Commit
•
ad061d0
1
Parent(s):
2fe6b35
End of training
Browse files- README.md +3 -3
- all_results.json +11 -11
- eval_results.json +6 -6
- runs/May24_00-39-21_175d33c654c0/events.out.tfevents.1716522212.175d33c654c0.42.1 +3 -0
- train_results.json +6 -6
- trainer_state.json +2262 -774
README.md
CHANGED
|
@@ -22,7 +22,7 @@ model-index:
|
|
| 22 |
metrics:
|
| 23 |
- name: Accuracy
|
| 24 |
type: accuracy
|
| 25 |
-
value: 0.
|
| 26 |
---
|
| 27 |
|
| 28 |
<!-- This model card has been generated automatically according to the information the Trainer had access to. You
|
|
@@ -32,8 +32,8 @@ should probably proofread and complete it, then remove this comment. -->
|
|
| 32 |
|
| 33 |
This model is a fine-tuned version of [motheecreator/vit-base-patch16-224-in21k-finetuned](https://huggingface.co/motheecreator/vit-base-patch16-224-in21k-finetuned) on the image_folder dataset.
|
| 34 |
It achieves the following results on the evaluation set:
|
| 35 |
-
- Loss: 0.
|
| 36 |
-
- Accuracy: 0.
|
| 37 |
|
| 38 |
## Model description
|
| 39 |
|
|
|
|
| 22 |
metrics:
|
| 23 |
- name: Accuracy
|
| 24 |
type: accuracy
|
| 25 |
+
value: 0.7839129072681704
|
| 26 |
---
|
| 27 |
|
| 28 |
<!-- This model card has been generated automatically according to the information the Trainer had access to. You
|
|
|
|
| 32 |
|
| 33 |
This model is a fine-tuned version of [motheecreator/vit-base-patch16-224-in21k-finetuned](https://huggingface.co/motheecreator/vit-base-patch16-224-in21k-finetuned) on the image_folder dataset.
|
| 34 |
It achieves the following results on the evaluation set:
|
| 35 |
+
- Loss: 0.6282
|
| 36 |
+
- Accuracy: 0.7839
|
| 37 |
|
| 38 |
## Model description
|
| 39 |
|
all_results.json
CHANGED
|
@@ -1,13 +1,13 @@
|
|
| 1 |
{
|
| 2 |
-
"epoch":
|
| 3 |
-
"eval_accuracy": 0.
|
| 4 |
-
"eval_loss": 0.
|
| 5 |
-
"eval_runtime":
|
| 6 |
-
"eval_samples_per_second":
|
| 7 |
-
"eval_steps_per_second": 3.
|
| 8 |
-
"total_flos":
|
| 9 |
-
"train_loss": 0.
|
| 10 |
-
"train_runtime":
|
| 11 |
-
"train_samples_per_second":
|
| 12 |
-
"train_steps_per_second": 0.
|
| 13 |
}
|
|
|
|
| 1 |
{
|
| 2 |
+
"epoch": 6.0,
|
| 3 |
+
"eval_accuracy": 0.7839129072681704,
|
| 4 |
+
"eval_loss": 0.62819904088974,
|
| 5 |
+
"eval_runtime": 233.6957,
|
| 6 |
+
"eval_samples_per_second": 109.27,
|
| 7 |
+
"eval_steps_per_second": 3.415,
|
| 8 |
+
"total_flos": 4.749417900430118e+19,
|
| 9 |
+
"train_loss": 0.321556950521748,
|
| 10 |
+
"train_runtime": 10205.5984,
|
| 11 |
+
"train_samples_per_second": 60.052,
|
| 12 |
+
"train_steps_per_second": 0.469
|
| 13 |
}
|
eval_results.json
CHANGED
|
@@ -1,8 +1,8 @@
|
|
| 1 |
{
|
| 2 |
-
"epoch":
|
| 3 |
-
"eval_accuracy": 0.
|
| 4 |
-
"eval_loss": 0.
|
| 5 |
-
"eval_runtime":
|
| 6 |
-
"eval_samples_per_second":
|
| 7 |
-
"eval_steps_per_second": 3.
|
| 8 |
}
|
|
|
|
| 1 |
{
|
| 2 |
+
"epoch": 6.0,
|
| 3 |
+
"eval_accuracy": 0.7839129072681704,
|
| 4 |
+
"eval_loss": 0.62819904088974,
|
| 5 |
+
"eval_runtime": 233.6957,
|
| 6 |
+
"eval_samples_per_second": 109.27,
|
| 7 |
+
"eval_steps_per_second": 3.415
|
| 8 |
}
|
runs/May24_00-39-21_175d33c654c0/events.out.tfevents.1716522212.175d33c654c0.42.1
ADDED
|
@@ -0,0 +1,3 @@
|
|
|
|
|
|
|
|
|
|
|
|
|
| 1 |
+
version https://git-lfs.github.com/spec/v1
|
| 2 |
+
oid sha256:28360fbd68ae4f82d76955a856b50410d40b864486e405dabf47d335bd0bb6b7
|
| 3 |
+
size 411
|
train_results.json
CHANGED
|
@@ -1,8 +1,8 @@
|
|
| 1 |
{
|
| 2 |
-
"epoch":
|
| 3 |
-
"total_flos":
|
| 4 |
-
"train_loss": 0.
|
| 5 |
-
"train_runtime":
|
| 6 |
-
"train_samples_per_second":
|
| 7 |
-
"train_steps_per_second": 0.
|
| 8 |
}
|
|
|
|
| 1 |
{
|
| 2 |
+
"epoch": 6.0,
|
| 3 |
+
"total_flos": 4.749417900430118e+19,
|
| 4 |
+
"train_loss": 0.321556950521748,
|
| 5 |
+
"train_runtime": 10205.5984,
|
| 6 |
+
"train_samples_per_second": 60.052,
|
| 7 |
+
"train_steps_per_second": 0.469
|
| 8 |
}
|
trainer_state.json
CHANGED
|
@@ -1,1463 +1,2951 @@
|
|
| 1 |
{
|
| 2 |
-
"best_metric": 0.
|
| 3 |
-
"best_model_checkpoint": "vit-base-patch16-224-in21k-finetuned/checkpoint-
|
| 4 |
-
"epoch":
|
| 5 |
"eval_steps": 500,
|
| 6 |
-
"global_step":
|
| 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.
|
| 13 |
-
"learning_rate":
|
| 14 |
-
"loss":
|
| 15 |
"step": 10
|
| 16 |
},
|
| 17 |
{
|
| 18 |
-
"epoch": 0.
|
| 19 |
-
"learning_rate":
|
| 20 |
-
"loss":
|
| 21 |
"step": 20
|
| 22 |
},
|
| 23 |
{
|
| 24 |
-
"epoch": 0.
|
| 25 |
-
"learning_rate":
|
| 26 |
-
"loss":
|
| 27 |
"step": 30
|
| 28 |
},
|
| 29 |
{
|
| 30 |
-
"epoch": 0.
|
| 31 |
-
"learning_rate":
|
| 32 |
-
"loss":
|
| 33 |
"step": 40
|
| 34 |
},
|
| 35 |
{
|
| 36 |
-
"epoch": 0.
|
| 37 |
-
"learning_rate":
|
| 38 |
-
"loss":
|
| 39 |
"step": 50
|
| 40 |
},
|
| 41 |
{
|
| 42 |
-
"epoch": 0.
|
| 43 |
-
"learning_rate":
|
| 44 |
-
"loss":
|
| 45 |
"step": 60
|
| 46 |
},
|
| 47 |
{
|
| 48 |
-
"epoch": 0.
|
| 49 |
-
"learning_rate": 1.
|
| 50 |
-
"loss": 0.
|
| 51 |
"step": 70
|
| 52 |
},
|
| 53 |
{
|
| 54 |
-
"epoch": 0.
|
| 55 |
-
"learning_rate": 1.
|
| 56 |
-
"loss": 0.
|
| 57 |
"step": 80
|
| 58 |
},
|
| 59 |
{
|
| 60 |
-
"epoch": 0.
|
| 61 |
-
"learning_rate":
|
| 62 |
-
"loss": 0.
|
| 63 |
"step": 90
|
| 64 |
},
|
| 65 |
{
|
| 66 |
-
"epoch": 0.
|
| 67 |
-
"learning_rate":
|
| 68 |
-
"loss": 0.
|
| 69 |
"step": 100
|
| 70 |
},
|
| 71 |
{
|
| 72 |
-
"epoch": 0.
|
| 73 |
-
"learning_rate":
|
| 74 |
-
"loss": 0.
|
| 75 |
"step": 110
|
| 76 |
},
|
| 77 |
{
|
| 78 |
-
"epoch": 0.
|
| 79 |
-
"learning_rate":
|
| 80 |
-
"loss": 0.
|
| 81 |
"step": 120
|
| 82 |
},
|
| 83 |
{
|
| 84 |
-
"epoch": 0.
|
| 85 |
-
"learning_rate": 2.
|
| 86 |
-
"loss": 0.
|
| 87 |
"step": 130
|
| 88 |
},
|
| 89 |
{
|
| 90 |
-
"epoch": 0.
|
| 91 |
-
"learning_rate":
|
| 92 |
-
"loss": 0.
|
| 93 |
"step": 140
|
| 94 |
},
|
| 95 |
{
|
| 96 |
-
"epoch": 0.
|
| 97 |
-
"learning_rate":
|
| 98 |
-
"loss": 0.
|
| 99 |
"step": 150
|
| 100 |
},
|
| 101 |
{
|
| 102 |
-
"epoch": 0.
|
| 103 |
-
"learning_rate":
|
| 104 |
-
"loss": 0.
|
| 105 |
"step": 160
|
| 106 |
},
|
| 107 |
{
|
| 108 |
-
"epoch": 0.
|
| 109 |
-
"learning_rate":
|
| 110 |
-
"loss": 0.
|
| 111 |
"step": 170
|
| 112 |
},
|
| 113 |
{
|
| 114 |
-
"epoch": 0.
|
| 115 |
-
"learning_rate":
|
| 116 |
-
"loss": 0.
|
| 117 |
"step": 180
|
| 118 |
},
|
| 119 |
{
|
| 120 |
-
"epoch": 0.
|
| 121 |
-
"learning_rate":
|
| 122 |
-
"loss": 0.
|
| 123 |
"step": 190
|
| 124 |
},
|
| 125 |
{
|
| 126 |
-
"epoch": 0.
|
| 127 |
-
"learning_rate":
|
| 128 |
-
"loss": 0.
|
| 129 |
"step": 200
|
| 130 |
},
|
| 131 |
{
|
| 132 |
-
"epoch": 0.
|
| 133 |
-
"learning_rate":
|
| 134 |
-
"loss": 0.
|
| 135 |
"step": 210
|
| 136 |
},
|
| 137 |
{
|
| 138 |
-
"epoch": 0.
|
| 139 |
-
"learning_rate":
|
| 140 |
-
"loss": 0.
|
| 141 |
"step": 220
|
| 142 |
},
|
| 143 |
{
|
| 144 |
-
"epoch":
|
| 145 |
-
"
|
| 146 |
-
"
|
| 147 |
-
"eval_runtime": 60.2714,
|
| 148 |
-
"eval_samples_per_second": 119.095,
|
| 149 |
-
"eval_steps_per_second": 3.733,
|
| 150 |
-
"step": 224
|
| 151 |
-
},
|
| 152 |
-
{
|
| 153 |
-
"epoch": 1.02,
|
| 154 |
-
"learning_rate": 4.985119047619048e-05,
|
| 155 |
-
"loss": 0.5157,
|
| 156 |
"step": 230
|
| 157 |
},
|
| 158 |
{
|
| 159 |
-
"epoch":
|
| 160 |
-
"learning_rate":
|
| 161 |
-
"loss": 0.
|
| 162 |
"step": 240
|
| 163 |
},
|
| 164 |
{
|
| 165 |
-
"epoch":
|
| 166 |
-
"learning_rate":
|
| 167 |
-
"loss": 0.
|
| 168 |
"step": 250
|
| 169 |
},
|
| 170 |
{
|
| 171 |
-
"epoch":
|
| 172 |
-
"learning_rate": 4.
|
| 173 |
-
"loss": 0.
|
| 174 |
"step": 260
|
| 175 |
},
|
| 176 |
{
|
| 177 |
-
"epoch":
|
| 178 |
-
"learning_rate": 4.
|
| 179 |
-
"loss": 0.
|
| 180 |
"step": 270
|
| 181 |
},
|
| 182 |
{
|
| 183 |
-
"epoch":
|
| 184 |
-
"learning_rate": 4.
|
| 185 |
-
"loss": 0.
|
| 186 |
"step": 280
|
| 187 |
},
|
| 188 |
{
|
| 189 |
-
"epoch":
|
| 190 |
-
"learning_rate": 4.
|
| 191 |
-
"loss": 0.
|
| 192 |
"step": 290
|
| 193 |
},
|
| 194 |
{
|
| 195 |
-
"epoch":
|
| 196 |
-
"learning_rate": 4.
|
| 197 |
-
"loss": 0.
|
| 198 |
"step": 300
|
| 199 |
},
|
| 200 |
{
|
| 201 |
-
"epoch":
|
| 202 |
-
"learning_rate": 4.
|
| 203 |
-
"loss": 0.
|
| 204 |
"step": 310
|
| 205 |
},
|
| 206 |
{
|
| 207 |
-
"epoch":
|
| 208 |
-
"learning_rate":
|
| 209 |
-
"loss": 0.
|
| 210 |
"step": 320
|
| 211 |
},
|
| 212 |
{
|
| 213 |
-
"epoch":
|
| 214 |
-
"learning_rate": 4.
|
| 215 |
-
"loss": 0.
|
| 216 |
"step": 330
|
| 217 |
},
|
| 218 |
{
|
| 219 |
-
"epoch":
|
| 220 |
-
"learning_rate": 4.
|
| 221 |
-
"loss": 0.
|
| 222 |
"step": 340
|
| 223 |
},
|
| 224 |
{
|
| 225 |
-
"epoch":
|
| 226 |
-
"learning_rate": 4.
|
| 227 |
-
"loss": 0.
|
| 228 |
"step": 350
|
| 229 |
},
|
| 230 |
{
|
| 231 |
-
"epoch":
|
| 232 |
-
"learning_rate": 4.
|
| 233 |
-
"loss": 0.
|
| 234 |
"step": 360
|
| 235 |
},
|
| 236 |
{
|
| 237 |
-
"epoch":
|
| 238 |
-
"learning_rate": 4.
|
| 239 |
-
"loss": 0.
|
| 240 |
"step": 370
|
| 241 |
},
|
| 242 |
{
|
| 243 |
-
"epoch":
|
| 244 |
-
"learning_rate": 4.
|
| 245 |
-
"loss": 0.
|
| 246 |
"step": 380
|
| 247 |
},
|
| 248 |
{
|
| 249 |
-
"epoch":
|
| 250 |
-
"learning_rate": 4.
|
| 251 |
-
"loss": 0.
|
| 252 |
"step": 390
|
| 253 |
},
|
| 254 |
{
|
| 255 |
-
"epoch":
|
| 256 |
-
"learning_rate": 4.
|
| 257 |
-
"loss": 0.
|
| 258 |
"step": 400
|
| 259 |
},
|
| 260 |
{
|
| 261 |
-
"epoch":
|
| 262 |
-
"learning_rate": 4.
|
| 263 |
-
"loss": 0.
|
| 264 |
"step": 410
|
| 265 |
},
|
| 266 |
{
|
| 267 |
-
"epoch":
|
| 268 |
-
"learning_rate": 4.
|
| 269 |
-
"loss": 0.
|
| 270 |
"step": 420
|
| 271 |
},
|
| 272 |
{
|
| 273 |
-
"epoch":
|
| 274 |
-
"learning_rate": 4.
|
| 275 |
-
"loss": 0.
|
| 276 |
"step": 430
|
| 277 |
},
|
| 278 |
{
|
| 279 |
-
"epoch":
|
| 280 |
-
"learning_rate": 4.
|
| 281 |
-
"loss": 0.
|
| 282 |
"step": 440
|
| 283 |
},
|
| 284 |
{
|
| 285 |
-
"epoch":
|
| 286 |
-
"
|
| 287 |
-
"
|
| 288 |
-
"eval_runtime": 60.4491,
|
| 289 |
-
"eval_samples_per_second": 118.744,
|
| 290 |
-
"eval_steps_per_second": 3.722,
|
| 291 |
-
"step": 449
|
| 292 |
-
},
|
| 293 |
-
{
|
| 294 |
-
"epoch": 2.0,
|
| 295 |
-
"learning_rate": 4.439484126984127e-05,
|
| 296 |
-
"loss": 0.4971,
|
| 297 |
"step": 450
|
| 298 |
},
|
| 299 |
{
|
| 300 |
-
"epoch":
|
| 301 |
-
"learning_rate": 4.
|
| 302 |
-
"loss": 0.
