
-
"loss": 0.0578,
|
| 1801 |
-
"step": 2560
|
| 1802 |
-
},
|
| 1803 |
-
{
|
| 1804 |
-
"epoch": 3.21,
|
| 1805 |
-
"grad_norm": 1.9882813699118946,
|
| 1806 |
-
"learning_rate": 1.7983991995998e-05,
|
| 1807 |
-
"loss": 0.0524,
|
| 1808 |
-
"step": 2570
|
| 1809 |
-
},
|
| 1810 |
-
{
|
| 1811 |
-
"epoch": 3.23,
|
| 1812 |
-
"grad_norm": 1.6533759029481423,
|
| 1813 |
-
"learning_rate": 1.7858929464732367e-05,
|
| 1814 |
-
"loss": 0.0481,
|
| 1815 |
-
"step": 2580
|
| 1816 |
-
},
|
| 1817 |
-
{
|
| 1818 |
-
"epoch": 3.24,
|
| 1819 |
-
"grad_norm": 1.2640217174463038,
|
| 1820 |
-
"learning_rate": 1.7733866933466734e-05,
|
| 1821 |
-
"loss": 0.0475,
|
| 1822 |
-
"step": 2590
|
| 1823 |
-
},
|
| 1824 |
-
{
|
| 1825 |
-
"epoch": 3.25,
|
| 1826 |
-
"grad_norm": 1.8551424723405705,
|
| 1827 |
-
"learning_rate": 1.76088044022011e-05,
|
| 1828 |
-
"loss": 0.0522,
|
| 1829 |
-
"step": 2600
|
| 1830 |
-
},
|
| 1831 |
-
{
|
| 1832 |
-
"epoch": 3.26,
|
| 1833 |
-
"grad_norm": 1.7963597761468264,
|
| 1834 |
-
"learning_rate": 1.748374187093547e-05,
|
| 1835 |
-
"loss": 0.0541,
|
| 1836 |
-
"step": 2610
|
| 1837 |
-
},
|
| 1838 |
-
{
|
| 1839 |
-
"epoch": 3.27,
|
| 1840 |
-
"grad_norm": 1.7321785430153591,
|
| 1841 |
-
"learning_rate": 1.7358679339669837e-05,
|
| 1842 |
-
"loss": 0.0507,
|
| 1843 |
-
"step": 2620
|
| 1844 |
-
},
|
| 1845 |
-
{
|
| 1846 |
-
"epoch": 3.29,
|
| 1847 |
-
"grad_norm": 1.955876601783095,
|
| 1848 |
-
"learning_rate": 1.72336168084042e-05,
|
| 1849 |
-
"loss": 0.0485,
|
| 1850 |
-
"step": 2630
|
| 1851 |
-
},
|
| 1852 |
-
{
|
| 1853 |
-
"epoch": 3.3,
|
| 1854 |
-
"grad_norm": 1.8485106736675452,
|
| 1855 |
-
"learning_rate": 1.710855427713857e-05,
|
| 1856 |
-
"loss": 0.0532,
|
| 1857 |
-
"step": 2640
|
| 1858 |
-
},
|
| 1859 |
-
{
|
| 1860 |
-
"epoch": 3.31,
|
| 1861 |
-
"grad_norm": 1.5996584855623148,
|
| 1862 |
-
"learning_rate": 1.6983491745872937e-05,
|
| 1863 |
-
"loss": 0.0577,
|
| 1864 |
-
"step": 2650
|
| 1865 |
-
},
|
| 1866 |
-
{
|
| 1867 |
-
"epoch": 3.33,
|
| 1868 |
-
"grad_norm": 1.5254709610927397,
|
| 1869 |
-
"learning_rate": 1.6858429214607304e-05,
|
| 1870 |
-
"loss": 0.0456,
|
| 1871 |
-
"step": 2660
|
| 1872 |
-
},
|
| 1873 |
-
{
|
| 1874 |
-
"epoch": 3.34,
|
| 1875 |
-
"grad_norm": 1.4432293700437517,
|
| 1876 |
-
"learning_rate": 1.6733366683341674e-05,
|
| 1877 |
-
"loss": 0.0546,
|
| 1878 |
-
"step": 2670
|
| 1879 |
-
},
|
| 1880 |
-
{
|
| 1881 |
-
"epoch": 3.35,
|
| 1882 |
-
"grad_norm": 1.6784150027520048,
|
| 1883 |
-
"learning_rate": 1.660830415207604e-05,
|
| 1884 |
-
"loss": 0.0535,
|
| 1885 |
-
"step": 2680
|
| 1886 |
-
},
|
| 1887 |
-
{
|
| 1888 |
-
"epoch": 3.36,
|
| 1889 |
-
"grad_norm": 1.7809353600733642,
|
| 1890 |
-
"learning_rate": 1.6483241620810404e-05,
|
| 1891 |
-
"loss": 0.0451,
|
| 1892 |
-
"step": 2690
|
| 1893 |
-
},
|
| 1894 |
-
{
|
| 1895 |
-
"epoch": 3.38,
|
| 1896 |
-
"grad_norm": 2.581167749013489,
|
| 1897 |
-
"learning_rate": 1.6358179089544774e-05,
|
| 1898 |
-
"loss": 0.0547,
|
| 1899 |
-
"step": 2700
|
| 1900 |
-
},
|
| 1901 |
-
{
|
| 1902 |
-
"epoch": 3.39,
|
| 1903 |
-
"grad_norm": 2.0047117997419015,
|
| 1904 |
-
"learning_rate": 1.623311655827914e-05,
|
| 1905 |
-
"loss": 0.0545,
|
| 1906 |
-
"step": 2710
|
| 1907 |
-
},
|
| 1908 |
-
{
|
| 1909 |
-
"epoch": 3.4,
|
| 1910 |
-
"grad_norm": 1.4654369729791492,
|
| 1911 |
-
"learning_rate": 1.6108054027013507e-05,
|
| 1912 |
-
"loss": 0.048,
|
| 1913 |
-
"step": 2720
|
| 1914 |
-
},
|
| 1915 |
-
{
|
| 1916 |
-
"epoch": 3.41,
|
| 1917 |
-
"grad_norm": 1.878234552879185,
|
| 1918 |
-
"learning_rate": 1.5982991495747877e-05,
|
| 1919 |
-
"loss": 0.0508,
|
| 1920 |
-
"step": 2730
|
| 1921 |
-
},
|
| 1922 |
-
{
|
| 1923 |
-
"epoch": 3.42,
|
| 1924 |
-
"grad_norm": 1.4766487217509554,
|
| 1925 |
-
"learning_rate": 1.585792896448224e-05,
|
| 1926 |
-
"loss": 0.048,
|
| 1927 |
-
"step": 2740
|
| 1928 |
-
},
|
| 1929 |
-
{
|
| 1930 |
-
"epoch": 3.44,