|
| 303 |
"step": 460
|
| 304 |
},
|
| 305 |
{
|
| 306 |
-
"epoch":
|
| 307 |
-
"learning_rate": 4.
|
| 308 |
-
"loss": 0.
|
| 309 |
"step": 470
|
| 310 |
},
|
| 311 |
{
|
| 312 |
-
"epoch":
|
| 313 |
-
"learning_rate": 4.
|
| 314 |
-
"loss": 0.
|
| 315 |
"step": 480
|
| 316 |
},
|
| 317 |
{
|
| 318 |
-
"epoch":
|
| 319 |
-
"learning_rate": 4.
|
| 320 |
-
"loss": 0.
|
| 321 |
"step": 490
|
| 322 |
},
|
| 323 |
{
|
| 324 |
-
"epoch":
|
| 325 |
-
"learning_rate": 4.
|
| 326 |
-
"loss": 0.
|
| 327 |
"step": 500
|
| 328 |
},
|
| 329 |
{
|
| 330 |
-
"epoch":
|
| 331 |
-
"learning_rate": 4.
|
| 332 |
-
"loss": 0.
|
| 333 |
"step": 510
|
| 334 |
},
|
| 335 |
{
|
| 336 |
-
"epoch":
|
| 337 |
-
"learning_rate": 4.
|
| 338 |
-
"loss": 0.
|
| 339 |
"step": 520
|
| 340 |
},
|
| 341 |
{
|
| 342 |
-
"epoch":
|
| 343 |
-
"learning_rate": 4.
|
| 344 |
-
"loss": 0.
|
| 345 |
"step": 530
|
| 346 |
},
|
| 347 |
{
|
| 348 |
-
"epoch":
|
| 349 |
-
"learning_rate": 4.
|
| 350 |
-
"loss": 0.
|
| 351 |
"step": 540
|
| 352 |
},
|
| 353 |
{
|
| 354 |
-
"epoch":
|
| 355 |
-
"learning_rate": 4.
|
| 356 |
-
"loss": 0.
|
| 357 |
"step": 550
|
| 358 |
},
|
| 359 |
{
|
| 360 |
-
"epoch":
|
| 361 |
-
"learning_rate": 4.
|
| 362 |
-
"loss": 0.
|
| 363 |
"step": 560
|
| 364 |
},
|
| 365 |
{
|
| 366 |
-
"epoch":
|
| 367 |
-
"learning_rate": 4.
|
| 368 |
-
"loss": 0.
|
| 369 |
"step": 570
|
| 370 |
},
|
| 371 |
{
|
| 372 |
-
"epoch":
|
| 373 |
-
"learning_rate": 4.
|
| 374 |
-
"loss": 0.
|
| 375 |
"step": 580
|
| 376 |
},
|
| 377 |
{
|
| 378 |
-
"epoch":
|
| 379 |
-
"learning_rate": 4.
|
| 380 |
-
"loss": 0.
|
| 381 |
"step": 590
|
| 382 |
},
|
| 383 |
{
|
| 384 |
-
"epoch":
|
| 385 |
-
"learning_rate": 4.
|
| 386 |
-
"loss": 0.
|
| 387 |
"step": 600
|
| 388 |
},
|
| 389 |
{
|
| 390 |
-
"epoch":
|
| 391 |
-
"learning_rate": 4.
|
| 392 |
-
"loss": 0.
|
| 393 |
"step": 610
|
| 394 |
},
|
| 395 |
{
|
| 396 |
-
"epoch":
|
| 397 |
-
"learning_rate": 4.
|
| 398 |
-
"loss": 0.
|
| 399 |
"step": 620
|
| 400 |
},
|
| 401 |
{
|
| 402 |
-
"epoch":
|
| 403 |
-
"learning_rate":
|
| 404 |
-
"loss": 0.
|
| 405 |
"step": 630
|
| 406 |
},
|
| 407 |
{
|
| 408 |
-
"epoch":
|
| 409 |
-
"learning_rate":
|
| 410 |
-
"loss": 0.
|
| 411 |
"step": 640
|
| 412 |
},
|
| 413 |
{
|
| 414 |
-
"epoch":
|
| 415 |
-
"learning_rate":
|
| 416 |
-
"loss": 0.
|
| 417 |
"step": 650
|
| 418 |
},
|
| 419 |
{
|
| 420 |
-
"epoch":
|
| 421 |
-
"learning_rate":
|
| 422 |
-
"loss": 0.
|
| 423 |
"step": 660
|
| 424 |
},
|
| 425 |
{
|
| 426 |
-
"epoch":
|
| 427 |
-
"learning_rate":
|
| 428 |
-
"loss": 0.
|
| 429 |
"step": 670
|
| 430 |
},
|
| 431 |
{
|
| 432 |
-
"epoch":
|
| 433 |
-
"
|
| 434 |
-
"
|
| 435 |
-
"eval_runtime": 60.0956,
|
| 436 |
-
"eval_samples_per_second": 119.443,
|
| 437 |
-
"eval_steps_per_second": 3.744,
|
| 438 |
-
"step": 673
|
| 439 |
-
},
|
| 440 |
-
{
|
| 441 |
-
"epoch": 3.03,
|
| 442 |
-
"learning_rate": 3.8690476190476195e-05,
|
| 443 |
-
"loss": 0.3228,
|
| 444 |
"step": 680
|
| 445 |
},
|
| 446 |
{
|
| 447 |
-
"epoch":
|
| 448 |
-
"learning_rate":
|
| 449 |
-
"loss": 0.
|
| 450 |
"step": 690
|
| 451 |
},
|
| 452 |
{
|
| 453 |
-
"epoch":
|
| 454 |
-
"learning_rate":
|
| 455 |
-
"loss": 0.
|
| 456 |
"step": 700
|
| 457 |
},
|
| 458 |
{
|
| 459 |
-
"epoch":
|
| 460 |
-
"learning_rate":
|
| 461 |
-
"loss": 0.
|
| 462 |
"step": 710
|
| 463 |
},
|
| 464 |
{
|
| 465 |
-
"epoch":
|
| 466 |
-
"learning_rate":
|
| 467 |
-
"loss": 0.
|
| 468 |
"step": 720
|
| 469 |
},
|
| 470 |
{
|
| 471 |
-
"epoch":
|
| 472 |
-
"learning_rate":
|
| 473 |
-
"loss": 0.
|
| 474 |
"step": 730
|
| 475 |
},
|
| 476 |
{
|
| 477 |
-
"epoch":
|
| 478 |
-
"learning_rate":
|
| 479 |
-
"loss": 0.
|
| 480 |
"step": 740
|
| 481 |
},
|
| 482 |
{
|
| 483 |
-
"epoch":
|
| 484 |
-
"learning_rate":
|
| 485 |
-
"loss": 0.
|
| 486 |
"step": 750
|
| 487 |
},
|
| 488 |
{
|
| 489 |
-
"epoch":
|
| 490 |
-
"learning_rate":
|
| 491 |
-
"loss": 0.
|
| 492 |
"step": 760
|
| 493 |
},
|
| 494 |
{
|
| 495 |
-
"epoch":
|
| 496 |
-
"learning_rate":
|
| 497 |
-
"loss": 0.
|
| 498 |
"step": 770
|
| 499 |
},
|
| 500 |
{
|
| 501 |
-
"epoch":
|
| 502 |
-
"learning_rate":
|
| 503 |
-
"loss": 0.
|
| 504 |
"step": 780
|
| 505 |
},
|
| 506 |
{
|
| 507 |
-
"epoch":
|
| 508 |
-
"learning_rate":
|
| 509 |
-
"loss": 0.
|
| 510 |
"step": 790
|
| 511 |
},
|
| 512 |
{
|
| 513 |
-
"epoch":
|
| 514 |
-
"
|
| 515 |
-
"
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
| 516 |
"step": 800
|
| 517 |
},
|
| 518 |
{
|
| 519 |
-
"epoch":
|
| 520 |
-
"learning_rate":
|
| 521 |
-
"loss": 0.
|
| 522 |
"step": 810
|
| 523 |
},
|
| 524 |
{
|
| 525 |
-
"epoch":
|
| 526 |
-
"learning_rate":
|
| 527 |
-
"loss": 0.
|
| 528 |
"step": 820
|
| 529 |
},
|
| 530 |
{
|
| 531 |
-
"epoch":
|
| 532 |
-
"learning_rate":
|
| 533 |
-
"loss": 0.
|
| 534 |
"step": 830
|
| 535 |
},
|
| 536 |
{
|
| 537 |
-
"epoch":
|
| 538 |
-
"learning_rate":
|
| 539 |
-
"loss": 0.
|
| 540 |
"step": 840
|
| 541 |
},
|
| 542 |
{
|
| 543 |
-
"epoch":
|
| 544 |
-
"learning_rate":
|
| 545 |
-
"loss": 0.
|
| 546 |
"step": 850
|
| 547 |
},
|
| 548 |
{
|
| 549 |
-
"epoch":
|
| 550 |
-
"learning_rate":
|
| 551 |
-
"loss": 0.
|
| 552 |
"step": 860
|
| 553 |
},
|
| 554 |
{
|
| 555 |
-
"epoch":
|
| 556 |
-
"learning_rate":
|
| 557 |
-
"loss": 0.
|
| 558 |
"step": 870
|
| 559 |
},
|
| 560 |
{
|
| 561 |
-
"epoch":
|
| 562 |
-
"learning_rate":
|
| 563 |
-
"loss": 0.
|
| 564 |
"step": 880
|
| 565 |
},
|
| 566 |
{
|
| 567 |
-
"epoch":
|
| 568 |
-
"learning_rate":
|
| 569 |
-
"loss": 0.
|
| 570 |
"step": 890
|
| 571 |
},
|
| 572 |
{
|
| 573 |
-
"epoch":
|
| 574 |
-
"
|
| 575 |
-
"
|
| 576 |
-
"eval_runtime": 60.4275,
|
| 577 |
-
"eval_samples_per_second": 118.787,
|
| 578 |
-
"eval_steps_per_second": 3.723,
|
| 579 |
-
"step": 898
|
| 580 |
-
},
|
| 581 |
-
{
|
| 582 |
-
"epoch": 4.01,
|
| 583 |
-
"learning_rate": 3.3234126984126983e-05,
|
| 584 |
-
"loss": 0.3389,
|
| 585 |
"step": 900
|
| 586 |
},
|
| 587 |
{
|
| 588 |
-
"epoch":
|
| 589 |
-
"learning_rate": 3.
|
| 590 |
-
"loss": 0.
|
| 591 |
"step": 910
|
| 592 |
},
|
| 593 |
{
|
| 594 |
-
"epoch":
|
| 595 |
-
"learning_rate": 3.
|
| 596 |
-
"loss": 0.
|
| 597 |
"step": 920
|
| 598 |
},
|
| 599 |
{
|
| 600 |
-
"epoch":
|
| 601 |
-
"learning_rate": 3.
|
| 602 |
-
"loss": 0.
|
| 603 |
"step": 930
|
| 604 |
},
|
| 605 |
{
|
| 606 |
-
"epoch":
|
| 607 |
-
"learning_rate": 3.
|
| 608 |
-
"loss": 0.
|
| 609 |
"step": 940
|
| 610 |
},
|
| 611 |
{
|
| 612 |
-
"epoch":
|
| 613 |
-
"learning_rate": 3.
|
| 614 |
-
"loss": 0.
|
| 615 |
"step": 950
|
| 616 |
},
|
| 617 |
{
|
| 618 |
-
"epoch":
|
| 619 |
-
"learning_rate": 3.
|
| 620 |
-
"loss": 0.
|
| 621 |
"step": 960
|
| 622 |
},
|
| 623 |
{
|
| 624 |
-
"epoch":
|
| 625 |
-
"learning_rate": 3.
|
| 626 |
-
"loss": 0.
|
| 627 |
"step": 970
|
| 628 |
},
|
| 629 |
{
|
| 630 |
-
"epoch":
|
| 631 |
-
"learning_rate": 3.
|
| 632 |
-
"loss": 0.
|
| 633 |
"step": 980
|
| 634 |
},
|
| 635 |
{
|
| 636 |
-
"epoch":
|
| 637 |
-
"learning_rate": 3.
|
| 638 |
-
"loss": 0.
|
| 639 |
"step": 990
|
| 640 |
},
|
| 641 |
{
|
| 642 |
-
"epoch":
|
| 643 |
-
"learning_rate": 3.
|
| 644 |
-
"loss": 0.
|
| 645 |
"step": 1000
|
| 646 |
},
|
| 647 |
{
|
| 648 |
-
"epoch":
|
| 649 |
-
"learning_rate": 3.
|
| 650 |
-
"loss": 0.
|
| 651 |
"step": 1010
|
| 652 |
},
|
| 653 |
{
|
| 654 |
-
"epoch":
|
| 655 |
-
"learning_rate": 3.
|
| 656 |
-
"loss": 0.
|
| 657 |
"step": 1020
|
| 658 |
},
|
| 659 |
{
|
| 660 |
-
"epoch":
|
| 661 |
-
"learning_rate": 3.
|
| 662 |
-
"loss": 0.
|
| 663 |
"step": 1030
|
| 664 |
},
|
| 665 |
{
|
| 666 |
-
"epoch":
|
| 667 |
-
"learning_rate":
|
| 668 |
-
"loss": 0.
|
| 669 |
"step": 1040
|
| 670 |
},
|
| 671 |
{
|
| 672 |
-
"epoch":
|
| 673 |
-
"learning_rate":
|
| 674 |
-
"loss": 0.
|
| 675 |
"step": 1050
|
| 676 |
},
|
| 677 |
{
|
| 678 |
-
"epoch":
|
| 679 |
-
"learning_rate":
|
| 680 |
-
"loss": 0.
|
| 681 |
"step": 1060
|
| 682 |
},
|
| 683 |
{
|
| 684 |
-
"epoch":
|
| 685 |
-
"learning_rate":
|
| 686 |
-
"loss": 0.
|
| 687 |
"step": 1070
|
| 688 |
},
|
| 689 |
{
|
| 690 |
-
"epoch":
|
| 691 |
-
"learning_rate":
|
| 692 |
-
"loss": 0.
|
| 693 |
"step": 1080
|
| 694 |
},
|
| 695 |
{
|
| 696 |
-
"epoch":
|
| 697 |
-
"learning_rate":
|
| 698 |
-
"loss": 0.
|
| 699 |
"step": 1090
|
| 700 |
},
|
| 701 |
{
|
| 702 |
-
"epoch":
|
| 703 |
-
"learning_rate":
|
| 704 |
-
"loss": 0.
|
| 705 |
"step": 1100
|
| 706 |
},
|
| 707 |
{
|
| 708 |
-
"epoch":
|
| 709 |
-
"learning_rate":
|
| 710 |
-
"loss": 0.
|
| 711 |
"step": 1110
|
| 712 |
},
|
| 713 |
{
|
| 714 |
-
"epoch": 4
|
| 715 |
-
"learning_rate":
|
| 716 |
-
"loss": 0.
|
| 717 |
"step": 1120
|
| 718 |
},
|
| 719 |
{
|
| 720 |
-
"epoch":
|
| 721 |
-
"
|
| 722 |
-
"
|
| 723 |
-
"eval_runtime": 60.4436,
|
| 724 |
-
"eval_samples_per_second": 118.755,
|
| 725 |
-
"eval_steps_per_second": 3.722,
|
| 726 |
-
"step": 1122
|
| 727 |
-
},
|
| 728 |
-
{
|
| 729 |
-
"epoch": 5.03,
|
| 730 |
-
"learning_rate": 2.7529761904761907e-05,
|
| 731 |
-
"loss": 0.2474,
|
| 732 |
"step": 1130
|
| 733 |
},
|
| 734 |
{
|
| 735 |
-
"epoch":
|
| 736 |
-
"learning_rate":
|
| 737 |
-
"loss": 0.
|
| 738 |
"step": 1140
|
| 739 |
},
|
| 740 |
{
|
| 741 |
-
"epoch":
|
| 742 |
-
"learning_rate":
|
| 743 |
-
"loss": 0.
|
| 744 |
"step": 1150
|
| 745 |
},
|
| 746 |
{
|
| 747 |
-
"epoch":
|
| 748 |
-
"learning_rate":
|
| 749 |
-
"loss": 0.
|
| 750 |
"step": 1160
|
| 751 |
},
|
| 752 |
{
|
| 753 |
-
"epoch":
|
| 754 |
-
"learning_rate":
|
| 755 |
-
"loss": 0.
|
| 756 |
"step": 1170
|
| 757 |
},
|
| 758 |
{
|
| 759 |
-
"epoch":
|
| 760 |
-
"learning_rate":
|
| 761 |
-
"loss": 0.
|
| 762 |
"step": 1180
|
| 763 |
},
|
| 764 |
{
|
| 765 |
-
"epoch":
|
| 766 |
-
"learning_rate":
|
| 767 |
-
"loss": 0.
|
| 768 |
"step": 1190
|
| 769 |
},
|
| 770 |
{
|
| 771 |
-
"epoch": 5
|
| 772 |
-
"learning_rate":
|
| 773 |
-
"loss": 0.
|
| 774 |
"step": 1200
|
| 775 |
},
|
| 776 |
{
|
| 777 |
-
"epoch":
|
| 778 |
-
"learning_rate":
|
| 779 |
-
"loss": 0.
|
| 780 |
"step": 1210
|
| 781 |
},
|
| 782 |
{
|
| 783 |
-
"epoch":
|
| 784 |
-
"learning_rate":
|
| 785 |
-
"loss": 0.
|
| 786 |
"step": 1220
|
| 787 |
},
|
| 788 |
{
|
| 789 |
-
"epoch":
|
| 790 |
-
"learning_rate":
|
| 791 |
-
"loss": 0.
|
| 792 |
"step": 1230
|
| 793 |
},
|
| 794 |
{
|
| 795 |
-
"epoch":
|
| 796 |
-
"learning_rate":
|
| 797 |
-
"loss": 0.
|
| 798 |
"step": 1240
|
| 799 |
},
|
| 800 |
{
|
| 801 |
-
"epoch":
|
| 802 |
-
"learning_rate":
|
| 803 |
-
"loss": 0.
|
| 804 |
"step": 1250
|
| 805 |
},
|
| 806 |
{
|
| 807 |
-
"epoch":
|
| 808 |
-
"learning_rate":
|
| 809 |
-
"loss": 0.
|
| 810 |
"step": 1260
|
| 811 |
},
|
| 812 |
{
|
| 813 |
-
"epoch":
|
| 814 |
-
"learning_rate":
|
| 815 |
-
"loss": 0.
|
| 816 |
"step": 1270
|
| 817 |
},
|
| 818 |
{
|
| 819 |
-
"epoch":
|
| 820 |
-
"learning_rate":
|
| 821 |
-
"loss": 0.
|
| 822 |
"step": 1280
|
| 823 |
},
|
| 824 |
{
|
| 825 |
-
"epoch":
|
| 826 |
-
"learning_rate":
|
| 827 |
-
"loss": 0.
|
| 828 |
"step": 1290
|
| 829 |
},
|
| 830 |
{
|
| 831 |
-
"epoch":
|
| 832 |
-
"learning_rate":
|
| 833 |
-
"loss": 0.
|
| 834 |
"step": 1300
|
| 835 |
},
|
| 836 |
{
|
| 837 |
-
"epoch":
|
| 838 |
-
"learning_rate":
|
| 839 |
-
"loss": 0.
|
| 840 |
"step": 1310
|
| 841 |
},
|
| 842 |
{
|
| 843 |
-
"epoch":
|
| 844 |
-
"learning_rate":
|
| 845 |
-
"loss": 0.
|
| 846 |
"step": 1320
|
| 847 |
},
|
| 848 |
{
|
| 849 |
-
"epoch":
|
| 850 |
-
"learning_rate":
|
| 851 |
-
"loss": 0.
|
| 852 |
"step": 1330
|
| 853 |
},
|
| 854 |
{
|
| 855 |
-
"epoch":
|
| 856 |
-
"learning_rate":
|
| 857 |
-
"loss": 0.
|
| 858 |
"step": 1340
|
| 859 |
},
|
| 860 |
{
|
| 861 |
-
"epoch":
|
| 862 |
-
"
|
| 863 |
-
"
|
| 864 |
-
"eval_runtime": 60.3923,
|
| 865 |
-
"eval_samples_per_second": 118.856,
|
| 866 |
-
"eval_steps_per_second": 3.726,
|
| 867 |
-
"step": 1347
|
| 868 |
-
},
|
| 869 |
-
{
|
| 870 |
-
"epoch": 6.01,
|
| 871 |
-
"learning_rate": 2.20734126984127e-05,
|
| 872 |
-
"loss": 0.2192,
|
| 873 |
"step": 1350
|
| 874 |
},
|
| 875 |
{
|
| 876 |
-
"epoch":
|
| 877 |
-
"learning_rate":
|
| 878 |
-
"loss": 0.
|
| 879 |
"step": 1360
|
| 880 |
},
|
| 881 |
{
|
| 882 |
-
"epoch":
|
| 883 |
-
"learning_rate":
|
| 884 |
-
"loss": 0.
|
| 885 |
"step": 1370
|
| 886 |
},
|
| 887 |
{
|
| 888 |
-
"epoch":
|
| 889 |
-
"learning_rate":
|
| 890 |
-
"loss": 0.
|
| 891 |
"step": 1380
|
| 892 |
},
|
| 893 |
{
|
| 894 |
-
"epoch":
|
| 895 |
-
"learning_rate":
|
| 896 |
-
"loss": 0.
|
| 897 |
"step": 1390
|
| 898 |
},
|
| 899 |
{
|
| 900 |
-
"epoch":
|
| 901 |
-
"learning_rate":
|
| 902 |
-
"loss": 0.
|
| 903 |
"step": 1400
|
| 904 |
},
|
| 905 |
{
|
| 906 |
-
"epoch":
|
| 907 |
-
"learning_rate":
|
| 908 |
-
"loss": 0.
|
| 909 |
"step": 1410
|
| 910 |
},
|
| 911 |
{
|
| 912 |
-
"epoch":
|
| 913 |
-
"learning_rate":
|
| 914 |
-
"loss": 0.
|
| 915 |
"step": 1420
|
| 916 |
},
|
| 917 |
{
|
| 918 |
-
"epoch":
|
| 919 |
-
"learning_rate":
|
| 920 |
-
"loss": 0.
|
| 921 |
"step": 1430
|
| 922 |
},
|
| 923 |
{
|
| 924 |
-
"epoch":
|
| 925 |
-
"learning_rate":
|
| 926 |
-
"loss": 0.
|
| 927 |
"step": 1440
|
| 928 |
},
|
| 929 |
{
|
| 930 |
-
"epoch":
|
| 931 |
-
"learning_rate":
|
| 932 |
-
"loss": 0.
|
| 933 |
"step": 1450
|
| 934 |
},
|
| 935 |
{
|
| 936 |
-
"epoch":
|
| 937 |
-
"learning_rate":
|
| 938 |
-
"loss": 0.
|
| 939 |
"step": 1460
|
| 940 |
},
|
| 941 |
{
|
| 942 |
-
"epoch":
|
| 943 |
-
"learning_rate":
|
| 944 |
-
"loss": 0.
|
| 945 |
"step": 1470
|
| 946 |
},
|
| 947 |
{
|
| 948 |
-
"epoch":
|
| 949 |
-
"learning_rate":
|
| 950 |
-
"loss": 0.
|
| 951 |
"step": 1480
|
| 952 |
},
|
| 953 |
{
|
| 954 |
-
"epoch":
|
| 955 |
-
"learning_rate":
|
| 956 |
-
"loss": 0.
|
| 957 |
"step": 1490
|
| 958 |
},
|
| 959 |
{
|
| 960 |
-
"epoch":
|
| 961 |
-
"learning_rate":
|
| 962 |
-
"loss": 0.
|
| 963 |
"step": 1500
|
| 964 |
},
|
| 965 |
{
|
| 966 |
-
"epoch":
|
| 967 |
-
"learning_rate":
|
| 968 |
-
"loss": 0.
|
| 969 |
"step": 1510
|
| 970 |
},
|
| 971 |
{
|
| 972 |
-
"epoch":
|
| 973 |
-
"learning_rate":
|
| 974 |
-
"loss": 0.
|
| 975 |
"step": 1520
|
| 976 |
},
|
| 977 |
{
|
| 978 |
-
"epoch":
|
| 979 |
-
"learning_rate":
|
| 980 |
-
"loss": 0.
|
| 981 |
"step": 1530
|
| 982 |
},
|
| 983 |
{
|
| 984 |
-
"epoch":
|
| 985 |
-
"learning_rate":
|
| 986 |
-
"loss": 0.
|
| 987 |
"step": 1540
|
| 988 |
},
|
| 989 |
{
|
| 990 |
-
"epoch":
|
| 991 |
-
"learning_rate":
|
| 992 |
-
"loss": 0.
|
| 993 |
"step": 1550
|
| 994 |
},
|
| 995 |
{
|
| 996 |
-
"epoch":
|
| 997 |
-
"learning_rate":
|
| 998 |
-
"loss": 0.
|
| 999 |
"step": 1560
|
| 1000 |
},
|
| 1001 |
{
|
| 1002 |
-
"epoch":
|
| 1003 |
-
"learning_rate":
|
| 1004 |
-
"loss": 0.
|
| 1005 |
"step": 1570
|
| 1006 |
},
|
| 1007 |
{
|
| 1008 |
-
"epoch":
|
| 1009 |
-
"
|
| 1010 |
-
"
|
| 1011 |
-
"eval_runtime": 59.9724,
|
| 1012 |
-
"eval_samples_per_second": 119.688,
|
| 1013 |
-
"eval_steps_per_second": 3.752,
|
| 1014 |
-
"step": 1571
|
| 1015 |
-
},
|
| 1016 |
-
{
|
| 1017 |
-
"epoch": 7.04,
|
| 1018 |
-
"learning_rate": 1.636904761904762e-05,
|
| 1019 |
-
"loss": 0.1495,
|
| 1020 |
"step": 1580
|
| 1021 |
},
|
| 1022 |
{
|
| 1023 |
-
"epoch":
|
| 1024 |
-
"learning_rate":
|
| 1025 |
-
"loss": 0.
|
| 1026 |
"step": 1590
|
| 1027 |
},
|
| 1028 |
{
|
| 1029 |
-
"epoch":
|
| 1030 |
-
"
|
| 1031 |
-
"
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
| 1032 |
"step": 1600
|
| 1033 |
},
|
| 1034 |
{
|
| 1035 |
-
"epoch":
|
| 1036 |
-
"learning_rate":
|
| 1037 |
-
"loss": 0.
|
| 1038 |
"step": 1610
|
| 1039 |
},
|
| 1040 |
{
|
| 1041 |
-
"epoch":
|
| 1042 |
-
"learning_rate":
|
| 1043 |
-
"loss": 0.
|
| 1044 |
"step": 1620
|
| 1045 |
},
|
| 1046 |
{
|
| 1047 |
-
"epoch":
|
| 1048 |
-
"learning_rate":
|
| 1049 |
-
"loss": 0.
|
| 1050 |
"step": 1630
|
| 1051 |
},
|
| 1052 |
{
|
| 1053 |
-
"epoch":
|
| 1054 |
-
"learning_rate":
|
| 1055 |
-
"loss": 0.
|
| 1056 |
"step": 1640
|
| 1057 |
},
|
| 1058 |
{
|
| 1059 |
-
"epoch":
|
| 1060 |
-
"learning_rate":
|
| 1061 |
-
"loss": 0.
|
| 1062 |
"step": 1650
|
| 1063 |
},
|
| 1064 |
{
|
| 1065 |
-
"epoch":
|
| 1066 |
-
"learning_rate":
|
| 1067 |
-
"loss": 0.
|
| 1068 |
"step": 1660
|
| 1069 |
},
|
| 1070 |
{
|
| 1071 |
-
"epoch":
|
| 1072 |
-
"learning_rate":
|
| 1073 |
-
"loss": 0.
|
| 1074 |
"step": 1670
|
| 1075 |
},
|
| 1076 |
{
|
| 1077 |
-
"epoch":
|
| 1078 |
-
"learning_rate":
|
| 1079 |
-
"loss": 0.
|
| 1080 |
"step": 1680
|
| 1081 |
},
|
| 1082 |
{
|
| 1083 |
-
"epoch":
|
| 1084 |
-
"learning_rate":
|
| 1085 |
-
"loss": 0.
|
| 1086 |
"step": 1690
|
| 1087 |
},
|
| 1088 |
{
|
| 1089 |
-
"epoch":
|
| 1090 |
-
"learning_rate":
|
| 1091 |
-
"loss": 0.
|
| 1092 |
"step": 1700
|
| 1093 |
},
|
| 1094 |
{
|
| 1095 |
-
"epoch":
|
| 1096 |
-
"learning_rate":
|
| 1097 |
-
"loss": 0.
|
| 1098 |
"step": 1710
|
| 1099 |
},
|
| 1100 |
{
|
| 1101 |
-
"epoch":
|
| 1102 |
-
"learning_rate":
|
| 1103 |
-
"loss": 0.
|
| 1104 |
"step": 1720
|
| 1105 |
},
|
| 1106 |
{
|
| 1107 |
-
"epoch":
|
| 1108 |
-
"learning_rate":
|
| 1109 |
-
"loss": 0.
|
| 1110 |
"step": 1730
|
| 1111 |
},
|
| 1112 |
{
|
| 1113 |
-
"epoch":
|
| 1114 |
-
"learning_rate":
|
| 1115 |
-
"loss": 0.
|
| 1116 |
"step": 1740
|
| 1117 |
},
|
| 1118 |
{
|
| 1119 |
-
"epoch":
|
| 1120 |
-
"learning_rate":
|
| 1121 |
-
"loss": 0.
|
| 1122 |
"step": 1750
|
| 1123 |
},
|
| 1124 |
{
|
| 1125 |
-
"epoch":
|
| 1126 |
-
"learning_rate":
|
| 1127 |
-
"loss": 0.
|
| 1128 |
"step": 1760
|
| 1129 |
},
|
| 1130 |
{
|
| 1131 |
-
"epoch":
|
| 1132 |
-
"learning_rate":
|
| 1133 |
-
"loss": 0.
|
| 1134 |
"step": 1770
|
| 1135 |
},
|
| 1136 |
{
|
| 1137 |
-
"epoch":
|
| 1138 |
-
"learning_rate":
|
| 1139 |
-
"loss": 0.
|
| 1140 |
"step": 1780
|
| 1141 |
},
|
| 1142 |
{
|
| 1143 |
-
"epoch":
|
| 1144 |
-
"learning_rate":
|
| 1145 |
-
"loss": 0.
|
| 1146 |
"step": 1790
|
| 1147 |
},
|
| 1148 |
{
|
| 1149 |
-
"epoch":
|
| 1150 |
-
"
|
| 1151 |
-
"
|
| 1152 |
-
"eval_runtime": 59.6867,
|
| 1153 |
-
"eval_samples_per_second": 120.261,
|
| 1154 |
-
"eval_steps_per_second": 3.77,
|
| 1155 |
-
"step": 1796
|
| 1156 |
-
},
|
| 1157 |
-
{
|
| 1158 |
-
"epoch": 8.02,
|
| 1159 |
-
"learning_rate": 1.0912698412698414e-05,
|
| 1160 |
-
"loss": 0.1218,
|
| 1161 |
"step": 1800
|
| 1162 |
},
|
| 1163 |
{
|
| 1164 |
-
"epoch":
|
| 1165 |
-
"learning_rate":
|
| 1166 |
-
"loss": 0.
|
| 1167 |
"step": 1810
|
| 1168 |
},
|
| 1169 |
{
|
| 1170 |
-
"epoch":
|
| 1171 |
-
"learning_rate":
|
| 1172 |
-
"loss": 0.
|
| 1173 |
"step": 1820
|
| 1174 |
},
|
| 1175 |
{
|
| 1176 |
-
"epoch":
|
| 1177 |
-
"learning_rate":
|
| 1178 |
-
"loss": 0.
|
| 1179 |
"step": 1830
|
| 1180 |
},
|
| 1181 |
{
|
| 1182 |
-
"epoch":
|
| 1183 |
-
"learning_rate":
|
| 1184 |
-
"loss": 0.
|
| 1185 |
"step": 1840
|
| 1186 |
},
|
| 1187 |
{
|
| 1188 |
-
"epoch":
|
| 1189 |
-
"learning_rate":
|
| 1190 |
-
"loss": 0.
|
| 1191 |
"step": 1850
|
| 1192 |
},
|
| 1193 |
{
|
| 1194 |
-
"epoch":
|
| 1195 |
-
"learning_rate":
|
| 1196 |
-
"loss": 0.
|
| 1197 |
"step": 1860
|
| 1198 |
},
|
| 1199 |
{
|
| 1200 |
-
"epoch":
|
| 1201 |
-
"learning_rate":
|
| 1202 |
-
"loss": 0.
|
| 1203 |
"step": 1870
|
| 1204 |
},
|
| 1205 |
{
|
| 1206 |
-
"epoch":
|
| 1207 |
-
"learning_rate":
|
| 1208 |
-
"loss": 0.
|
| 1209 |
"step": 1880
|
| 1210 |
},
|
| 1211 |
{
|
| 1212 |
-
"epoch":
|
| 1213 |
-
"learning_rate":
|
| 1214 |
-
"loss": 0.
|
| 1215 |
"step": 1890
|
| 1216 |
},
|
| 1217 |
{
|
| 1218 |
-
"epoch":
|
| 1219 |
-
"learning_rate":
|
| 1220 |
-
"loss": 0.