|
| 1931 |
-
"grad_norm": 2.015925067182647,
|
| 1932 |
-
"learning_rate": 1.573286643321661e-05,
|
| 1933 |
-
"loss": 0.0535,
|
| 1934 |
-
"step": 2750
|
| 1935 |
-
},
|
| 1936 |
-
{
|
| 1937 |
-
"epoch": 3.45,
|
| 1938 |
-
"grad_norm": 1.2839739616582218,
|
| 1939 |
-
"learning_rate": 1.5607803901950977e-05,
|
| 1940 |
-
"loss": 0.0511,
|
| 1941 |
-
"step": 2760
|
| 1942 |
-
},
|
| 1943 |
-
{
|
| 1944 |
-
"epoch": 3.46,
|
| 1945 |
-
"grad_norm": 1.2605091355298021,
|
| 1946 |
-
"learning_rate": 1.5482741370685344e-05,
|
| 1947 |
-
"loss": 0.0492,
|
| 1948 |
-
"step": 2770
|
| 1949 |
-
},
|
| 1950 |
-
{
|
| 1951 |
-
"epoch": 3.48,
|
| 1952 |
-
"grad_norm": 1.778839488026904,
|
| 1953 |
-
"learning_rate": 1.535767883941971e-05,
|
| 1954 |
-
"loss": 0.0508,
|
| 1955 |
-
"step": 2780
|
| 1956 |
-
},
|
| 1957 |
-
{
|
| 1958 |
-
"epoch": 3.49,
|
| 1959 |
-
"grad_norm": 1.873029054431649,
|
| 1960 |
-
"learning_rate": 1.5232616308154077e-05,
|
| 1961 |
-
"loss": 0.0515,
|
| 1962 |
-
"step": 2790
|
| 1963 |
-
},
|
| 1964 |
-
{
|
| 1965 |
-
"epoch": 3.5,
|
| 1966 |
-
"grad_norm": 1.489240202252438,
|
| 1967 |
-
"learning_rate": 1.5107553776888445e-05,
|
| 1968 |
-
"loss": 0.0469,
|
| 1969 |
-
"step": 2800
|
| 1970 |
-
},
|
| 1971 |
-
{
|
| 1972 |
-
"epoch": 3.51,
|
| 1973 |
-
"grad_norm": 1.1951564612172105,
|
| 1974 |
-
"learning_rate": 1.4982491245622812e-05,
|
| 1975 |
-
"loss": 0.0424,
|
| 1976 |
-
"step": 2810
|
| 1977 |
-
},
|
| 1978 |
-
{
|
| 1979 |
-
"epoch": 3.52,
|
| 1980 |
-
"grad_norm": 1.427494221957884,
|
| 1981 |
-
"learning_rate": 1.485742871435718e-05,
|
| 1982 |
-
"loss": 0.0452,
|
| 1983 |
-
"step": 2820
|
| 1984 |
-
},
|
| 1985 |
-
{
|
| 1986 |
-
"epoch": 3.54,
|
| 1987 |
-
"grad_norm": 1.767643003168799,
|
| 1988 |
-
"learning_rate": 1.4732366183091547e-05,
|
| 1989 |
-
"loss": 0.054,
|
| 1990 |
-
"step": 2830
|
| 1991 |
-
},
|
| 1992 |
-
{
|
| 1993 |
-
"epoch": 3.55,
|
| 1994 |
-
"grad_norm": 1.8461197963044351,
|
| 1995 |
-
"learning_rate": 1.4607303651825912e-05,
|
| 1996 |
-
"loss": 0.05,
|
| 1997 |
-
"step": 2840
|
| 1998 |
-
},
|
| 1999 |
-
{
|
| 2000 |
-
"epoch": 3.56,
|
| 2001 |
-
"grad_norm": 1.6976823547575175,
|
| 2002 |
-
"learning_rate": 1.448224112056028e-05,
|
| 2003 |
-
"loss": 0.0484,
|
| 2004 |
-
"step": 2850
|
| 2005 |
-
},
|
| 2006 |
-
{
|
| 2007 |
-
"epoch": 3.58,
|
| 2008 |
-
"grad_norm": 2.1165349313587103,
|
| 2009 |
-
"learning_rate": 1.4357178589294649e-05,
|
| 2010 |
-
"loss": 0.0495,
|
| 2011 |
-
"step": 2860
|
| 2012 |
-
},
|
| 2013 |
-
{
|
| 2014 |
-
"epoch": 3.59,
|
| 2015 |
-
"grad_norm": 1.4367595921945118,
|
| 2016 |
-
"learning_rate": 1.4232116058029015e-05,
|
| 2017 |
-
"loss": 0.0449,
|
| 2018 |
-
"step": 2870
|
| 2019 |
-
},
|
| 2020 |
-
{
|
| 2021 |
-
"epoch": 3.6,
|
| 2022 |
-
"grad_norm": 1.5701725835763933,
|
| 2023 |
-
"learning_rate": 1.4107053526763384e-05,
|
| 2024 |
-
"loss": 0.0535,
|
| 2025 |
-
"step": 2880
|
| 2026 |
-
},
|
| 2027 |
-
{
|
| 2028 |
-
"epoch": 3.61,
|
| 2029 |
-
"grad_norm": 1.3890521981336823,
|
| 2030 |
-
"learning_rate": 1.3981990995497749e-05,
|
| 2031 |
-
"loss": 0.0497,
|
| 2032 |
-
"step": 2890
|
| 2033 |
-
},
|
| 2034 |
-
{
|
| 2035 |
-
"epoch": 3.62,
|
| 2036 |
-
"grad_norm": 2.3482814470474187,
|
| 2037 |
-
"learning_rate": 1.3856928464232117e-05,
|
| 2038 |
-
"loss": 0.0479,
|
| 2039 |
-
"step": 2900
|
| 2040 |
-
},
|
| 2041 |
-
{
|
| 2042 |
-
"epoch": 3.64,
|
| 2043 |
-
"grad_norm": 1.818720567602742,
|
| 2044 |
-
"learning_rate": 1.3731865932966484e-05,
|
| 2045 |
-
"loss": 0.0456,
|
| 2046 |
-
"step": 2910
|
| 2047 |
-
},
|
| 2048 |
-
{
|
| 2049 |
-
"epoch": 3.65,
|
| 2050 |
-
"grad_norm": 1.580084990497098,
|
| 2051 |
-
"learning_rate": 1.3606803401700852e-05,
|
| 2052 |
-
"loss": 0.0581,
|
| 2053 |
-
"step": 2920
|
| 2054 |
-
},
|
| 2055 |
-
{
|
| 2056 |
-
"epoch": 3.66,
|
| 2057 |
-
"grad_norm": 1.570405568263889,
|
| 2058 |
-
"learning_rate": 1.348174087043522e-05,