|
| 1221 |
"step": 1900
|
| 1222 |
},
|
| 1223 |
{
|
| 1224 |
-
"epoch":
|
| 1225 |
-
"learning_rate":
|
| 1226 |
-
"loss": 0.
|
| 1227 |
"step": 1910
|
| 1228 |
},
|
| 1229 |
{
|
| 1230 |
-
"epoch":
|
| 1231 |
-
"learning_rate":
|
| 1232 |
-
"loss": 0.
|
| 1233 |
"step": 1920
|
| 1234 |
},
|
| 1235 |
{
|
| 1236 |
-
"epoch":
|
| 1237 |
-
"learning_rate":
|
| 1238 |
-
"loss": 0.
|
| 1239 |
"step": 1930
|
| 1240 |
},
|
| 1241 |
{
|
| 1242 |
-
"epoch":
|
| 1243 |
-
"learning_rate":
|
| 1244 |
-
"loss": 0.
|
| 1245 |
"step": 1940
|
| 1246 |
},
|
| 1247 |
{
|
| 1248 |
-
"epoch":
|
| 1249 |
-
"learning_rate":
|
| 1250 |
-
"loss": 0.
|
| 1251 |
"step": 1950
|
| 1252 |
},
|
| 1253 |
{
|
| 1254 |
-
"epoch":
|
| 1255 |
-
"learning_rate":
|
| 1256 |
-
"loss": 0.
|
| 1257 |
"step": 1960
|
| 1258 |
},
|
| 1259 |
{
|
| 1260 |
-
"epoch":
|
| 1261 |
-
"learning_rate":
|
| 1262 |
-
"loss": 0.
|
| 1263 |
"step": 1970
|
| 1264 |
},
|
| 1265 |
{
|
| 1266 |
-
"epoch":
|
| 1267 |
-
"learning_rate":
|
| 1268 |
-
"loss": 0.
|
| 1269 |
"step": 1980
|
| 1270 |
},
|
| 1271 |
{
|
| 1272 |
-
"epoch":
|
| 1273 |
-
"learning_rate":
|
| 1274 |
-
"loss": 0.
|
| 1275 |
"step": 1990
|
| 1276 |
},
|
| 1277 |
{
|
| 1278 |
-
"epoch":
|
| 1279 |
-
"learning_rate":
|
| 1280 |
-
"loss": 0.
|
| 1281 |
"step": 2000
|
| 1282 |
},
|
| 1283 |
{
|
| 1284 |
-
"epoch":
|
| 1285 |
-
"learning_rate":
|
| 1286 |
-
"loss": 0.
|
| 1287 |
"step": 2010
|
| 1288 |
},
|
| 1289 |
{
|
| 1290 |
-
"epoch":
|
| 1291 |
-
"learning_rate":
|
| 1292 |
-
"loss": 0.
|
| 1293 |
"step": 2020
|
| 1294 |
},
|
| 1295 |
{
|
| 1296 |
-
"epoch":
|
| 1297 |
-
"
|
| 1298 |
-
"
|
| 1299 |
-
"eval_runtime": 59.8317,
|
| 1300 |
-
"eval_samples_per_second": 119.97,
|
| 1301 |
-
"eval_steps_per_second": 3.761,
|
| 1302 |
-
"step": 2020
|
| 1303 |
-
},
|
| 1304 |
-
{
|
| 1305 |
-
"epoch": 9.04,
|
| 1306 |
-
"learning_rate": 5.208333333333334e-06,
|
| 1307 |
-
"loss": 0.078,
|
| 1308 |
"step": 2030
|
| 1309 |
},
|
| 1310 |
{
|
| 1311 |
-
"epoch":
|
| 1312 |
-
"learning_rate":
|
| 1313 |
-
"loss": 0.
|
| 1314 |
"step": 2040
|
| 1315 |
},
|
| 1316 |
{
|
| 1317 |
-
"epoch":
|
| 1318 |
-
"learning_rate":
|
| 1319 |
-
"loss": 0.
|
| 1320 |
"step": 2050
|
| 1321 |
},
|
| 1322 |
{
|
| 1323 |
-
"epoch":
|
| 1324 |
-
"learning_rate":
|
| 1325 |
-
"loss": 0.
|
| 1326 |
"step": 2060
|
| 1327 |
},
|
| 1328 |
{
|
| 1329 |
-
"epoch":
|
| 1330 |
-
"learning_rate":
|
| 1331 |
-
"loss": 0.
|
| 1332 |
"step": 2070
|
| 1333 |
},
|
| 1334 |
{
|
| 1335 |
-
"epoch":
|
| 1336 |
-
"learning_rate": 3.
|
| 1337 |
-
"loss": 0.
|
| 1338 |
"step": 2080
|
| 1339 |
},
|
| 1340 |
{
|
| 1341 |
-
"epoch":
|
| 1342 |
-
"learning_rate": 3.
|
| 1343 |
-
"loss": 0.
|
| 1344 |
"step": 2090
|
| 1345 |
},
|
| 1346 |
{
|
| 1347 |
-
"epoch":
|
| 1348 |
-
"learning_rate": 3.
|
| 1349 |
-
"loss": 0.
|
| 1350 |
"step": 2100
|
| 1351 |
},
|
| 1352 |
{
|
| 1353 |
-
"epoch":
|
| 1354 |
-
"learning_rate": 3.
|
| 1355 |
-
"loss": 0.
|
| 1356 |
"step": 2110
|
| 1357 |
},
|
| 1358 |
{
|
| 1359 |
-
"epoch":
|
| 1360 |
-
"learning_rate":
|
| 1361 |
-
"loss": 0.
|
| 1362 |
"step": 2120
|
| 1363 |
},
|
| 1364 |
{
|
| 1365 |
-
"epoch":
|
| 1366 |
-
"learning_rate":
|
| 1367 |
-
"loss": 0.
|
| 1368 |
"step": 2130
|
| 1369 |
},
|
| 1370 |
{
|
| 1371 |
-
"epoch":
|
| 1372 |
-
"learning_rate":
|
| 1373 |
-
"loss": 0.
|
| 1374 |
"step": 2140
|
| 1375 |
},
|
| 1376 |
{
|
| 1377 |
-
"epoch":
|
| 1378 |
-
"learning_rate":
|
| 1379 |
-
"loss": 0.
|
| 1380 |
"step": 2150
|
| 1381 |
},
|
| 1382 |
{
|
| 1383 |
-
"epoch":
|
| 1384 |
-
"learning_rate":
|
| 1385 |
-
"loss": 0.
|
| 1386 |
"step": 2160
|
| 1387 |
},
|
| 1388 |
{
|
| 1389 |
-
"epoch":
|
| 1390 |
-
"learning_rate":
|
| 1391 |
-
"loss": 0.
|
| 1392 |
"step": 2170
|
| 1393 |
},
|
| 1394 |
{
|
| 1395 |
-
"epoch":
|
| 1396 |
-
"learning_rate":
|
| 1397 |
-
"loss": 0.
|
| 1398 |
"step": 2180
|
| 1399 |
},
|
| 1400 |
{
|
| 1401 |
-
"epoch":
|
| 1402 |
-
"learning_rate":
|
| 1403 |
-
"loss": 0.
|
| 1404 |
"step": 2190
|
| 1405 |
},
|
| 1406 |
{
|
| 1407 |
-
"epoch":
|
| 1408 |
-
"learning_rate":
|
| 1409 |
-
"loss": 0.
|
| 1410 |
"step": 2200
|
| 1411 |
},
|
| 1412 |
{
|
| 1413 |
-
"epoch":
|
| 1414 |
-
"learning_rate":
|
| 1415 |
-
"loss": 0.
|
| 1416 |
"step": 2210
|
| 1417 |
},
|
| 1418 |
{
|
| 1419 |
-
"epoch":
|
| 1420 |
-
"learning_rate":
|
| 1421 |
-
"loss": 0.
|
| 1422 |
"step": 2220
|
| 1423 |
},
|
| 1424 |
{
|
| 1425 |
-
"epoch":
|
| 1426 |
-
"learning_rate": 2.
|
| 1427 |
-
"loss": 0.
|
| 1428 |
"step": 2230
|
| 1429 |
},
|
| 1430 |
{
|
| 1431 |
-
"epoch":
|
| 1432 |
-
"learning_rate":
|
| 1433 |
-
"loss": 0.
|
| 1434 |
"step": 2240
|
| 1435 |
},
|
| 1436 |
{
|
| 1437 |
-
"epoch":
|
| 1438 |
-
"
|
| 1439 |
-
"
|
| 1440 |
-
"
|
| 1441 |
-
|
| 1442 |
-
|
| 1443 |
-
"
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
| 1444 |
},
|
| 1445 |
{
|
| 1446 |
-
"epoch":
|
| 1447 |
-
"
|
| 1448 |
-
"
|
| 1449 |
-
"
|
| 1450 |
-
|
| 1451 |
-
|
| 1452 |
-
"
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
| 1453 |
}
|
| 1454 |
],
|
| 1455 |
"logging_steps": 10,
|
| 1456 |
-
"max_steps":
|
| 1457 |
"num_input_tokens_seen": 0,
|
| 1458 |
-
"num_train_epochs":
|
| 1459 |
"save_steps": 500,
|
| 1460 |
-
"total_flos":
|
| 1461 |
"train_batch_size": 32,
|
| 1462 |
"trial_name": null,
|
| 1463 |
"trial_params": null
|
|
|
|
| 1 |
{
|
| 2 |
+
"best_metric": 0.7839129072681704,
|
| 3 |
+
"best_model_checkpoint": "vit-base-patch16-224-in21k-finetuned/checkpoint-3192",
|
| 4 |
+
"epoch": 6.0,
|
| 5 |
"eval_steps": 500,
|
| 6 |
+
"global_step": 4788,
|
| 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 |
+
"learning_rate": 1.5625e-06,
|
| 14 |
+
"loss": 1.5128,
|
| 15 |
"step": 10
|
| 16 |
},
|
| 17 |
{
|
| 18 |
+
"epoch": 0.03,
|
| 19 |
+
"learning_rate": 3.125e-06,
|
| 20 |
+
"loss": 1.3411,
|
| 21 |
"step": 20
|
| 22 |
},
|
| 23 |
{
|
| 24 |
+
"epoch": 0.04,
|
| 25 |
+
"learning_rate": 4.6875000000000004e-06,
|
| 26 |
+
"loss": 1.2153,
|
| 27 |
"step": 30
|
| 28 |
},
|
| 29 |
{
|
| 30 |
+
"epoch": 0.05,
|
| 31 |
+
"learning_rate": 6.25e-06,
|
| 32 |
+
"loss": 1.2697,
|
| 33 |
"step": 40
|
| 34 |
},
|
| 35 |
{
|
| 36 |
+
"epoch": 0.06,
|
| 37 |
+
"learning_rate": 7.8125e-06,
|
| 38 |
+
"loss": 1.1394,
|
| 39 |
"step": 50
|
| 40 |
},
|
| 41 |
{
|
| 42 |
+
"epoch": 0.08,
|
| 43 |
+
"learning_rate": 9.375000000000001e-06,
|
| 44 |
+
"loss": 1.0215,
|
| 45 |
"step": 60
|
| 46 |
},
|
| 47 |
{
|
| 48 |
+
"epoch": 0.09,
|
| 49 |
+
"learning_rate": 1.09375e-05,
|
| 50 |
+
"loss": 0.9762,
|
| 51 |
"step": 70
|
| 52 |
},
|
| 53 |
{
|
| 54 |
+
"epoch": 0.1,
|
| 55 |
+
"learning_rate": 1.25e-05,
|
| 56 |
+
"loss": 0.8893,
|
| 57 |
"step": 80
|
| 58 |
},
|
| 59 |
{
|
| 60 |
+
"epoch": 0.11,
|
| 61 |
+
"learning_rate": 1.4062500000000001e-05,
|
| 62 |
+
"loss": 0.9721,
|
| 63 |
"step": 90
|
| 64 |
},
|
| 65 |
{
|
| 66 |
+
"epoch": 0.13,
|
| 67 |
+
"learning_rate": 1.5625e-05,
|
| 68 |
+
"loss": 0.9281,
|
| 69 |
"step": 100
|
| 70 |
},
|
| 71 |
{
|
| 72 |
+
"epoch": 0.14,
|
| 73 |
+
"learning_rate": 1.71875e-05,
|
| 74 |
+
"loss": 0.9426,
|
| 75 |
"step": 110
|
| 76 |
},
|
| 77 |
{
|
| 78 |
+
"epoch": 0.15,
|
| 79 |
+
"learning_rate": 1.8750000000000002e-05,
|
| 80 |
+
"loss": 0.8554,
|
| 81 |
"step": 120
|
| 82 |
},
|
| 83 |
{
|
| 84 |
+
"epoch": 0.16,
|
| 85 |
+
"learning_rate": 2.0312500000000002e-05,
|
| 86 |
+
"loss": 0.808,
|
| 87 |
"step": 130
|
| 88 |
},
|
| 89 |
{
|
| 90 |
+
"epoch": 0.18,
|
| 91 |
+
"learning_rate": 2.1875e-05,
|
| 92 |
+
"loss": 0.8681,
|
| 93 |
"step": 140
|
| 94 |
},
|
| 95 |
{
|
| 96 |
+
"epoch": 0.19,
|
| 97 |
+
"learning_rate": 2.34375e-05,
|
| 98 |
+
"loss": 0.897,
|
| 99 |
"step": 150
|
| 100 |
},
|
| 101 |
{
|
| 102 |
+
"epoch": 0.2,
|
| 103 |
+
"learning_rate": 2.5e-05,
|
| 104 |
+
"loss": 0.861,
|
| 105 |
"step": 160
|
| 106 |
},
|
| 107 |
{
|
| 108 |
+
"epoch": 0.21,
|
| 109 |
+
"learning_rate": 2.6562500000000002e-05,
|
| 110 |
+
"loss": 0.9079,
|
| 111 |
"step": 170
|
| 112 |
},
|
| 113 |
{
|
| 114 |
+
"epoch": 0.23,
|
| 115 |
+
"learning_rate": 2.8125000000000003e-05,
|
| 116 |
+
"loss": 0.9046,
|
| 117 |
"step": 180
|
| 118 |
},
|
| 119 |
{
|
| 120 |
+
"epoch": 0.24,
|
| 121 |
+
"learning_rate": 2.96875e-05,
|
| 122 |
+
"loss": 0.8711,
|
| 123 |
"step": 190
|
| 124 |
},
|
| 125 |
{
|
| 126 |
+
"epoch": 0.25,
|
| 127 |
+
"learning_rate": 3.125e-05,
|
| 128 |
+
"loss": 0.8316,
|
| 129 |
"step": 200
|
| 130 |
},
|
| 131 |
{
|
| 132 |
+
"epoch": 0.26,
|
| 133 |
+
"learning_rate": 3.2812500000000005e-05,
|
| 134 |
+
"loss": 0.829,
|
| 135 |
"step": 210
|
| 136 |
},
|
| 137 |
{
|
| 138 |
+
"epoch": 0.28,
|
| 139 |
+
"learning_rate": 3.4375e-05,
|
| 140 |
+
"loss": 0.8234,
|
| 141 |
"step": 220
|
| 142 |
},
|
| 143 |
{
|
| 144 |
+
"epoch": 0.29,
|
| 145 |
+
"learning_rate": 3.59375e-05,
|
| 146 |
+
"loss": 0.8658,
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
| 147 |
"step": 230
|
| 148 |
},
|
| 149 |
{
|
| 150 |
+
"epoch": 0.3,
|
| 151 |
+
"learning_rate": 3.7500000000000003e-05,
|
| 152 |
+
"loss": 0.8653,
|
| 153 |
"step": 240
|
| 154 |
},
|
| 155 |
{
|
| 156 |
+
"epoch": 0.31,
|
| 157 |
+
"learning_rate": 3.90625e-05,
|
| 158 |
+
"loss": 0.8014,
|
| 159 |
"step": 250
|
| 160 |
},
|
| 161 |
{
|
| 162 |
+
"epoch": 0.33,
|
| 163 |
+
"learning_rate": 4.0625000000000005e-05,
|
| 164 |
+
"loss": 0.7934,
|
| 165 |
"step": 260
|
| 166 |
},
|
| 167 |
{
|
| 168 |
+
"epoch": 0.34,
|
| 169 |
+
"learning_rate": 4.21875e-05,
|
| 170 |
+
"loss": 0.8023,
|
| 171 |
"step": 270
|
| 172 |
},
|
| 173 |
{
|
| 174 |
+
"epoch": 0.35,
|
| 175 |
+
"learning_rate": 4.375e-05,
|
| 176 |
+
"loss": 0.8294,
|
| 177 |
"step": 280
|
| 178 |
},
|
| 179 |
{
|
| 180 |
+
"epoch": 0.36,
|
| 181 |
+
"learning_rate": 4.5312500000000004e-05,
|
| 182 |
+
"loss": 0.7873,
|
| 183 |
"step": 290
|
| 184 |
},
|
| 185 |
{
|
| 186 |
+
"epoch": 0.38,
|
| 187 |
+
"learning_rate": 4.6875e-05,
|
| 188 |
+
"loss": 0.8071,
|
| 189 |
"step": 300
|
| 190 |
},
|
| 191 |
{
|
| 192 |
+
"epoch": 0.39,
|
| 193 |
+
"learning_rate": 4.8437500000000005e-05,
|
| 194 |
+
"loss": 0.8086,
|
| 195 |
"step": 310
|
| 196 |
},
|
| 197 |
{
|
| 198 |
+
"epoch": 0.4,
|
| 199 |
+
"learning_rate": 5e-05,
|
| 200 |
+
"loss": 0.8296,
|
| 201 |
"step": 320
|
| 202 |
},
|
| 203 |
{
|
| 204 |
+
"epoch": 0.41,
|
| 205 |
+
"learning_rate": 4.9825905292479115e-05,
|
| 206 |
+
"loss": 0.8096,
|
| 207 |
"step": 330
|
| 208 |
},
|
| 209 |
{
|
| 210 |
+
"epoch": 0.43,
|
| 211 |
+
"learning_rate": 4.965181058495822e-05,
|
| 212 |
+
"loss": 0.861,
|
| 213 |
"step": 340
|
| 214 |
},
|
| 215 |
{
|
| 216 |
+
"epoch": 0.44,
|
| 217 |
+
"learning_rate": 4.9477715877437334e-05,
|
| 218 |
+
"loss": 0.8047,
|
| 219 |
"step": 350
|
| 220 |
},
|
| 221 |
{
|
| 222 |
+
"epoch": 0.45,
|
| 223 |
+
"learning_rate": 4.930362116991643e-05,
|
| 224 |
+
"loss": 0.8182,
|
| 225 |
"step": 360
|
| 226 |
},
|
| 227 |
{
|
| 228 |
+
"epoch": 0.46,
|
| 229 |
+
"learning_rate": 4.9129526462395546e-05,
|
| 230 |
+
"loss": 0.8412,
|
| 231 |
"step": 370
|
| 232 |
},
|
| 233 |
{
|
| 234 |
+
"epoch": 0.48,
|
| 235 |
+
"learning_rate": 4.895543175487465e-05,
|
| 236 |
+
"loss": 0.8438,
|
| 237 |
"step": 380
|
| 238 |
},
|
| 239 |
{
|
| 240 |
+
"epoch": 0.49,
|
| 241 |
+
"learning_rate": 4.8781337047353764e-05,
|
| 242 |
+
"loss": 0.7776,
|
| 243 |
"step": 390
|
| 244 |
},
|
| 245 |
{
|
| 246 |
+
"epoch": 0.5,
|
| 247 |
+
"learning_rate": 4.860724233983287e-05,
|
| 248 |
+
"loss": 0.8141,
|
| 249 |
"step": 400
|
| 250 |
},
|
| 251 |
{
|
| 252 |
+
"epoch": 0.51,
|
| 253 |
+
"learning_rate": 4.843314763231198e-05,
|
| 254 |
+
"loss": 0.9028,
|
| 255 |
"step": 410
|
| 256 |
},
|
| 257 |
{
|
| 258 |
+
"epoch": 0.53,
|
| 259 |
+
"learning_rate": 4.825905292479109e-05,
|
| 260 |
+
"loss": 0.8347,
|
| 261 |
"step": 420
|
| 262 |
},
|
| 263 |
{
|
| 264 |
+
"epoch": 0.54,
|
| 265 |
+
"learning_rate": 4.80849582172702e-05,
|
| 266 |
+
"loss": 0.8383,
|
| 267 |
"step": 430
|
| 268 |
},
|
| 269 |
{
|
| 270 |
+
"epoch": 0.55,
|
| 271 |
+
"learning_rate": 4.79108635097493e-05,
|
| 272 |
+
"loss": 0.8052,
|
| 273 |
"step": 440
|
| 274 |
},
|
| 275 |
{
|
| 276 |
+
"epoch": 0.56,
|
| 277 |
+
"learning_rate": 4.7736768802228414e-05,
|
| 278 |
+
"loss": 0.8401,
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
| 279 |
"step": 450
|
| 280 |
},
|
| 281 |
{
|
| 282 |
+
"epoch": 0.58,
|
| 283 |
+
"learning_rate": 4.756267409470752e-05,
|
| 284 |
+
"loss": 0.7737,
|
| 285 |
"step": 460
|
| 286 |
},
|
| 287 |
{
|
| 288 |
+
"epoch": 0.59,
|
| 289 |
+
"learning_rate": 4.738857938718663e-05,
|
| 290 |
+
"loss": 0.7741,
|
| 291 |
"step": 470
|
| 292 |
},
|
| 293 |
{
|
| 294 |
+
"epoch": 0.6,
|
| 295 |
+
"learning_rate": 4.721448467966574e-05,
|
| 296 |
+
"loss": 0.8272,
|
| 297 |
"step": 480
|
| 298 |
},
|
| 299 |
{
|
| 300 |
+
"epoch": 0.61,
|
| 301 |
+
"learning_rate": 4.704038997214485e-05,
|
| 302 |
+
"loss": 0.7843,
|
| 303 |
"step": 490
|
| 304 |
},
|
| 305 |
{
|
| 306 |
+
"epoch": 0.63,
|
| 307 |
+
"learning_rate": 4.686629526462396e-05,
|
| 308 |
+
"loss": 0.8069,
|
| 309 |
"step": 500
|
| 310 |
},
|
| 311 |
{
|
| 312 |
+
"epoch": 0.64,
|
| 313 |
+
"learning_rate": 4.669220055710306e-05,
|
| 314 |
+
"loss": 0.7725,
|
| 315 |
"step": 510
|
| 316 |
},
|
| 317 |
{
|
| 318 |
+
"epoch": 0.65,
|
| 319 |
+
"learning_rate": 4.6518105849582176e-05,
|
| 320 |
+
"loss": 0.8299,
|
| 321 |
"step": 520
|
| 322 |
},
|
| 323 |
{
|
| 324 |
+
"epoch": 0.66,
|
| 325 |
+
"learning_rate": 4.634401114206128e-05,
|
| 326 |
+
"loss": 0.7864,
|
| 327 |
"step": 530
|
| 328 |
},
|
| 329 |
{
|
| 330 |
+
"epoch": 0.68,
|
| 331 |
+
"learning_rate": 4.6169916434540394e-05,
|
| 332 |
+
"loss": 0.8226,
|
| 333 |
"step": 540
|
| 334 |
},
|
| 335 |
{
|
| 336 |
+
"epoch": 0.69,
|
| 337 |
+
"learning_rate": 4.59958217270195e-05,
|
| 338 |
+
"loss": 0.785,
|
| 339 |
"step": 550
|
| 340 |
},
|
| 341 |
{
|
| 342 |
+
"epoch": 0.7,
|
| 343 |
+
"learning_rate": 4.582172701949861e-05,
|
| 344 |
+
"loss": 0.8298,
|
| 345 |
"step": 560
|
| 346 |
},
|
| 347 |
{
|
| 348 |
+
"epoch": 0.71,
|
| 349 |
+
"learning_rate": 4.564763231197772e-05,
|
| 350 |
+
"loss": 0.7736,
|
| 351 |
"step": 570
|
| 352 |
},
|
| 353 |
{
|
| 354 |
+
"epoch": 0.73,
|
| 355 |
+
"learning_rate": 4.547353760445683e-05,
|
| 356 |
+
"loss": 0.8041,
|
| 357 |
"step": 580
|
| 358 |
},
|
| 359 |
{
|
| 360 |
+
"epoch": 0.74,
|
| 361 |
+
"learning_rate": 4.529944289693593e-05,
|
| 362 |
+
"loss": 0.7805,
|
| 363 |
"step": 590
|
| 364 |
},
|
| 365 |
{
|
| 366 |
+
"epoch": 0.75,
|
| 367 |
+
"learning_rate": 4.5125348189415044e-05,
|
| 368 |
+
"loss": 0.8247,
|
| 369 |
"step": 600
|
| 370 |
},
|
| 371 |
{
|
| 372 |
+
"epoch": 0.76,
|
| 373 |
+
"learning_rate": 4.495125348189415e-05,
|
| 374 |
+
"loss": 0.7941,
|
| 375 |
"step": 610
|
| 376 |
},
|
| 377 |
{
|
| 378 |
+
"epoch": 0.78,
|
| 379 |
+
"learning_rate": 4.477715877437326e-05,
|
| 380 |
+
"loss": 0.7644,
|
| 381 |
"step": 620
|
| 382 |
},
|
| 383 |
{
|
| 384 |
+
"epoch": 0.79,
|
| 385 |
+
"learning_rate": 4.460306406685237e-05,
|
| 386 |
+
"loss": 0.77,
|
| 387 |
"step": 630
|
| 388 |
},
|
| 389 |
{
|
| 390 |
+
"epoch": 0.8,
|
| 391 |
+
"learning_rate": 4.442896935933148e-05,
|
| 392 |
+
"loss": 0.778,
|
| 393 |
"step": 640
|
| 394 |
},
|
| 395 |
{
|
| 396 |
+
"epoch": 0.81,
|
| 397 |
+
"learning_rate": 4.425487465181059e-05,
|
| 398 |
+
"loss": 0.8061,
|
| 399 |
"step": 650
|
| 400 |
},
|
| 401 |
{
|
| 402 |
+
"epoch": 0.83,
|
| 403 |
+
"learning_rate": 4.40807799442897e-05,
|
| 404 |
+
"loss": 0.7544,
|
| 405 |
"step": 660
|
| 406 |
},
|
| 407 |
{
|
| 408 |
+
"epoch": 0.84,
|
| 409 |
+
"learning_rate": 4.39066852367688e-05,
|
| 410 |
+
"loss": 0.8129,
|
| 411 |
"step": 670
|
| 412 |
},
|
| 413 |
{
|
| 414 |
+
"epoch": 0.85,
|
| 415 |
+
"learning_rate": 4.373259052924791e-05,
|
| 416 |
+
"loss": 0.7776,
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
| 417 |
"step": 680
|
| 418 |
},
|
| 419 |
{
|
| 420 |
+
"epoch": 0.86,
|
| 421 |
+
"learning_rate": 4.355849582172702e-05,
|
| 422 |
+
"loss": 0.8078,
|
| 423 |
"step": 690
|
| 424 |
},
|
| 425 |
{
|
| 426 |
+
"epoch": 0.88,
|
| 427 |
+
"learning_rate": 4.338440111420613e-05,
|
| 428 |
+
"loss": 0.7845,
|
| 429 |
"step": 700
|
| 430 |
},
|
| 431 |
{
|
| 432 |
+
"epoch": 0.89,
|
| 433 |
+
"learning_rate": 4.321030640668524e-05,
|
| 434 |
+
"loss": 0.7826,
|
| 435 |
"step": 710
|
| 436 |
},
|
| 437 |
{
|
| 438 |
+
"epoch": 0.9,
|
| 439 |
+
"learning_rate": 4.303621169916435e-05,
|
| 440 |
+
"loss": 0.7844,
|
| 441 |
"step": 720
|
| 442 |
},
|
| 443 |
{
|
| 444 |
+
"epoch": 0.91,
|
| 445 |
+
"learning_rate": 4.286211699164346e-05,
|
| 446 |
+
"loss": 0.7905,
|
| 447 |
"step": 730
|
| 448 |
},
|
| 449 |
{
|
| 450 |
+
"epoch": 0.93,
|
| 451 |
+
"learning_rate": 4.268802228412256e-05,
|
| 452 |
+
"loss": 0.7858,
|
| 453 |
"step": 740
|
| 454 |
},
|
| 455 |
{
|
| 456 |
+
"epoch": 0.94,
|
| 457 |
+
"learning_rate": 4.2513927576601673e-05,
|
| 458 |
+
"loss": 0.822,
|
| 459 |
"step": 750
|
| 460 |
},
|
| 461 |
{
|
| 462 |
+
"epoch": 0.95,
|
| 463 |
+
"learning_rate": 4.233983286908078e-05,
|
| 464 |
+
"loss": 0.7246,
|
| 465 |
"step": 760
|
| 466 |
},
|
| 467 |
{
|
| 468 |
+
"epoch": 0.96,
|
| 469 |
+
"learning_rate": 4.216573816155989e-05,
|
| 470 |
+
"loss": 0.797,
|
| 471 |
"step": 770
|
| 472 |
},
|
| 473 |
{
|
| 474 |
+
"epoch": 0.98,
|
| 475 |
+
"learning_rate": 4.1991643454039e-05,
|
| 476 |
+
"loss": 0.7749,
|
| 477 |
"step": 780
|
| 478 |
},
|
| 479 |
{
|
| 480 |
+
"epoch": 0.99,
|
| 481 |
+
"learning_rate": 4.181754874651811e-05,
|
| 482 |
+
"loss": 0.7964,
|
| 483 |
"step": 790
|
| 484 |
},
|
| 485 |
{
|
| 486 |
+
"epoch": 1.0,
|
| 487 |
+
"eval_accuracy": 0.7270911654135338,
|
| 488 |
+
"eval_loss": 0.7869051694869995,
|
| 489 |
+
"eval_runtime": 240.7482,
|
| 490 |
+
"eval_samples_per_second": 106.069,
|
| 491 |
+
"eval_steps_per_second": 3.315,
|
| 492 |
+
"step": 798
|
| 493 |
+
},
|
| 494 |
+
{
|
| 495 |
+
"epoch": 1.0,
|
| 496 |
+
"learning_rate": 4.164345403899722e-05,
|
| 497 |
+
"loss": 0.7452,
|
| 498 |
"step": 800
|
| 499 |
},
|
| 500 |
{
|
| 501 |
+
"epoch": 1.02,
|
| 502 |
+
"learning_rate": 4.146935933147633e-05,
|
| 503 |
+
"loss": 0.6894,
|
| 504 |
"step": 810
|
| 505 |
},
|
| 506 |
{
|
| 507 |
+
"epoch": 1.03,
|
| 508 |
+
"learning_rate": 4.129526462395543e-05,
|
| 509 |
+
"loss": 0.7115,
|
| 510 |
"step": 820
|
| 511 |
},
|
| 512 |
{
|
| 513 |
+
"epoch": 1.04,
|
| 514 |
+
"learning_rate": 4.112116991643454e-05,
|
| 515 |
+
"loss": 0.6841,
|
| 516 |
"step": 830
|
| 517 |
},
|
| 518 |
{
|
| 519 |
+
"epoch": 1.05,
|
| 520 |
+
"learning_rate": 4.094707520891365e-05,
|
| 521 |
+
"loss": 0.6697,
|
| 522 |
"step": 840
|
| 523 |
},
|
| 524 |
{
|
| 525 |
+
"epoch": 1.07,
|
| 526 |
+
"learning_rate": 4.077298050139276e-05,
|
| 527 |
+
"loss": 0.6796,
|
| 528 |
"step": 850
|
| 529 |
},
|
| 530 |
{
|
| 531 |
+
"epoch": 1.08,
|
| 532 |
+
"learning_rate": 4.0598885793871866e-05,
|
| 533 |
+
"loss": 0.7095,
|
| 534 |
"step": 860
|
| 535 |
},
|
| 536 |
{
|
| 537 |
+
"epoch": 1.09,
|
| 538 |
+
"learning_rate": 4.042479108635098e-05,
|
| 539 |
+
"loss": 0.667,
|
| 540 |
"step": 870
|
| 541 |
},
|
| 542 |
{
|
| 543 |
+
"epoch": 1.1,
|
| 544 |
+
"learning_rate": 4.0250696378830085e-05,
|
| 545 |
+
"loss": 0.678,
|
| 546 |
"step": 880
|
| 547 |
},
|
| 548 |
{
|
| 549 |
+
"epoch": 1.12,
|
| 550 |
+
"learning_rate": 4.007660167130919e-05,
|
| 551 |
+
"loss": 0.6771,
|
| 552 |
"step": 890
|
| 553 |
},
|
| 554 |
{
|
| 555 |
+
"epoch": 1.13,
|
| 556 |
+
"learning_rate": 3.9902506963788303e-05,
|
| 557 |
+
"loss": 0.7135,
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
| 558 |
"step": 900
|
| 559 |
},
|
| 560 |
{
|
| 561 |
+
"epoch": 1.14,
|
| 562 |
+
"learning_rate": 3.972841225626741e-05,
|
| 563 |
+
"loss": 0.6982,
|
| 564 |
"step": 910
|
| 565 |
},
|
| 566 |
{
|
| 567 |
+
"epoch": 1.15,
|
| 568 |
+
"learning_rate": 3.955431754874652e-05,
|
| 569 |
+
"loss": 0.7658,
|
| 570 |
"step": 920
|
| 571 |
},
|
| 572 |
{
|
| 573 |