|
| 2059 |
-
"loss": 0.0467,
|
| 2060 |
-
"step": 2930
|
| 2061 |
-
},
|
| 2062 |
-
{
|
| 2063 |
-
"epoch": 3.67,
|
| 2064 |
-
"grad_norm": 1.3862471087385182,
|
| 2065 |
-
"learning_rate": 1.3356678339169585e-05,
|
| 2066 |
-
"loss": 0.0427,
|
| 2067 |
-
"step": 2940
|
| 2068 |
-
},
|
| 2069 |
-
{
|
| 2070 |
-
"epoch": 3.69,
|
| 2071 |
-
"grad_norm": 1.905843535066407,
|
| 2072 |
-
"learning_rate": 1.3231615807903952e-05,
|
| 2073 |
-
"loss": 0.0451,
|
| 2074 |
-
"step": 2950
|
| 2075 |
-
},
|
| 2076 |
-
{
|
| 2077 |
-
"epoch": 3.7,
|
| 2078 |
-
"grad_norm": 2.0096560078524535,
|
| 2079 |
-
"learning_rate": 1.310655327663832e-05,
|
| 2080 |
-
"loss": 0.0478,
|
| 2081 |
-
"step": 2960
|
| 2082 |
-
},
|
| 2083 |
-
{
|
| 2084 |
-
"epoch": 3.71,
|
| 2085 |
-
"grad_norm": 1.8387148364015753,
|
| 2086 |
-
"learning_rate": 1.2981490745372687e-05,
|
| 2087 |
-
"loss": 0.0468,
|
| 2088 |
-
"step": 2970
|
| 2089 |
-
},
|
| 2090 |
-
{
|
| 2091 |
-
"epoch": 3.73,
|
| 2092 |
-
"grad_norm": 1.3511210696441773,
|
| 2093 |
-
"learning_rate": 1.2856428214107055e-05,
|
| 2094 |
-
"loss": 0.0433,
|
| 2095 |
-
"step": 2980
|
| 2096 |
-
},
|
| 2097 |
-
{
|
| 2098 |
-
"epoch": 3.74,
|
| 2099 |
-
"grad_norm": 1.343344871307373,
|
| 2100 |
-
"learning_rate": 1.273136568284142e-05,
|
| 2101 |
-
"loss": 0.0491,
|
| 2102 |
-
"step": 2990
|
| 2103 |
-
},
|
| 2104 |
-
{
|
| 2105 |
-
"epoch": 3.75,
|
| 2106 |
-
"grad_norm": 1.6097943398521046,
|
| 2107 |
-
"learning_rate": 1.2606303151575788e-05,
|
| 2108 |
-
"loss": 0.0441,
|
| 2109 |
-
"step": 3000
|
| 2110 |
-
},
|
| 2111 |
-
{
|
| 2112 |
-
"epoch": 3.76,
|
| 2113 |
-
"grad_norm": 1.430264215935809,
|
| 2114 |
-
"learning_rate": 1.2481240620310155e-05,
|
| 2115 |
-
"loss": 0.0442,
|
| 2116 |
-
"step": 3010
|
| 2117 |
-
},
|
| 2118 |
-
{
|
| 2119 |
-
"epoch": 3.77,
|
| 2120 |
-
"grad_norm": 1.986829606239943,
|
| 2121 |
-
"learning_rate": 1.2356178089044523e-05,
|
| 2122 |
-
"loss": 0.0438,
|
| 2123 |
-
"step": 3020
|
| 2124 |
-
},
|
| 2125 |
-
{
|
| 2126 |
-
"epoch": 3.79,
|
| 2127 |
-
"grad_norm": 1.541613560698392,
|
| 2128 |
-
"learning_rate": 1.223111555777889e-05,
|
| 2129 |
-
"loss": 0.05,
|
| 2130 |
-
"step": 3030
|
| 2131 |
-
},
|
| 2132 |
-
{
|
| 2133 |
-
"epoch": 3.8,
|
| 2134 |
-
"grad_norm": 1.4820071156597356,
|
| 2135 |
-
"learning_rate": 1.2106053026513257e-05,
|
| 2136 |
-
"loss": 0.0442,
|
| 2137 |
-
"step": 3040
|
| 2138 |
-
},
|
| 2139 |
-
{
|
| 2140 |
-
"epoch": 3.81,
|
| 2141 |
-
"grad_norm": 1.9180888276280081,
|
| 2142 |
-
"learning_rate": 1.1980990495247623e-05,
|
| 2143 |
-
"loss": 0.0453,
|
| 2144 |
-
"step": 3050
|
| 2145 |
-
},
|
| 2146 |
-
{
|
| 2147 |
-
"epoch": 3.83,
|
| 2148 |
-
"grad_norm": 1.5819512241044464,
|
| 2149 |
-
"learning_rate": 1.1855927963981992e-05,
|
| 2150 |
-
"loss": 0.0444,
|
| 2151 |
-
"step": 3060
|
| 2152 |
-
},
|
| 2153 |
-
{
|
| 2154 |
-
"epoch": 3.84,
|
| 2155 |
-
"grad_norm": 1.502604448747085,
|
| 2156 |
-
"learning_rate": 1.1730865432716358e-05,
|
| 2157 |
-
"loss": 0.0408,
|
| 2158 |
-
"step": 3070
|
| 2159 |
-
},
|
| 2160 |
-
{
|
| 2161 |
-
"epoch": 3.85,
|
| 2162 |
-
"grad_norm": 1.7278176832888912,
|
| 2163 |
-
"learning_rate": 1.1605802901450725e-05,
|
| 2164 |
-
"loss": 0.0392,
|
| 2165 |
-
"step": 3080
|
| 2166 |
-
},
|
| 2167 |
-
{
|
| 2168 |
-
"epoch": 3.86,
|
| 2169 |
-
"grad_norm": 1.3651777045585691,
|
| 2170 |
-
"learning_rate": 1.1480740370185093e-05,
|
| 2171 |
-
"loss": 0.0388,
|
| 2172 |
-
"step": 3090
|
| 2173 |
-
},
|
| 2174 |
-
{
|
| 2175 |
-
"epoch": 3.88,
|
| 2176 |
-
"grad_norm": 1.7185388608772323,
|
| 2177 |
-
"learning_rate": 1.1355677838919462e-05,
|
| 2178 |
-
"loss": 0.0432,
|
| 2179 |
-
"step": 3100
|
| 2180 |
-
},
|
| 2181 |
-
{
|
| 2182 |
-
"epoch": 3.89,
|
| 2183 |
-
"grad_norm": 1.5103350474787707,
|
| 2184 |
-
"learning_rate": 1.1230615307653827e-05,
|
| 2185 |
-
"loss": 0.0441,
|
| 2186 |
-
"step": 3110
|
| 2187 |
-
},
|
| 2188 |
-
{
|
| 2189 |
-
"epoch": 3.9,
|
| 2190 |
-
"grad_norm": 1.4597562206231804,
|
| 2191 |
-
"learning_rate": 1.1105552776388195e-05,
|
| 2192 |
-
"loss": 0.0475,
|
| 2193 |
-
"step": 3120
|
| 2194 |
-
},
|
| 2195 |
-
{
|
| 2196 |
-
"epoch": 3.91,
|
| 2197 |
-
"grad_norm": 1.3774614843117499,
|
| 2198 |
-
"learning_rate": 1.0980490245122562e-05,
|
| 2199 |
-
"loss": 0.0398,
|
| 2200 |
-
"step": 3130
|
| 2201 |
-
},
|
| 2202 |
-
{
|
| 2203 |
-
"epoch": 3.92,
|
| 2204 |
-
"grad_norm": 1.606767412261371,
|
| 2205 |
-
"learning_rate": 1.0855427713856928e-05,
|
| 2206 |
-
"loss": 0.0471,
|
| 2207 |
-
"step": 3140
|
| 2208 |
-
},
|
| 2209 |
-
{
|
| 2210 |
-
"epoch": 3.94,
|
| 2211 |
-
"grad_norm": 1.3861408157622845,
|
| 2212 |
-
"learning_rate": 1.0730365182591297e-05,
|
| 2213 |
-
"loss": 0.0443,
|
| 2214 |
-
"step": 3150
|
| 2215 |
-
},
|
| 2216 |
-
{
|
| 2217 |
-
"epoch": 3.95,
|
| 2218 |
-
"grad_norm": 1.5538838672128699,
|
| 2219 |
-
"learning_rate": 1.0605302651325663e-05,
|
| 2220 |
-
"loss": 0.0472,
|
| 2221 |
-
"step": 3160
|
| 2222 |
-
},
|
| 2223 |
-
{
|
| 2224 |
-
"epoch": 3.96,
|
| 2225 |
-
"grad_norm": 1.5946894102329936,
|
| 2226 |
-
"learning_rate": 1.0480240120060032e-05,
|
| 2227 |
-
"loss": 0.0433,
|
| 2228 |
-
"step": 3170
|
| 2229 |
-
},
|
| 2230 |
-
{
|
| 2231 |
-
"epoch": 3.98,
|
| 2232 |
-
"grad_norm": 1.2468458913211395,
|
| 2233 |
-
"learning_rate": 1.0355177588794397e-05,
|
| 2234 |
-
"loss": 0.0383,
|
| 2235 |
-
"step": 3180
|
| 2236 |
-
},
|
| 2237 |
-
{
|
| 2238 |
-
"epoch": 3.99,
|
| 2239 |
-
"grad_norm": 1.5664978761582478,
|
| 2240 |
-
"learning_rate": 1.0230115057528765e-05,
|
| 2241 |
-
"loss": 0.0459,
|
| 2242 |
-
"step": 3190
|
| 2243 |
-
},
|
| 2244 |
-
{
|
| 2245 |
-
"epoch": 4.0,
|
| 2246 |
-
"grad_norm": 1.809011654121981,
|
| 2247 |
-
"learning_rate": 1.0105052526263133e-05,
|
| 2248 |
-
"loss": 0.0419,
|
| 2249 |
-
"step": 3200
|
| 2250 |
-
},
|
| 2251 |
-
{
|
| 2252 |
-
"epoch": 4.01,
|
| 2253 |
-
"grad_norm": 1.0160616596239802,
|
| 2254 |
-
"learning_rate": 9.979989994997498e-06,
|
| 2255 |
-
"loss": 0.0166,
|
| 2256 |
-
"step": 3210
|
| 2257 |
-
},
|
| 2258 |
-
{
|
| 2259 |
-
"epoch": 4.03,
|
| 2260 |
-
"grad_norm": 0.9631860678272716,
|
| 2261 |
-
"learning_rate": 9.854927463731867e-06,
|
| 2262 |
-
"loss": 0.0147,
|
| 2263 |
-
"step": 3220
|
| 2264 |
-
},
|
| 2265 |
-
{
|
| 2266 |
-
"epoch": 4.04,
|
| 2267 |
-
"grad_norm": 1.265768914223235,
|
| 2268 |
-
"learning_rate": 9.729864932466233e-06,
|
| 2269 |
-
"loss": 0.0157,
|
| 2270 |
-
"step": 3230
|
| 2271 |
-
},
|
| 2272 |
-
{
|
| 2273 |
-
"epoch": 4.05,
|
| 2274 |
-
"grad_norm": 1.110439058880837,
|
| 2275 |
-
"learning_rate": 9.604802401200602e-06,
|
| 2276 |
-
"loss": 0.0188,
|
| 2277 |
-
"step": 3240
|
| 2278 |
-
},
|
| 2279 |
-
{
|
| 2280 |
-
"epoch": 4.06,
|
| 2281 |
-
"grad_norm": 0.7057898144627774,
|
| 2282 |
-
"learning_rate": 9.479739869934968e-06,
|
| 2283 |
-
"loss": 0.0133,
|
| 2284 |
-
"step": 3250
|
| 2285 |
-
},
|
| 2286 |
-
{
|
| 2287 |
-
"epoch": 4.08,
|
| 2288 |
-
"grad_norm": 0.8140341140571917,
|
| 2289 |
-
"learning_rate": 9.354677338669335e-06,
|
| 2290 |
-
"loss": 0.0139,
|
| 2291 |
-
"step": 3260
|
| 2292 |
-
},
|
| 2293 |
-
{
|
| 2294 |
-
"epoch": 4.09,
|
| 2295 |
-
"grad_norm": 0.9005386330187762,
|
| 2296 |
-
"learning_rate": 9.229614807403703e-06,
|
| 2297 |
-
"loss": 0.0178,
|
| 2298 |
-
"step": 3270
|
| 2299 |
-
},
|
| 2300 |
-
{
|
| 2301 |
-
"epoch": 4.1,
|
| 2302 |
-
"grad_norm": 0.6990527589994958,
|
| 2303 |
-
"learning_rate": 9.104552276138068e-06,
|
| 2304 |
-
"loss": 0.0166,
|
| 2305 |
-
"step": 3280
|
| 2306 |
-
},
|
| 2307 |
-
{
|
| 2308 |
-
"epoch": 4.11,
|
| 2309 |
-
"grad_norm": 0.8374763300383368,
|
| 2310 |
-
"learning_rate": 8.979489744872437e-06,
|
| 2311 |
-
"loss": 0.0126,
|
| 2312 |
-
"step": 3290
|
| 2313 |
-
},
|
| 2314 |
-
{
|
| 2315 |
-
"epoch": 4.12,
|
| 2316 |
-
"grad_norm": 0.6026036059465715,
|
| 2317 |
-