+
"epoch": 1.17,
|
| 574 |
+
"learning_rate": 3.938022284122563e-05,
|
| 575 |
+
"loss": 0.6841,
|
| 576 |
"step": 930
|
| 577 |
},
|
| 578 |
{
|
| 579 |
+
"epoch": 1.18,
|
| 580 |
+
"learning_rate": 3.920612813370474e-05,
|
| 581 |
+
"loss": 0.7158,
|
| 582 |
"step": 940
|
| 583 |
},
|
| 584 |
{
|
| 585 |
+
"epoch": 1.19,
|
| 586 |
+
"learning_rate": 3.903203342618385e-05,
|
| 587 |
+
"loss": 0.7176,
|
| 588 |
"step": 950
|
| 589 |
},
|
| 590 |
{
|
| 591 |
+
"epoch": 1.2,
|
| 592 |
+
"learning_rate": 3.885793871866296e-05,
|
| 593 |
+
"loss": 0.7032,
|
| 594 |
"step": 960
|
| 595 |
},
|
| 596 |
{
|
| 597 |
+
"epoch": 1.22,
|
| 598 |
+
"learning_rate": 3.868384401114206e-05,
|
| 599 |
+
"loss": 0.7131,
|
| 600 |
"step": 970
|
| 601 |
},
|
| 602 |
{
|
| 603 |
+
"epoch": 1.23,
|
| 604 |
+
"learning_rate": 3.850974930362117e-05,
|
| 605 |
+
"loss": 0.7125,
|
| 606 |
"step": 980
|
| 607 |
},
|
| 608 |
{
|
| 609 |
+
"epoch": 1.24,
|
| 610 |
+
"learning_rate": 3.833565459610028e-05,
|
| 611 |
+
"loss": 0.7089,
|
| 612 |
"step": 990
|
| 613 |
},
|
| 614 |
{
|
| 615 |
+
"epoch": 1.25,
|
| 616 |
+
"learning_rate": 3.816155988857939e-05,
|
| 617 |
+
"loss": 0.7246,
|
| 618 |
"step": 1000
|
| 619 |
},
|
| 620 |
{
|
| 621 |
+
"epoch": 1.27,
|
| 622 |
+
"learning_rate": 3.7987465181058496e-05,
|
| 623 |
+
"loss": 0.6763,
|
| 624 |
"step": 1010
|
| 625 |
},
|
| 626 |
{
|
| 627 |
+
"epoch": 1.28,
|
| 628 |
+
"learning_rate": 3.781337047353761e-05,
|
| 629 |
+
"loss": 0.6693,
|
| 630 |
"step": 1020
|
| 631 |
},
|
| 632 |
{
|
| 633 |
+
"epoch": 1.29,
|
| 634 |
+
"learning_rate": 3.7639275766016715e-05,
|
| 635 |
+
"loss": 0.6912,
|
| 636 |
"step": 1030
|
| 637 |
},
|
| 638 |
{
|
| 639 |
+
"epoch": 1.3,
|
| 640 |
+
"learning_rate": 3.746518105849583e-05,
|
| 641 |
+
"loss": 0.6866,
|
| 642 |
"step": 1040
|
| 643 |
},
|
| 644 |
{
|
| 645 |
+
"epoch": 1.32,
|
| 646 |
+
"learning_rate": 3.7291086350974927e-05,
|
| 647 |
+
"loss": 0.6532,
|
| 648 |
"step": 1050
|
| 649 |
},
|
| 650 |
{
|
| 651 |
+
"epoch": 1.33,
|
| 652 |
+
"learning_rate": 3.711699164345404e-05,
|
| 653 |
+
"loss": 0.7283,
|
| 654 |
"step": 1060
|
| 655 |
},
|
| 656 |
{
|
| 657 |
+
"epoch": 1.34,
|
| 658 |
+
"learning_rate": 3.6942896935933145e-05,
|
| 659 |
+
"loss": 0.7032,
|
| 660 |
"step": 1070
|
| 661 |
},
|
| 662 |
{
|
| 663 |
+
"epoch": 1.35,
|
| 664 |
+
"learning_rate": 3.676880222841226e-05,
|
| 665 |
+
"loss": 0.6854,
|
| 666 |
"step": 1080
|
| 667 |
},
|
| 668 |
{
|
| 669 |
+
"epoch": 1.37,
|
| 670 |
+
"learning_rate": 3.659470752089137e-05,
|
| 671 |
+
"loss": 0.6769,
|
| 672 |
"step": 1090
|
| 673 |
},
|
| 674 |
{
|
| 675 |
+
"epoch": 1.38,
|
| 676 |
+
"learning_rate": 3.642061281337048e-05,
|
| 677 |
+
"loss": 0.7349,
|
| 678 |
"step": 1100
|
| 679 |
},
|
| 680 |
{
|
| 681 |
+
"epoch": 1.39,
|
| 682 |
+
"learning_rate": 3.624651810584959e-05,
|
| 683 |
+
"loss": 0.7274,
|
| 684 |
"step": 1110
|
| 685 |
},
|
| 686 |
{
|
| 687 |
+
"epoch": 1.4,
|
| 688 |
+
"learning_rate": 3.607242339832869e-05,
|
| 689 |
+
"loss": 0.6858,
|
| 690 |
"step": 1120
|
| 691 |
},
|
| 692 |
{
|
| 693 |
+
"epoch": 1.42,
|
| 694 |
+
"learning_rate": 3.58983286908078e-05,
|
| 695 |
+
"loss": 0.7229,
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
| 696 |
"step": 1130
|
| 697 |
},
|
| 698 |
{
|
| 699 |
+
"epoch": 1.43,
|
| 700 |
+
"learning_rate": 3.572423398328691e-05,
|
| 701 |
+
"loss": 0.6924,
|
| 702 |
"step": 1140
|
| 703 |
},
|
| 704 |
{
|
| 705 |
+
"epoch": 1.44,
|
| 706 |
+
"learning_rate": 3.555013927576602e-05,
|
| 707 |
+
"loss": 0.686,
|
| 708 |
"step": 1150
|
| 709 |
},
|
| 710 |
{
|
| 711 |
+
"epoch": 1.45,
|
| 712 |
+
"learning_rate": 3.5376044568245126e-05,
|
| 713 |
+
"loss": 0.7268,
|
| 714 |
"step": 1160
|
| 715 |
},
|
| 716 |
{
|
| 717 |
+
"epoch": 1.47,
|
| 718 |
+
"learning_rate": 3.520194986072424e-05,
|
| 719 |
+
"loss": 0.7107,
|
| 720 |
"step": 1170
|
| 721 |
},
|
| 722 |
{
|
| 723 |
+
"epoch": 1.48,
|
| 724 |
+
"learning_rate": 3.5027855153203345e-05,
|
| 725 |
+
"loss": 0.7095,
|
| 726 |
"step": 1180
|
| 727 |
},
|
| 728 |
{
|
| 729 |
+
"epoch": 1.49,
|
| 730 |
+
"learning_rate": 3.485376044568246e-05,
|
| 731 |
+
"loss": 0.6764,
|
| 732 |
"step": 1190
|
| 733 |
},
|
| 734 |
{
|
| 735 |
+
"epoch": 1.5,
|
| 736 |
+
"learning_rate": 3.4679665738161556e-05,
|
| 737 |
+
"loss": 0.7076,
|
| 738 |
"step": 1200
|
| 739 |
},
|
| 740 |
{
|
| 741 |
+
"epoch": 1.52,
|
| 742 |
+
"learning_rate": 3.450557103064067e-05,
|
| 743 |
+
"loss": 0.7194,
|
| 744 |
"step": 1210
|
| 745 |
},
|
| 746 |
{
|
| 747 |
+
"epoch": 1.53,
|
| 748 |
+
"learning_rate": 3.4331476323119775e-05,
|
| 749 |
+
"loss": 0.7164,
|
| 750 |
"step": 1220
|
| 751 |
},
|
| 752 |
{
|
| 753 |
+
"epoch": 1.54,
|
| 754 |
+
"learning_rate": 3.415738161559889e-05,
|
| 755 |
+
"loss": 0.6958,
|
| 756 |
"step": 1230
|
| 757 |
},
|
| 758 |
{
|
| 759 |
+
"epoch": 1.55,
|
| 760 |
+
"learning_rate": 3.3983286908077994e-05,
|
| 761 |
+
"loss": 0.662,
|
| 762 |
"step": 1240
|
| 763 |
},
|
| 764 |
{
|
| 765 |
+
"epoch": 1.57,
|
| 766 |
+
"learning_rate": 3.3809192200557107e-05,
|
| 767 |
+
"loss": 0.6475,
|
| 768 |
"step": 1250
|
| 769 |
},
|
| 770 |
{
|
| 771 |
+
"epoch": 1.58,
|
| 772 |
+
"learning_rate": 3.363509749303621e-05,
|
| 773 |
+
"loss": 0.7169,
|
| 774 |
"step": 1260
|
| 775 |
},
|
| 776 |
{
|
| 777 |
+
"epoch": 1.59,
|
| 778 |
+
"learning_rate": 3.346100278551532e-05,
|
| 779 |
+
"loss": 0.6948,
|
| 780 |
"step": 1270
|
| 781 |
},
|
| 782 |
{
|
| 783 |
+
"epoch": 1.6,
|
| 784 |
+
"learning_rate": 3.328690807799443e-05,
|
| 785 |
+
"loss": 0.7413,
|
| 786 |
"step": 1280
|
| 787 |
},
|
| 788 |
{
|
| 789 |
+
"epoch": 1.62,
|
| 790 |
+
"learning_rate": 3.311281337047354e-05,
|
| 791 |
+
"loss": 0.6824,
|
| 792 |
"step": 1290
|
| 793 |
},
|
| 794 |
{
|
| 795 |
+
"epoch": 1.63,
|
| 796 |
+
"learning_rate": 3.293871866295265e-05,
|
| 797 |
+
"loss": 0.7024,
|
| 798 |
"step": 1300
|
| 799 |
},
|
| 800 |
{
|
| 801 |
+
"epoch": 1.64,
|
| 802 |
+
"learning_rate": 3.2764623955431756e-05,
|
| 803 |
+
"loss": 0.6748,
|
| 804 |
"step": 1310
|
| 805 |
},
|
| 806 |
{
|
| 807 |
+
"epoch": 1.65,
|
| 808 |
+
"learning_rate": 3.259052924791087e-05,
|
| 809 |
+
"loss": 0.7112,
|
| 810 |
"step": 1320
|
| 811 |
},
|
| 812 |
{
|
| 813 |
+
"epoch": 1.67,
|
| 814 |
+
"learning_rate": 3.2416434540389975e-05,
|
| 815 |
+
"loss": 0.6878,
|
| 816 |
"step": 1330
|
| 817 |
},
|
| 818 |
{
|
| 819 |
+
"epoch": 1.68,
|
| 820 |
+
"learning_rate": 3.224233983286909e-05,
|
| 821 |
+
"loss": 0.7114,
|
| 822 |
"step": 1340
|
| 823 |
},
|
| 824 |
{
|
| 825 |
+
"epoch": 1.69,
|
| 826 |
+
"learning_rate": 3.2068245125348186e-05,
|
| 827 |
+
"loss": 0.6924,
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
| 828 |
"step": 1350
|
| 829 |
},
|
| 830 |
{
|
| 831 |
+
"epoch": 1.7,
|
| 832 |
+
"learning_rate": 3.18941504178273e-05,
|
| 833 |
+
"loss": 0.7354,
|
| 834 |
"step": 1360
|
| 835 |
},
|
| 836 |
{
|
| 837 |
+
"epoch": 1.72,
|
| 838 |
+
"learning_rate": 3.1720055710306405e-05,
|
| 839 |
+
"loss": 0.6757,
|
| 840 |
"step": 1370
|
| 841 |
},
|
| 842 |
{
|
| 843 |
+
"epoch": 1.73,
|
| 844 |
+
"learning_rate": 3.154596100278552e-05,
|
| 845 |
+
"loss": 0.6675,
|
| 846 |
"step": 1380
|
| 847 |
},
|
| 848 |
{
|
| 849 |
+
"epoch": 1.74,
|
| 850 |
+
"learning_rate": 3.1371866295264624e-05,
|
| 851 |
+
"loss": 0.6923,
|
| 852 |
"step": 1390
|
| 853 |
},
|
| 854 |
{
|
| 855 |
+
"epoch": 1.75,
|
| 856 |
+
"learning_rate": 3.1197771587743737e-05,
|
| 857 |
+
"loss": 0.665,
|
| 858 |
"step": 1400
|
| 859 |
},
|
| 860 |
{
|
| 861 |
+
"epoch": 1.77,
|
| 862 |
+
"learning_rate": 3.102367688022284e-05,
|
| 863 |
+
"loss": 0.6473,
|
| 864 |
"step": 1410
|
| 865 |
},
|
| 866 |
{
|
| 867 |
+
"epoch": 1.78,
|
| 868 |
+
"learning_rate": 3.0849582172701955e-05,
|
| 869 |
+
"loss": 0.6194,
|
| 870 |
"step": 1420
|
| 871 |
},
|
| 872 |
{
|
| 873 |
+
"epoch": 1.79,
|
| 874 |
+
"learning_rate": 3.0675487465181054e-05,
|
| 875 |
+
"loss": 0.6535,
|
| 876 |
"step": 1430
|
| 877 |
},
|
| 878 |
{
|
| 879 |
+
"epoch": 1.8,
|
| 880 |
+
"learning_rate": 3.050139275766017e-05,
|
| 881 |
+
"loss": 0.6704,
|
| 882 |
"step": 1440
|
| 883 |
},
|
| 884 |
{
|
| 885 |
+
"epoch": 1.82,
|
| 886 |
+
"learning_rate": 3.0327298050139273e-05,
|
| 887 |
+
"loss": 0.713,
|
| 888 |
"step": 1450
|
| 889 |
},
|
| 890 |
{
|
| 891 |
+
"epoch": 1.83,
|
| 892 |
+
"learning_rate": 3.0153203342618386e-05,
|
| 893 |
+
"loss": 0.7435,
|
| 894 |
"step": 1460
|
| 895 |
},
|
| 896 |
{
|
| 897 |
+
"epoch": 1.84,
|
| 898 |
+
"learning_rate": 2.99791086350975e-05,
|
| 899 |
+
"loss": 0.7073,
|
| 900 |
"step": 1470
|
| 901 |
},
|
| 902 |
{
|
| 903 |
+
"epoch": 1.85,
|
| 904 |
+
"learning_rate": 2.98050139275766e-05,
|
| 905 |
+
"loss": 0.6691,
|
| 906 |
"step": 1480
|
| 907 |
},
|
| 908 |
{
|
| 909 |
+
"epoch": 1.87,
|
| 910 |
+
"learning_rate": 2.9630919220055714e-05,
|
| 911 |
+
"loss": 0.6859,
|
| 912 |
"step": 1490
|
| 913 |
},
|
| 914 |
{
|
| 915 |
+
"epoch": 1.88,
|
| 916 |
+
"learning_rate": 2.945682451253482e-05,
|
| 917 |
+
"loss": 0.7689,
|
| 918 |
"step": 1500
|
| 919 |
},
|
| 920 |
{
|
| 921 |
+
"epoch": 1.89,
|
| 922 |
+
"learning_rate": 2.9282729805013932e-05,
|
| 923 |
+
"loss": 0.6777,
|
| 924 |
"step": 1510
|
| 925 |
},
|
| 926 |
{
|
| 927 |
+
"epoch": 1.9,
|
| 928 |
+
"learning_rate": 2.9108635097493035e-05,
|
| 929 |
+
"loss": 0.6909,
|
| 930 |
"step": 1520
|
| 931 |
},
|
| 932 |
{
|
| 933 |
+
"epoch": 1.92,
|
| 934 |
+
"learning_rate": 2.8934540389972148e-05,
|
| 935 |
+
"loss": 0.7016,
|
| 936 |
"step": 1530
|
| 937 |
},
|
| 938 |
{
|
| 939 |
+
"epoch": 1.93,
|
| 940 |
+
"learning_rate": 2.8760445682451254e-05,
|
| 941 |
+
"loss": 0.701,
|
| 942 |
"step": 1540
|
| 943 |
},
|
| 944 |
{
|
| 945 |
+
"epoch": 1.94,
|
| 946 |
+
"learning_rate": 2.8586350974930366e-05,
|
| 947 |
+
"loss": 0.7016,
|
| 948 |
"step": 1550
|
| 949 |
},
|
| 950 |
{
|
| 951 |
+
"epoch": 1.95,
|
| 952 |
+
"learning_rate": 2.841225626740947e-05,
|
| 953 |
+
"loss": 0.6674,