"learning_rate": 8.854427213606805e-06,
|
| 2318 |
-
"loss": 0.0194,
|
| 2319 |
-
"step": 3300
|
| 2320 |
-
},
|
| 2321 |
-
{
|
| 2322 |
-
"epoch": 4.14,
|
| 2323 |
-
"grad_norm": 0.6384568858739467,
|
| 2324 |
-
"learning_rate": 8.72936468234117e-06,
|
| 2325 |
-
"loss": 0.0185,
|
| 2326 |
-
"step": 3310
|
| 2327 |
-
},
|
| 2328 |
-
{
|
| 2329 |
-
"epoch": 4.15,
|
| 2330 |
-
"grad_norm": 0.7979584602452928,
|
| 2331 |
-
"learning_rate": 8.604302151075538e-06,
|
| 2332 |
-
"loss": 0.0127,
|
| 2333 |
-
"step": 3320
|
| 2334 |
-
},
|
| 2335 |
-
{
|
| 2336 |
-
"epoch": 4.16,
|
| 2337 |
-
"grad_norm": 1.0334510433669584,
|
| 2338 |
-
"learning_rate": 8.479239619809905e-06,
|
| 2339 |
-
"loss": 0.0153,
|
| 2340 |
-
"step": 3330
|
| 2341 |
-
},
|
| 2342 |
-
{
|
| 2343 |
-
"epoch": 4.17,
|
| 2344 |
-
"grad_norm": 0.788681126956416,
|
| 2345 |
-
"learning_rate": 8.354177088544273e-06,
|
| 2346 |
-
"loss": 0.0196,
|
| 2347 |
-
"step": 3340
|
| 2348 |
-
},
|
| 2349 |
-
{
|
| 2350 |
-
"epoch": 4.19,
|
| 2351 |
-
"grad_norm": 0.6685423145269257,
|
| 2352 |
-
"learning_rate": 8.22911455727864e-06,
|
| 2353 |
-
"loss": 0.0157,
|
| 2354 |
-
"step": 3350
|
| 2355 |
-
},
|
| 2356 |
-
{
|
| 2357 |
-
"epoch": 4.2,
|
| 2358 |
-
"grad_norm": 0.7287135532877159,
|
| 2359 |
-
"learning_rate": 8.104052026013006e-06,
|
| 2360 |
-
"loss": 0.0161,
|
| 2361 |
-
"step": 3360
|
| 2362 |
-
},
|
| 2363 |
-
{
|
| 2364 |
-
"epoch": 4.21,
|
| 2365 |
-
"grad_norm": 0.7064113955783903,
|
| 2366 |
-
"learning_rate": 7.978989494747375e-06,
|
| 2367 |
-
"loss": 0.0147,
|
| 2368 |
-
"step": 3370
|
| 2369 |
-
},
|
| 2370 |
-
{
|
| 2371 |
-
"epoch": 4.22,
|
| 2372 |
-
"grad_norm": 0.6607611553955022,
|
| 2373 |
-
"learning_rate": 7.85392696348174e-06,
|
| 2374 |
-
"loss": 0.0133,
|
| 2375 |
-
"step": 3380
|
| 2376 |
-
},
|
| 2377 |
-
{
|
| 2378 |
-
"epoch": 4.24,
|
| 2379 |
-
"grad_norm": 1.067040055954057,
|
| 2380 |
-
"learning_rate": 7.728864432216108e-06,
|
| 2381 |
-
"loss": 0.0148,
|
| 2382 |
-
"step": 3390
|
| 2383 |
-
},
|
| 2384 |
-
{
|
| 2385 |
-
"epoch": 4.25,
|
| 2386 |
-
"grad_norm": 1.0324328603459556,
|
| 2387 |
-
"learning_rate": 7.603801900950476e-06,
|
| 2388 |
-
"loss": 0.0142,
|
| 2389 |
-
"step": 3400
|
| 2390 |
-
},
|
| 2391 |
-
{
|
| 2392 |
-
"epoch": 4.26,
|
| 2393 |
-
"grad_norm": 0.9592195055570732,
|
| 2394 |
-
"learning_rate": 7.478739369684842e-06,
|
| 2395 |
-
"loss": 0.016,
|
| 2396 |
-
"step": 3410
|
| 2397 |
-
},
|
| 2398 |
-
{
|
| 2399 |
-
"epoch": 4.28,
|
| 2400 |
-
"grad_norm": 0.9230560560572981,
|
| 2401 |
-
"learning_rate": 7.35367683841921e-06,
|
| 2402 |
-
"loss": 0.0144,
|
| 2403 |
-
"step": 3420
|
| 2404 |
-
},
|
| 2405 |
-
{
|
| 2406 |
-
"epoch": 4.29,
|
| 2407 |
-
"grad_norm": 0.6457139232953377,
|
| 2408 |
-
"learning_rate": 7.228614307153578e-06,
|
| 2409 |
-
"loss": 0.0145,
|
| 2410 |
-
"step": 3430
|
| 2411 |
-
},
|
| 2412 |
-
{
|
| 2413 |
-
"epoch": 4.3,
|
| 2414 |
-
"grad_norm": 0.3527501285452713,
|
| 2415 |
-
"learning_rate": 7.103551775887944e-06,
|
| 2416 |
-
"loss": 0.0144,
|
| 2417 |
-
"step": 3440
|
| 2418 |
-
},
|
| 2419 |
-
{
|
| 2420 |
-
"epoch": 4.31,
|
| 2421 |
-
"grad_norm": 0.6072232084037488,
|
| 2422 |
-
"learning_rate": 6.978489244622312e-06,
|
| 2423 |
-
"loss": 0.0134,
|
| 2424 |
-
"step": 3450
|
| 2425 |
-
},
|
| 2426 |
-
{
|
| 2427 |
-
"epoch": 4.33,
|
| 2428 |
-
"grad_norm": 0.6554608368424355,
|
| 2429 |
-
"learning_rate": 6.853426713356678e-06,
|
| 2430 |
-
"loss": 0.0136,
|
| 2431 |
-
"step": 3460
|
| 2432 |
-
},
|
| 2433 |
-
{
|
| 2434 |
-
"epoch": 4.34,
|
| 2435 |
-
"grad_norm": 0.7345293166981702,
|
| 2436 |
-
"learning_rate": 6.7283641820910456e-06,
|
| 2437 |
-
"loss": 0.017,
|
| 2438 |
-
"step": 3470
|
| 2439 |
-
},
|
| 2440 |
-
{
|
| 2441 |
-
"epoch": 4.35,
|
| 2442 |
-
"grad_norm": 0.6630110101404149,
|
| 2443 |
-
"learning_rate": 6.603301650825414e-06,
|
| 2444 |
-
"loss": 0.0138,
|
| 2445 |
-
"step": 3480
|
| 2446 |
-
},
|
| 2447 |
-
{
|
| 2448 |
-
"epoch": 4.36,
|
| 2449 |
-
"grad_norm": 0.6014101157456893,
|
| 2450 |
-
"learning_rate": 6.47823911955978e-06,
|
| 2451 |
-
"loss": 0.0131,
|
| 2452 |
-
"step": 3490
|
| 2453 |
-
},
|
| 2454 |
-
{
|
| 2455 |
-
"epoch": 4.38,
|
| 2456 |
-
"grad_norm": 0.78056530512705,
|
| 2457 |
-
"learning_rate": 6.353176588294148e-06,
|
| 2458 |
-
"loss": 0.0152,
|
| 2459 |
-
"step": 3500
|
| 2460 |
-
},
|
| 2461 |
-
{
|
| 2462 |
-
"epoch": 4.39,
|
| 2463 |
-
"grad_norm": 0.625830313370077,
|
| 2464 |
-
"learning_rate": 6.228114057028515e-06,
|
| 2465 |
-
"loss": 0.0136,
|
| 2466 |
-
"step": 3510
|
| 2467 |
-
},
|
| 2468 |
-
{
|
| 2469 |
-
"epoch": 4.4,
|
| 2470 |
-
"grad_norm": 1.04947606820658,
|
| 2471 |
-
"learning_rate": 6.103051525762881e-06,
|
| 2472 |
-
"loss": 0.0158,
|
| 2473 |
-
"step": 3520
|
| 2474 |
-
},
|
| 2475 |
-
{
|
| 2476 |
-
"epoch": 4.41,
|
| 2477 |
-
"grad_norm": 0.6295492545834148,
|
| 2478 |
-
"learning_rate": 5.977988994497249e-06,
|
| 2479 |
-
"loss": 0.0143,
|
| 2480 |
-
"step": 3530
|
| 2481 |
-
},
|
| 2482 |
-
{
|
| 2483 |
-
"epoch": 4.42,
|
| 2484 |
-
"grad_norm": 0.6227005381610481,
|
| 2485 |
-
"learning_rate": 5.8529264632316155e-06,
|
| 2486 |
-
"loss": 0.0133,
|
| 2487 |
-
"step": 3540
|
| 2488 |
-
},
|
| 2489 |
-
{
|
| 2490 |
-
"epoch": 4.44,
|
| 2491 |
-
"grad_norm": 0.8799433375127763,
|
| 2492 |
-
"learning_rate": 5.727863931965984e-06,
|
| 2493 |
-
"loss": 0.0154,
|
| 2494 |
-
"step": 3550
|
| 2495 |
-
},
|
| 2496 |
-
{
|
| 2497 |
-
"epoch": 4.45,
|
| 2498 |
-
"grad_norm": 0.5503927714060292,
|
| 2499 |
-
"learning_rate": 5.6028014007003505e-06,
|
| 2500 |
-
"loss": 0.0114,
|
| 2501 |
-
"step": 3560
|
| 2502 |
-
},
|
| 2503 |
-
{
|
| 2504 |
-
"epoch": 4.46,
|
| 2505 |
-
"grad_norm": 1.165400459405073,
|
| 2506 |
-
"learning_rate": 5.477738869434718e-06,
|
| 2507 |
-
"loss": 0.0135,
|
| 2508 |
-
"step": 3570
|
| 2509 |
-
},
|
| 2510 |
-
{
|
| 2511 |
-
"epoch": 4.47,
|
| 2512 |
-
"grad_norm": 0.5467153588664332,
|
| 2513 |
-
"learning_rate": 5.352676338169085e-06,
|
| 2514 |
-
"loss": 0.0131,
|
| 2515 |
-
"step": 3580
|
| 2516 |
-
},
|
| 2517 |
-
{
|
| 2518 |
-
"epoch": 4.49,
|
| 2519 |
-
"grad_norm": 0.9263285375289982,
|
| 2520 |
-
"learning_rate": 5.227613806903451e-06,
|
| 2521 |
-
"loss": 0.0128,
|
| 2522 |
-
"step": 3590
|
| 2523 |
-
},
|
| 2524 |
-
{
|
| 2525 |
-
"epoch": 4.5,
|
| 2526 |
-
"grad_norm": 0.8415201179153123,
|
| 2527 |
-
"learning_rate": 5.10255127563782e-06,
|
| 2528 |
-
"loss": 0.0112,
|
| 2529 |
-
"step": 3600
|
| 2530 |
-
},
|
| 2531 |
-
{
|
| 2532 |
-
"epoch": 4.51,
|
| 2533 |
-
"grad_norm": 0.6593144935941556,
|
| 2534 |
-
"learning_rate": 4.977488744372186e-06,
|
| 2535 |
-
"loss": 0.0154,
|
| 2536 |
-
"step": 3610
|
| 2537 |
-
},
|
| 2538 |
-
{
|
| 2539 |
-
"epoch": 4.53,
|
| 2540 |
-
"grad_norm": 0.914570463491157,
|
| 2541 |
-
"learning_rate": 4.852426213106554e-06,
|
| 2542 |
-
"loss": 0.0131,
|
| 2543 |
-
"step": 3620
|
| 2544 |
-
},
|
| 2545 |
-
{
|
| 2546 |
-
"epoch": 4.54,
|
| 2547 |
-
"grad_norm": 0.5307450980608305,
|
| 2548 |
-
"learning_rate": 4.7273636818409205e-06,
|
| 2549 |
-
"loss": 0.0152,
|
| 2550 |
-
"step": 3630
|
| 2551 |
-
},
|
| 2552 |
-
{
|
| 2553 |
-
"epoch": 4.55,
|
| 2554 |
-
"grad_norm": 0.5720744936742393,
|
| 2555 |
-
"learning_rate": 4.602301150575288e-06,
|
| 2556 |
-
"loss": 0.01,
|
| 2557 |
-
"step": 3640
|
| 2558 |
-
},
|
| 2559 |
-
{
|
| 2560 |
-
"epoch": 4.56,
|
| 2561 |
-
"grad_norm": 0.6906646371919622,
|
| 2562 |
-
"learning_rate": 4.4772386193096554e-06,
|
| 2563 |
-
"loss": 0.0158,
|
| 2564 |
-
"step": 3650
|
| 2565 |
-
},
|
| 2566 |
-
{
|
| 2567 |
-
"epoch": 4.58,
|
| 2568 |
-
"grad_norm": 0.750217922657539,
|
| 2569 |
-
"learning_rate": 4.352176088044022e-06,
|
| 2570 |
-
"loss": 0.0137,
|
| 2571 |
-
"step": 3660
|
| 2572 |
-
},
|
| 2573 |
-
{
|
| 2574 |
-
"epoch": 4.59,
|
| 2575 |
-
"grad_norm": 0.6507958418548583,
|
| 2576 |
-
"learning_rate": 4.22711355677839e-06,