|
| 954 |
"step": 1560
|
| 955 |
},
|
| 956 |
{
|
| 957 |
+
"epoch": 1.97,
|
| 958 |
+
"learning_rate": 2.8238161559888582e-05,
|
| 959 |
+
"loss": 0.6663,
|
| 960 |
"step": 1570
|
| 961 |
},
|
| 962 |
{
|
| 963 |
+
"epoch": 1.98,
|
| 964 |
+
"learning_rate": 2.8064066852367688e-05,
|
| 965 |
+
"loss": 0.6621,
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
| 966 |
"step": 1580
|
| 967 |
},
|
| 968 |
{
|
| 969 |
+
"epoch": 1.99,
|
| 970 |
+
"learning_rate": 2.78899721448468e-05,
|
| 971 |
+
"loss": 0.6567,
|
| 972 |
"step": 1590
|
| 973 |
},
|
| 974 |
{
|
| 975 |
+
"epoch": 2.0,
|
| 976 |
+
"eval_accuracy": 0.738016917293233,
|
| 977 |
+
"eval_loss": 0.7538765668869019,
|
| 978 |
+
"eval_runtime": 243.4161,
|
| 979 |
+
"eval_samples_per_second": 104.907,
|
| 980 |
+
"eval_steps_per_second": 3.278,
|
| 981 |
+
"step": 1596
|
| 982 |
+
},
|
| 983 |
+
{
|
| 984 |
+
"epoch": 2.01,
|
| 985 |
+
"learning_rate": 3.69923416105825e-05,
|
| 986 |
+
"loss": 0.651,
|
| 987 |
"step": 1600
|
| 988 |
},
|
| 989 |
{
|
| 990 |
+
"epoch": 2.02,
|
| 991 |
+
"learning_rate": 3.687630540728708e-05,
|
| 992 |
+
"loss": 0.5874,
|
| 993 |
"step": 1610
|
| 994 |
},
|
| 995 |
{
|
| 996 |
+
"epoch": 2.03,
|
| 997 |
+
"learning_rate": 3.6760269203991646e-05,
|
| 998 |
+
"loss": 0.6426,
|
| 999 |
"step": 1620
|
| 1000 |
},
|
| 1001 |
{
|
| 1002 |
+
"epoch": 2.04,
|
| 1003 |
+
"learning_rate": 3.664423300069622e-05,
|
| 1004 |
+
"loss": 0.6418,
|
| 1005 |
"step": 1630
|
| 1006 |
},
|
| 1007 |
{
|
| 1008 |
+
"epoch": 2.06,
|
| 1009 |
+
"learning_rate": 3.652819679740079e-05,
|
| 1010 |
+
"loss": 0.6023,
|
| 1011 |
"step": 1640
|
| 1012 |
},
|
| 1013 |
{
|
| 1014 |
+
"epoch": 2.07,
|
| 1015 |
+
"learning_rate": 3.641216059410536e-05,
|
| 1016 |
+
"loss": 0.6012,
|
| 1017 |
"step": 1650
|
| 1018 |
},
|
| 1019 |
{
|
| 1020 |
+
"epoch": 2.08,
|
| 1021 |
+
"learning_rate": 3.6296124390809935e-05,
|
| 1022 |
+
"loss": 0.6276,
|
| 1023 |
"step": 1660
|
| 1024 |
},
|
| 1025 |
{
|
| 1026 |
+
"epoch": 2.09,
|
| 1027 |
+
"learning_rate": 3.6180088187514504e-05,
|
| 1028 |
+
"loss": 0.5936,
|
| 1029 |
"step": 1670
|
| 1030 |
},
|
| 1031 |
{
|
| 1032 |
+
"epoch": 2.11,
|
| 1033 |
+
"learning_rate": 3.606405198421908e-05,
|
| 1034 |
+
"loss": 0.6586,
|
| 1035 |
"step": 1680
|
| 1036 |
},
|
| 1037 |
{
|
| 1038 |
+
"epoch": 2.12,
|
| 1039 |
+
"learning_rate": 3.594801578092365e-05,
|
| 1040 |
+
"loss": 0.6888,
|
| 1041 |
"step": 1690
|
| 1042 |
},
|
| 1043 |
{
|
| 1044 |
+
"epoch": 2.13,
|
| 1045 |
+
"learning_rate": 3.5831979577628224e-05,
|
| 1046 |
+
"loss": 0.6365,
|
| 1047 |
"step": 1700
|
| 1048 |
},
|
| 1049 |
{
|
| 1050 |
+
"epoch": 2.14,
|
| 1051 |
+
"learning_rate": 3.571594337433279e-05,
|
| 1052 |
+
"loss": 0.627,
|
| 1053 |
"step": 1710
|
| 1054 |
},
|
| 1055 |
{
|
| 1056 |
+
"epoch": 2.16,
|
| 1057 |
+
"learning_rate": 3.559990717103737e-05,
|
| 1058 |
+
"loss": 0.6088,
|
| 1059 |
"step": 1720
|
| 1060 |
},
|
| 1061 |
{
|
| 1062 |
+
"epoch": 2.17,
|
| 1063 |
+
"learning_rate": 3.548387096774194e-05,
|
| 1064 |
+
"loss": 0.6739,
|
| 1065 |
"step": 1730
|
| 1066 |
},
|
| 1067 |
{
|
| 1068 |
+
"epoch": 2.18,
|
| 1069 |
+
"learning_rate": 3.536783476444651e-05,
|
| 1070 |
+
"loss": 0.6618,
|
| 1071 |
"step": 1740
|
| 1072 |
},
|
| 1073 |
{
|
| 1074 |
+
"epoch": 2.19,
|
| 1075 |
+
"learning_rate": 3.5251798561151075e-05,
|
| 1076 |
+
"loss": 0.5868,
|
| 1077 |
"step": 1750
|
| 1078 |
},
|
| 1079 |
{
|
| 1080 |
+
"epoch": 2.21,
|
| 1081 |
+
"learning_rate": 3.513576235785565e-05,
|
| 1082 |
+
"loss": 0.6162,
|
| 1083 |
"step": 1760
|
| 1084 |
},
|
| 1085 |
{
|
| 1086 |
+
"epoch": 2.22,
|
| 1087 |
+
"learning_rate": 3.501972615456022e-05,
|
| 1088 |
+
"loss": 0.6023,
|
| 1089 |
"step": 1770
|
| 1090 |
},
|
| 1091 |
{
|
| 1092 |
+
"epoch": 2.23,
|
| 1093 |
+
"learning_rate": 3.4903689951264796e-05,
|
| 1094 |
+
"loss": 0.6005,
|
| 1095 |
"step": 1780
|
| 1096 |
},
|
| 1097 |
{
|
| 1098 |
+
"epoch": 2.24,
|
| 1099 |
+
"learning_rate": 3.4787653747969365e-05,
|
| 1100 |
+
"loss": 0.6084,
|
| 1101 |
"step": 1790
|
| 1102 |
},
|
| 1103 |
{
|
| 1104 |
+
"epoch": 2.26,
|
| 1105 |
+
"learning_rate": 3.467161754467394e-05,
|
| 1106 |
+
"loss": 0.5963,
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
| 1107 |
"step": 1800
|
| 1108 |
},
|
| 1109 |
{
|
| 1110 |
+
"epoch": 2.27,
|
| 1111 |
+
"learning_rate": 3.455558134137851e-05,
|
| 1112 |
+
"loss": 0.6816,
|
| 1113 |
"step": 1810
|
| 1114 |
},
|
| 1115 |
{
|
| 1116 |
+
"epoch": 2.28,
|
| 1117 |
+
"learning_rate": 3.4439545138083085e-05,
|
| 1118 |
+
"loss": 0.6564,
|
| 1119 |
"step": 1820
|
| 1120 |
},
|
| 1121 |
{
|
| 1122 |
+
"epoch": 2.29,
|
| 1123 |
+
"learning_rate": 3.4323508934787654e-05,
|
| 1124 |
+
"loss": 0.6332,
|
| 1125 |
"step": 1830
|
| 1126 |
},
|
| 1127 |
{
|
| 1128 |
+
"epoch": 2.31,
|
| 1129 |
+
"learning_rate": 3.420747273149223e-05,
|
| 1130 |
+
"loss": 0.6295,
|
| 1131 |
"step": 1840
|
| 1132 |
},
|
| 1133 |
{
|
| 1134 |
+
"epoch": 2.32,
|
| 1135 |
+
"learning_rate": 3.40914365281968e-05,
|
| 1136 |
+
"loss": 0.6467,
|
| 1137 |
"step": 1850
|
| 1138 |
},
|
| 1139 |
{
|
| 1140 |
+
"epoch": 2.33,
|
| 1141 |
+
"learning_rate": 3.397540032490137e-05,
|
| 1142 |
+
"loss": 0.622,
|
| 1143 |
"step": 1860
|
| 1144 |
},
|
| 1145 |
{
|
| 1146 |
+
"epoch": 2.34,
|
| 1147 |
+
"learning_rate": 3.385936412160594e-05,
|
| 1148 |
+
"loss": 0.662,
|
| 1149 |
"step": 1870
|
| 1150 |
},
|
| 1151 |
{
|
| 1152 |
+
"epoch": 2.36,
|
| 1153 |
+
"learning_rate": 3.374332791831051e-05,
|
| 1154 |
+
"loss": 0.6514,
|
| 1155 |
"step": 1880
|
| 1156 |
},
|
| 1157 |
{
|
| 1158 |
+
"epoch": 2.37,
|
| 1159 |
+
"learning_rate": 3.362729171501509e-05,
|
| 1160 |
+
"loss": 0.6493,
|
| 1161 |
"step": 1890
|
| 1162 |
},
|
| 1163 |
{
|
| 1164 |
+
"epoch": 2.38,
|
| 1165 |
+
"learning_rate": 3.351125551171966e-05,
|
| 1166 |
+
"loss": 0.6167,
|
| 1167 |
"step": 1900
|
| 1168 |
},
|
| 1169 |
{
|
| 1170 |
+
"epoch": 2.39,
|
| 1171 |
+
"learning_rate": 3.339521930842423e-05,
|
| 1172 |
+
"loss": 0.6575,
|
| 1173 |
"step": 1910
|
| 1174 |
},
|
| 1175 |
{
|
| 1176 |
+
"epoch": 2.41,
|
| 1177 |
+
"learning_rate": 3.32791831051288e-05,
|
| 1178 |
+
"loss": 0.5828,
|
| 1179 |
"step": 1920
|
| 1180 |
},
|
| 1181 |
{
|
| 1182 |
+
"epoch": 2.42,
|
| 1183 |
+
"learning_rate": 3.316314690183338e-05,
|
| 1184 |
+
"loss": 0.6323,
|
| 1185 |
"step": 1930
|
| 1186 |
},
|
| 1187 |
{
|
| 1188 |
+
"epoch": 2.43,
|
| 1189 |
+
"learning_rate": 3.3047110698537946e-05,
|
| 1190 |
+
"loss": 0.6285,
|
| 1191 |
"step": 1940
|
| 1192 |
},
|
| 1193 |
{
|
| 1194 |
+
"epoch": 2.44,
|
| 1195 |
+
"learning_rate": 3.293107449524252e-05,
|
| 1196 |
+
"loss": 0.5822,
|
| 1197 |
"step": 1950
|
| 1198 |
},
|
| 1199 |
{
|
| 1200 |
+
"epoch": 2.46,
|
| 1201 |
+
"learning_rate": 3.2815038291947084e-05,
|
| 1202 |
+
"loss": 0.6441,
|
| 1203 |
"step": 1960
|
| 1204 |
},
|
| 1205 |
{
|
| 1206 |
+
"epoch": 2.47,
|
| 1207 |
+
"learning_rate": 3.269900208865166e-05,
|
| 1208 |
+
"loss": 0.6513,
|
| 1209 |
"step": 1970
|
| 1210 |
},
|
| 1211 |
{
|
| 1212 |
+
"epoch": 2.48,
|
| 1213 |
+
"learning_rate": 3.258296588535623e-05,
|
| 1214 |
+
"loss": 0.6371,
|
| 1215 |
"step": 1980
|
| 1216 |
},
|
| 1217 |
{
|
| 1218 |
+
"epoch": 2.49,
|
| 1219 |
+
"learning_rate": 3.2466929682060804e-05,
|
| 1220 |
+
"loss": 0.6149,
|
| 1221 |
"step": 1990
|
| 1222 |
},
|
| 1223 |
{
|
| 1224 |
+
"epoch": 2.51,
|
| 1225 |
+
"learning_rate": 3.235089347876537e-05,
|
| 1226 |
+
"loss": 0.6348,
|
| 1227 |
"step": 2000
|
| 1228 |
},
|
| 1229 |
{
|
| 1230 |
+
"epoch": 2.52,
|
| 1231 |
+
"learning_rate": 3.223485727546995e-05,
|
| 1232 |
+
"loss": 0.6046,
|
| 1233 |
"step": 2010
|
| 1234 |
},
|
| 1235 |
{
|
| 1236 |
+
"epoch": 2.53,
|
| 1237 |
+
"learning_rate": 3.211882107217452e-05,
|
| 1238 |
+
"loss": 0.6119,
|
| 1239 |
"step": 2020
|
| 1240 |
},
|
| 1241 |
{
|
| 1242 |
+
"epoch": 2.54,
|
| 1243 |
+
"learning_rate": 3.200278486887909e-05,
|
| 1244 |
+
"loss": 0.6382,
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
| 1245 |
"step": 2030
|
| 1246 |
},
|
| 1247 |
{
|
| 1248 |
+
"epoch": 2.56,
|
| 1249 |
+
"learning_rate": 3.188674866558366e-05,
|
| 1250 |
+
"loss": 0.6272,
|
| 1251 |
"step": 2040
|
| 1252 |
},
|
| 1253 |
{
|
| 1254 |
+
"epoch": 2.57,
|
| 1255 |
+
"learning_rate": 3.177071246228824e-05,
|
| 1256 |
+
"loss": 0.6405,
|
| 1257 |
"step": 2050
|
| 1258 |
},
|
| 1259 |
{
|
| 1260 |
+
"epoch": 2.58,
|
| 1261 |
+
"learning_rate": 3.165467625899281e-05,
|
| 1262 |
+
"loss": 0.6679,
|
| 1263 |
"step": 2060
|
| 1264 |
},
|
| 1265 |
{
|
| 1266 |
+
"epoch": 2.59,
|
| 1267 |
+
"learning_rate": 3.1538640055697376e-05,
|
| 1268 |
+
"loss": 0.6367,
|
| 1269 |
"step": 2070
|
| 1270 |
},
|
| 1271 |
{
|
| 1272 |
+
"epoch": 2.61,
|
| 1273 |
+
"learning_rate": 3.142260385240195e-05,
|
| 1274 |
+
"loss": 0.6656,
|
| 1275 |
"step": 2080
|
| 1276 |
},
|
| 1277 |
{
|
| 1278 |
+
"epoch": 2.62,
|
| 1279 |
+
"learning_rate": 3.130656764910652e-05,
|
| 1280 |
+
"loss": 0.6427,
|
| 1281 |
"step": 2090
|
| 1282 |
},
|
| 1283 |
{
|
| 1284 |
+
"epoch": 2.63,
|
| 1285 |
+
"learning_rate": 3.1190531445811096e-05,
|
| 1286 |
+
"loss": 0.6316,
|
| 1287 |
"step": 2100
|
| 1288 |
},
|
| 1289 |
{
|
| 1290 |
+
"epoch": 2.64,
|
| 1291 |
+
"learning_rate": 3.1074495242515665e-05,
|
| 1292 |
+
"loss": 0.6047,
|
| 1293 |
"step": 2110
|
| 1294 |
},
|
| 1295 |
{
|
| 1296 |
+
"epoch": 2.66,
|
| 1297 |
+
"learning_rate": 3.095845903922024e-05,
|
| 1298 |
+
"loss": 0.6562,
|
| 1299 |
"step": 2120
|
| 1300 |
},
|
| 1301 |
{
|
| 1302 |
+
"epoch": 2.67,
|
| 1303 |
+
"learning_rate": 3.084242283592481e-05,
|
| 1304 |
+
"loss": 0.6164,
|
| 1305 |
"step": 2130
|
| 1306 |
},
|
| 1307 |
{
|
| 1308 |
+
"epoch": 2.68,
|
| 1309 |
+
"learning_rate": 3.0726386632629385e-05,
|
| 1310 |
+
"loss": 0.638,
|
| 1311 |
"step": 2140
|
| 1312 |
},
|
| 1313 |
{
|
| 1314 |
+
"epoch": 2.69,
|
| 1315 |
+
"learning_rate": 3.0610350429333954e-05,
|
| 1316 |
+
"loss": 0.6587,
|
| 1317 |
"step": 2150
|