|
| 2577 |
-
"loss": 0.0108,
|
| 2578 |
-
"step": 3670
|
| 2579 |
-
},
|
| 2580 |
-
{
|
| 2581 |
-
"epoch": 4.6,
|
| 2582 |
-
"grad_norm": 0.4417488321613792,
|
| 2583 |
-
"learning_rate": 4.102051025512756e-06,
|
| 2584 |
-
"loss": 0.0125,
|
| 2585 |
-
"step": 3680
|
| 2586 |
-
},
|
| 2587 |
-
{
|
| 2588 |
-
"epoch": 4.61,
|
| 2589 |
-
"grad_norm": 0.8625362167494579,
|
| 2590 |
-
"learning_rate": 3.976988494247124e-06,
|
| 2591 |
-
"loss": 0.0157,
|
| 2592 |
-
"step": 3690
|
| 2593 |
-
},
|
| 2594 |
-
{
|
| 2595 |
-
"epoch": 4.62,
|
| 2596 |
-
"grad_norm": 0.7856141458518725,
|
| 2597 |
-
"learning_rate": 3.851925962981491e-06,
|
| 2598 |
-
"loss": 0.0109,
|
| 2599 |
-
"step": 3700
|
| 2600 |
-
},
|
| 2601 |
-
{
|
| 2602 |
-
"epoch": 4.64,
|
| 2603 |
-
"grad_norm": 0.6760897290661452,
|
| 2604 |
-
"learning_rate": 3.726863431715858e-06,
|
| 2605 |
-
"loss": 0.012,
|
| 2606 |
-
"step": 3710
|
| 2607 |
-
},
|
| 2608 |
-
{
|
| 2609 |
-
"epoch": 4.65,
|
| 2610 |
-
"grad_norm": 0.8928787426379452,
|
| 2611 |
-
"learning_rate": 3.601800900450225e-06,
|
| 2612 |
-
"loss": 0.0118,
|
| 2613 |
-
"step": 3720
|
| 2614 |
-
},
|
| 2615 |
-
{
|
| 2616 |
-
"epoch": 4.66,
|
| 2617 |
-
"grad_norm": 0.9518079569519863,
|
| 2618 |
-
"learning_rate": 3.476738369184592e-06,
|
| 2619 |
-
"loss": 0.0136,
|
| 2620 |
-
"step": 3730
|
| 2621 |
-
},
|
| 2622 |
-
{
|
| 2623 |
-
"epoch": 4.67,
|
| 2624 |
-
"grad_norm": 1.1730903426890404,
|
| 2625 |
-
"learning_rate": 3.35167583791896e-06,
|
| 2626 |
-
"loss": 0.0162,
|
| 2627 |
-
"step": 3740
|
| 2628 |
-
},
|
| 2629 |
-
{
|
| 2630 |
-
"epoch": 4.69,
|
| 2631 |
-
"grad_norm": 0.7279487019873496,
|
| 2632 |
-
"learning_rate": 3.226613306653327e-06,
|
| 2633 |
-
"loss": 0.0115,
|
| 2634 |
-
"step": 3750
|
| 2635 |
-
},
|
| 2636 |
-
{
|
| 2637 |
-
"epoch": 4.7,
|
| 2638 |
-
"grad_norm": 0.9401650695234949,
|
| 2639 |
-
"learning_rate": 3.101550775387694e-06,
|
| 2640 |
-
"loss": 0.0096,
|
| 2641 |
-
"step": 3760
|
| 2642 |
-
},
|
| 2643 |
-
{
|
| 2644 |
-
"epoch": 4.71,
|
| 2645 |
-
"grad_norm": 0.6111755424427511,
|
| 2646 |
-
"learning_rate": 2.976488244122061e-06,
|
| 2647 |
-
"loss": 0.0172,
|
| 2648 |
-
"step": 3770
|
| 2649 |
-
},
|
| 2650 |
-
{
|
| 2651 |
-
"epoch": 4.72,
|
| 2652 |
-
"grad_norm": 0.3495626578205311,
|
| 2653 |
-
"learning_rate": 2.8514257128564283e-06,
|
| 2654 |
-
"loss": 0.0123,
|
| 2655 |
-
"step": 3780
|
| 2656 |
-
},
|
| 2657 |
-
{
|
| 2658 |
-
"epoch": 4.74,
|
| 2659 |
-
"grad_norm": 0.4299090247891363,
|
| 2660 |
-
"learning_rate": 2.7263631815907953e-06,
|
| 2661 |
-
"loss": 0.0111,
|
| 2662 |
-
"step": 3790
|
| 2663 |
-
},
|
| 2664 |
-
{
|
| 2665 |
-
"epoch": 4.75,
|
| 2666 |
-
"grad_norm": 0.6609027411517305,
|
| 2667 |
-
"learning_rate": 2.601300650325163e-06,
|
| 2668 |
-
"loss": 0.0143,
|
| 2669 |
-
"step": 3800
|
| 2670 |
-
},
|
| 2671 |
-
{
|
| 2672 |
-
"epoch": 4.76,
|
| 2673 |
-
"grad_norm": 0.6019036947918726,
|
| 2674 |
-
"learning_rate": 2.47623811905953e-06,
|
| 2675 |
-
"loss": 0.0106,
|
| 2676 |
-
"step": 3810
|
| 2677 |
-
},
|
| 2678 |
-
{
|
| 2679 |
-
"epoch": 4.78,
|
| 2680 |
-
"grad_norm": 0.8931005105357686,
|
| 2681 |
-
"learning_rate": 2.351175587793897e-06,
|
| 2682 |
-
"loss": 0.0103,
|
| 2683 |
-
"step": 3820
|
| 2684 |
-
},
|
| 2685 |
-
{
|
| 2686 |
-
"epoch": 4.79,
|
| 2687 |
-
"grad_norm": 0.7662844250959648,
|
| 2688 |
-
"learning_rate": 2.226113056528264e-06,
|
| 2689 |
-
"loss": 0.0112,
|
| 2690 |
-
"step": 3830
|
| 2691 |
-
},
|
| 2692 |
-
{
|
| 2693 |
-
"epoch": 4.8,
|
| 2694 |
-
"grad_norm": 1.0339128043635872,
|
| 2695 |
-
"learning_rate": 2.1010505252626315e-06,
|
| 2696 |
-
"loss": 0.0136,
|
| 2697 |
-
"step": 3840
|
| 2698 |
-
},
|
| 2699 |
-
{
|
| 2700 |
-
"epoch": 4.81,
|
| 2701 |
-
"grad_norm": 0.6241448751379605,
|
| 2702 |
-
"learning_rate": 1.9759879939969986e-06,
|
| 2703 |
-
"loss": 0.0143,
|
| 2704 |
-
"step": 3850
|
| 2705 |
-
},
|
| 2706 |
-
{
|
| 2707 |
-
"epoch": 4.83,
|
| 2708 |
-
"grad_norm": 0.6831824891172988,
|
| 2709 |
-
"learning_rate": 1.8509254627313657e-06,
|
| 2710 |
-
"loss": 0.0121,
|
| 2711 |
-
"step": 3860
|
| 2712 |
-
},
|
| 2713 |
-
{
|
| 2714 |
-
"epoch": 4.84,
|
| 2715 |
-
"grad_norm": 0.7889281432328193,
|
| 2716 |
-
"learning_rate": 1.725862931465733e-06,
|
| 2717 |
-
"loss": 0.0134,
|
| 2718 |
-
"step": 3870
|
| 2719 |
-
},
|
| 2720 |
-
{
|
| 2721 |
-
"epoch": 4.85,
|
| 2722 |
-
"grad_norm": 1.013124998305625,
|
| 2723 |
-
"learning_rate": 1.6008004002001e-06,
|
| 2724 |
-
"loss": 0.0117,
|
| 2725 |
-
"step": 3880
|
| 2726 |
-
},
|
| 2727 |
-
{
|
| 2728 |
-
"epoch": 4.86,
|
| 2729 |
-
"grad_norm": 0.7583032179361607,
|
| 2730 |
-
"learning_rate": 1.4757378689344673e-06,
|
| 2731 |
-
"loss": 0.0117,
|
| 2732 |
-
"step": 3890
|
| 2733 |
-
},
|
| 2734 |
-
{
|
| 2735 |
-
"epoch": 4.88,
|
| 2736 |
-
"grad_norm": 0.7353255835981027,
|
| 2737 |
-
"learning_rate": 1.3506753376688344e-06,
|
| 2738 |
-
"loss": 0.0104,
|
| 2739 |
-
"step": 3900
|
| 2740 |
-
},
|
| 2741 |
-
{
|
| 2742 |
-
"epoch": 4.89,
|
| 2743 |
-
"grad_norm": 0.8789979335860438,
|
| 2744 |
-
"learning_rate": 1.2256128064032017e-06,
|
| 2745 |
-
"loss": 0.0127,
|
| 2746 |
-
"step": 3910
|
| 2747 |
-
},
|
| 2748 |
-
{
|
| 2749 |
-
"epoch": 4.9,
|
| 2750 |
-
"grad_norm": 0.7867461850365051,
|
| 2751 |
-
"learning_rate": 1.1005502751375688e-06,
|
| 2752 |
-
"loss": 0.0108,
|
| 2753 |
-
"step": 3920
|
| 2754 |
-
},
|
| 2755 |
-
{
|
| 2756 |
-
"epoch": 4.91,
|
| 2757 |
-
"grad_norm": 0.6118577850304773,
|
| 2758 |
-
"learning_rate": 9.75487743871936e-07,
|
| 2759 |
-
"loss": 0.0114,
|
| 2760 |
-
"step": 3930
|
| 2761 |
-
},
|
| 2762 |
-
{
|
| 2763 |
-
"epoch": 4.92,
|
| 2764 |
-
"grad_norm": 0.45045792869989826,
|
| 2765 |
-
"learning_rate": 8.504252126063032e-07,
|
| 2766 |
-
"loss": 0.0103,
|
| 2767 |
-
"step": 3940
|
| 2768 |
-
},
|
| 2769 |
-
{
|
| 2770 |
-
"epoch": 4.94,
|
| 2771 |
-
"grad_norm": 0.5684631882075091,
|
| 2772 |
-
"learning_rate": 7.253626813406703e-07,
|
| 2773 |
-
"loss": 0.0129,
|
| 2774 |
-
"step": 3950
|
| 2775 |
-
},
|
| 2776 |
-
{
|
| 2777 |
-
"epoch": 4.95,
|
| 2778 |
-
"grad_norm": 0.7805656487508835,
|
| 2779 |
-
"learning_rate": 6.003001500750376e-07,
|
| 2780 |
-
"loss": 0.0147,
|
| 2781 |
-
"step": 3960
|
| 2782 |
-
},
|
| 2783 |
-
{
|
| 2784 |
-
"epoch": 4.96,
|
| 2785 |
-
"grad_norm": 0.41471654590454743,
|
| 2786 |
-
"learning_rate": 4.7523761880940473e-07,
|
| 2787 |
-
"loss": 0.0138,
|
| 2788 |
-
"step": 3970
|
| 2789 |
-
},
|
| 2790 |
-
{
|
| 2791 |
-
"epoch": 4.97,
|
| 2792 |
-
"grad_norm": 0.5741494584607629,
|
| 2793 |
-
"learning_rate": 3.501750875437719e-07,
|
| 2794 |
-
"loss": 0.0121,
|
| 2795 |
-
"step": 3980
|
| 2796 |
-
},
|
| 2797 |
-
{
|
| 2798 |
-
"epoch": 4.99,
|
| 2799 |
-
"grad_norm": 1.0545689268922978,
|
| 2800 |
-
"learning_rate": 2.2511255627813906e-07,
|
| 2801 |
-
"loss": 0.0102,
|
| 2802 |
-
"step": 3990
|
| 2803 |
-
},
|
| 2804 |
-
{
|
| 2805 |
-
"epoch": 5.0,
|
| 2806 |
-
"grad_norm": 0.5955044024086137,
|
| 2807 |
-
"learning_rate": 1.0005002501250625e-07,
|
| 2808 |
-
"loss": 0.0094,
|
| 2809 |
-
"step": 4000
|
| 2810 |
-
},
|
| 2811 |
-
{
|
| 2812 |
-
"epoch": 5.0,
|
| 2813 |
-
"step": 4000,
|
| 2814 |
-
"total_flos": 18796948488192.0,
|
| 2815 |
-
"train_loss": 0.2349783019721508,
|
| 2816 |
-
"train_runtime": 19900.0608,
|
| 2817 |
-
"train_samples_per_second": 12.859,
|
| 2818 |
-
"train_steps_per_second": 0.201
|
| 2819 |
-
}
|
| 2820 |
-
],
|
| 2821 |
-
"logging_steps": 10,
|
| 2822 |
-
"max_steps": 4000,
|
| 2823 |
-
"num_input_tokens_seen": 0,
|
| 2824 |
-
"num_train_epochs": 5,
|
| 2825 |
-
"save_steps": 50000,
|
| 2826 |
-
"total_flos": 18796948488192.0,
|
| 2827 |
-
"train_batch_size": 8,
|
| 2828 |
-
"trial_name": null,
|
| 2829 |
-
"trial_params": null
|
| 2830 |
-
}
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
training_loss.png
ADDED
|