| 1318 |
},
|
| 1319 |
{
|
| 1320 |
+
"epoch": 2.71,
|
| 1321 |
+
"learning_rate": 3.0494314226038523e-05,
|
| 1322 |
+
"loss": 0.6784,
|
| 1323 |
"step": 2160
|
| 1324 |
},
|
| 1325 |
{
|
| 1326 |
+
"epoch": 2.72,
|
| 1327 |
+
"learning_rate": 3.0378278022743096e-05,
|
| 1328 |
+
"loss": 0.6727,
|
| 1329 |
"step": 2170
|
| 1330 |
},
|
| 1331 |
{
|
| 1332 |
+
"epoch": 2.73,
|
| 1333 |
+
"learning_rate": 3.0262241819447668e-05,
|
| 1334 |
+
"loss": 0.6163,
|
| 1335 |
"step": 2180
|
| 1336 |
},
|
| 1337 |
{
|
| 1338 |
+
"epoch": 2.74,
|
| 1339 |
+
"learning_rate": 3.014620561615224e-05,
|
| 1340 |
+
"loss": 0.6304,
|
| 1341 |
"step": 2190
|
| 1342 |
},
|
| 1343 |
{
|
| 1344 |
+
"epoch": 2.76,
|
| 1345 |
+
"learning_rate": 3.0030169412856812e-05,
|
| 1346 |
+
"loss": 0.6053,
|
| 1347 |
"step": 2200
|
| 1348 |
},
|
| 1349 |
{
|
| 1350 |
+
"epoch": 2.77,
|
| 1351 |
+
"learning_rate": 2.9914133209561385e-05,
|
| 1352 |
+
"loss": 0.6296,
|
| 1353 |
"step": 2210
|
| 1354 |
},
|
| 1355 |
{
|
| 1356 |
+
"epoch": 2.78,
|
| 1357 |
+
"learning_rate": 2.9798097006265957e-05,
|
| 1358 |
+
"loss": 0.643,
|
| 1359 |
"step": 2220
|
| 1360 |
},
|
| 1361 |
{
|
| 1362 |
+
"epoch": 2.79,
|
| 1363 |
+
"learning_rate": 2.968206080297053e-05,
|
| 1364 |
+
"loss": 0.6033,
|
| 1365 |
"step": 2230
|
| 1366 |
},
|
| 1367 |
{
|
| 1368 |
+
"epoch": 2.81,
|
| 1369 |
+
"learning_rate": 2.95660245996751e-05,
|
| 1370 |
+
"loss": 0.6418,
|
| 1371 |
"step": 2240
|
| 1372 |
},
|
| 1373 |
{
|
| 1374 |
+
"epoch": 2.82,
|
| 1375 |
+
"learning_rate": 2.9449988396379674e-05,
|
| 1376 |
+
"loss": 0.6206,
|
| 1377 |
+
"step": 2250
|
| 1378 |
+
},
|
| 1379 |
+
{
|
| 1380 |
+
"epoch": 2.83,
|
| 1381 |
+
"learning_rate": 2.933395219308424e-05,
|
| 1382 |
+
"loss": 0.6655,
|
| 1383 |
+
"step": 2260
|
| 1384 |
+
},
|
| 1385 |
+
{
|
| 1386 |
+
"epoch": 2.84,
|
| 1387 |
+
"learning_rate": 2.9217915989788812e-05,
|
| 1388 |
+
"loss": 0.5923,
|
| 1389 |
+
"step": 2270
|
| 1390 |
+
},
|
| 1391 |
+
{
|
| 1392 |
+
"epoch": 2.86,
|
| 1393 |
+
"learning_rate": 2.9101879786493384e-05,
|
| 1394 |
+
"loss": 0.6219,
|
| 1395 |
+
"step": 2280
|
| 1396 |
+
},
|
| 1397 |
+
{
|
| 1398 |
+
"epoch": 2.87,
|
| 1399 |
+
"learning_rate": 2.8985843583197956e-05,
|
| 1400 |
+
"loss": 0.6688,
|
| 1401 |
+
"step": 2290
|
| 1402 |
+
},
|
| 1403 |
+
{
|
| 1404 |
+
"epoch": 2.88,
|
| 1405 |
+
"learning_rate": 2.886980737990253e-05,
|
| 1406 |
+
"loss": 0.6476,
|
| 1407 |
+
"step": 2300
|
| 1408 |
+
},
|
| 1409 |
+
{
|
| 1410 |
+
"epoch": 2.89,
|
| 1411 |
+
"learning_rate": 2.87537711766071e-05,
|
| 1412 |
+
"loss": 0.6278,
|
| 1413 |
+
"step": 2310
|
| 1414 |
+
},
|
| 1415 |
+
{
|
| 1416 |
+
"epoch": 2.91,
|
| 1417 |
+
"learning_rate": 2.8637734973311677e-05,
|
| 1418 |
+
"loss": 0.5953,
|
| 1419 |
+
"step": 2320
|
| 1420 |
+
},
|
| 1421 |
+
{
|
| 1422 |
+
"epoch": 2.92,
|
| 1423 |
+
"learning_rate": 2.852169877001625e-05,
|
| 1424 |
+
"loss": 0.6612,
|
| 1425 |
+
"step": 2330
|
| 1426 |
+
},
|
| 1427 |
+
{
|
| 1428 |
+
"epoch": 2.93,
|
| 1429 |
+
"learning_rate": 2.840566256672082e-05,
|
| 1430 |
+
"loss": 0.6813,
|
| 1431 |
+
"step": 2340
|
| 1432 |
+
},
|
| 1433 |
+
{
|
| 1434 |
+
"epoch": 2.94,
|
| 1435 |
+
"learning_rate": 2.8289626363425394e-05,
|
| 1436 |
+
"loss": 0.6529,
|
| 1437 |
+
"step": 2350
|
| 1438 |
+
},
|
| 1439 |
+
{
|
| 1440 |
+
"epoch": 2.96,
|
| 1441 |
+
"learning_rate": 2.8173590160129966e-05,
|
| 1442 |
+
"loss": 0.5897,
|
| 1443 |
+
"step": 2360
|
| 1444 |
+
},
|
| 1445 |
+
{
|
| 1446 |
+
"epoch": 2.97,
|
| 1447 |
+
"learning_rate": 2.805755395683453e-05,
|
| 1448 |
+
"loss": 0.647,
|
| 1449 |
+
"step": 2370
|
| 1450 |
+
},
|
| 1451 |
+
{
|
| 1452 |
+
"epoch": 2.98,
|
| 1453 |
+
"learning_rate": 2.7941517753539104e-05,
|
| 1454 |
+
"loss": 0.6369,
|
| 1455 |
+
"step": 2380
|
| 1456 |
+
},
|
| 1457 |
+
{
|
| 1458 |
+
"epoch": 2.99,
|
| 1459 |
+
"learning_rate": 2.7825481550243676e-05,
|
| 1460 |
+
"loss": 0.6842,
|
| 1461 |
+
"step": 2390
|
| 1462 |
+
},
|
| 1463 |
+
{
|
| 1464 |
+
"epoch": 3.0,
|
| 1465 |
+
"eval_accuracy": 0.7837171052631579,
|
| 1466 |
+
"eval_loss": 0.6286582946777344,
|
| 1467 |
+
"eval_runtime": 244.0222,
|
| 1468 |
+
"eval_samples_per_second": 104.646,
|
| 1469 |
+
"eval_steps_per_second": 3.27,
|
| 1470 |
+
"step": 2394
|
| 1471 |
+
},
|
| 1472 |
+
{
|
| 1473 |
+
"epoch": 3.01,
|
| 1474 |
+
"learning_rate": 2.770944534694825e-05,
|
| 1475 |
+
"loss": 0.5777,
|
| 1476 |
+
"step": 2400
|
| 1477 |
+
},
|
| 1478 |
+
{
|
| 1479 |
+
"epoch": 3.02,
|
| 1480 |
+
"learning_rate": 2.759340914365282e-05,
|
| 1481 |
+
"loss": 0.5252,
|
| 1482 |
+
"step": 2410
|
| 1483 |
+
},
|
| 1484 |
+
{
|
| 1485 |
+
"epoch": 3.03,
|
| 1486 |
+
"learning_rate": 2.7477372940357393e-05,
|
| 1487 |
+
"loss": 0.5019,
|
| 1488 |
+
"step": 2420
|
| 1489 |
+
},
|
| 1490 |
+
{
|
| 1491 |
+
"epoch": 3.05,
|
| 1492 |
+
"learning_rate": 2.7361336737061965e-05,
|
| 1493 |
+
"loss": 0.5629,
|
| 1494 |
+
"step": 2430
|
| 1495 |
+
},
|
| 1496 |
+
{
|
| 1497 |
+
"epoch": 3.06,
|
| 1498 |
+
"learning_rate": 2.7245300533766538e-05,
|
| 1499 |
+
"loss": 0.4852,
|
| 1500 |
+
"step": 2440
|
| 1501 |
+
},
|
| 1502 |
+
{
|
| 1503 |
+
"epoch": 3.07,
|
| 1504 |
+
"learning_rate": 2.712926433047111e-05,
|
| 1505 |
+
"loss": 0.5356,
|
| 1506 |
+
"step": 2450
|
| 1507 |
+
},
|
| 1508 |
+
{
|
| 1509 |
+
"epoch": 3.08,
|
| 1510 |
+
"learning_rate": 2.7013228127175682e-05,
|
| 1511 |
+
"loss": 0.5717,
|
| 1512 |
+
"step": 2460
|
| 1513 |
+
},
|
| 1514 |
+
{
|
| 1515 |
+
"epoch": 3.1,
|
| 1516 |
+
"learning_rate": 2.6897191923880248e-05,
|
| 1517 |
+
"loss": 0.5351,
|
| 1518 |
+
"step": 2470
|
| 1519 |
+
},
|
| 1520 |
+
{
|
| 1521 |
+
"epoch": 3.11,
|
| 1522 |
+
"learning_rate": 2.678115572058482e-05,
|
| 1523 |
+
"loss": 0.4641,
|
| 1524 |
+
"step": 2480
|
| 1525 |
},
|
| 1526 |
{
|
| 1527 |
+
"epoch": 3.12,
|
| 1528 |
+
"learning_rate": 2.6665119517289396e-05,
|
| 1529 |
+
"loss": 0.5437,
|
| 1530 |
+
"step": 2490
|
| 1531 |
+
},
|
| 1532 |
+
{
|
| 1533 |
+
"epoch": 3.13,
|
| 1534 |
+
"learning_rate": 2.6549083313993968e-05,
|
| 1535 |
+
"loss": 0.5385,
|
| 1536 |
+
"step": 2500
|
| 1537 |
+
},
|
| 1538 |
+
{
|
| 1539 |
+
"epoch": 3.15,
|
| 1540 |
+
"learning_rate": 2.643304711069854e-05,
|
| 1541 |
+
"loss": 0.5425,
|
| 1542 |
+
"step": 2510
|
| 1543 |
+
},
|
| 1544 |
+
{
|
| 1545 |
+
"epoch": 3.16,
|
| 1546 |
+
"learning_rate": 2.6317010907403113e-05,
|
| 1547 |
+
"loss": 0.5478,
|
| 1548 |
+
"step": 2520
|
| 1549 |
+
},
|
| 1550 |
+
{
|
| 1551 |
+
"epoch": 3.17,
|
| 1552 |
+
"learning_rate": 2.6200974704107685e-05,
|
| 1553 |
+
"loss": 0.5799,
|
| 1554 |
+
"step": 2530
|
| 1555 |
+
},
|
| 1556 |
+
{
|
| 1557 |
+
"epoch": 3.18,
|
| 1558 |
+
"learning_rate": 2.6084938500812257e-05,
|
| 1559 |
+
"loss": 0.5189,
|
| 1560 |
+
"step": 2540
|
| 1561 |
+
},
|
| 1562 |
+
{
|
| 1563 |
+
"epoch": 3.2,
|
| 1564 |
+
"learning_rate": 2.596890229751683e-05,
|
| 1565 |
+
"loss": 0.5384,
|
| 1566 |
+
"step": 2550
|
| 1567 |
+
},
|
| 1568 |
+
{
|
| 1569 |
+
"epoch": 3.21,
|
| 1570 |
+
"learning_rate": 2.5852866094221402e-05,
|
| 1571 |
+
"loss": 0.5281,
|
| 1572 |
+
"step": 2560
|
| 1573 |
+
},
|
| 1574 |
+
{
|
| 1575 |
+
"epoch": 3.22,
|
| 1576 |
+
"learning_rate": 2.5736829890925968e-05,
|
| 1577 |
+
"loss": 0.4708,
|
| 1578 |
+
"step": 2570
|
| 1579 |
+
},
|
| 1580 |
+
{
|
| 1581 |
+
"epoch": 3.23,
|
| 1582 |
+
"learning_rate": 2.562079368763054e-05,
|
| 1583 |
+
"loss": 0.5123,
|
| 1584 |
+
"step": 2580
|
| 1585 |
+
},
|
| 1586 |
+
{
|
| 1587 |
+
"epoch": 3.25,
|
| 1588 |
+
"learning_rate": 2.5504757484335112e-05,
|
| 1589 |
+
"loss": 0.4915,
|
| 1590 |
+
"step": 2590
|
| 1591 |
+
},
|
| 1592 |
+
{
|
| 1593 |
+
"epoch": 3.26,
|
| 1594 |
+
"learning_rate": 2.5388721281039684e-05,
|
| 1595 |
+
"loss": 0.5612,
|
| 1596 |
+
"step": 2600
|
| 1597 |
+
},
|
| 1598 |
+
{
|
| 1599 |
+
"epoch": 3.27,
|
| 1600 |
+
"learning_rate": 2.5272685077744257e-05,
|
| 1601 |
+
"loss": 0.5616,
|
| 1602 |
+
"step": 2610
|
| 1603 |
+
},
|
| 1604 |
+
{
|
| 1605 |
+
"epoch": 3.28,
|
| 1606 |
+
"learning_rate": 2.515664887444883e-05,
|
| 1607 |
+
"loss": 0.5495,
|
| 1608 |
+
"step": 2620
|
| 1609 |
+
},
|
| 1610 |
+
{
|
| 1611 |
+
"epoch": 3.3,
|
| 1612 |
+
"learning_rate": 2.50406126711534e-05,
|
| 1613 |
+
"loss": 0.5311,
|
| 1614 |
+
"step": 2630
|
| 1615 |
+
},
|
| 1616 |
+
{
|
| 1617 |
+
"epoch": 3.31,
|
| 1618 |
+
"learning_rate": 2.4924576467857974e-05,
|
| 1619 |
+
"loss": 0.4424,
|
| 1620 |
+
"step": 2640
|
| 1621 |
+
},
|
| 1622 |
+
{
|
| 1623 |
+
"epoch": 3.32,
|
| 1624 |
+
"learning_rate": 2.4808540264562543e-05,
|
| 1625 |
+
"loss": 0.5303,
|
| 1626 |
+
"step": 2650
|
| 1627 |
+
},
|
| 1628 |
+
{
|
| 1629 |
+
"epoch": 3.33,
|
| 1630 |
+
"learning_rate": 2.4692504061267115e-05,
|
| 1631 |
+
"loss": 0.5391,
|
| 1632 |
+
"step": 2660
|
| 1633 |
+
},
|
| 1634 |
+
{
|
| 1635 |
+
"epoch": 3.35,
|
| 1636 |
+
"learning_rate": 2.4576467857971687e-05,
|
| 1637 |
+
"loss": 0.4972,
|
| 1638 |
+
"step": 2670
|
| 1639 |
+
},
|
| 1640 |
+
{
|
| 1641 |
+
"epoch": 3.36,
|
| 1642 |
+
"learning_rate": 2.446043165467626e-05,
|
| 1643 |
+
"loss": 0.5716,
|
| 1644 |
+
"step": 2680
|
| 1645 |
+
},
|
| 1646 |
+
{
|
| 1647 |
+
"epoch": 3.37,
|
| 1648 |
+
"learning_rate": 2.4344395451380832e-05,
|
| 1649 |
+
"loss": 0.4912,
|
| 1650 |
+
"step": 2690
|
| 1651 |
+
},
|
| 1652 |
+
{
|
| 1653 |
+
"epoch": 3.38,
|
| 1654 |
+
"learning_rate": 2.4228359248085404e-05,
|
| 1655 |
+
"loss": 0.5043,
|
| 1656 |
+
"step": 2700
|
| 1657 |
+
},
|
| 1658 |
+
{
|
| 1659 |
+
"epoch": 3.4,
|
| 1660 |
+
"learning_rate": 2.4112323044789976e-05,
|
| 1661 |
+
"loss": 0.5148,
|
| 1662 |
+
"step": 2710
|
| 1663 |
+
},
|
| 1664 |
+
{
|
| 1665 |
+
"epoch": 3.41,
|
| 1666 |
+
"learning_rate": 2.399628684149455e-05,
|
| 1667 |
+
"loss": 0.5139,
|
| 1668 |
+
"step": 2720
|
| 1669 |
+
},
|
| 1670 |
+
{
|
| 1671 |
+
"epoch": 3.42,
|
| 1672 |
+
"learning_rate": 2.388025063819912e-05,
|
| 1673 |
+
"loss": 0.514,
|
| 1674 |
+
"step": 2730
|
| 1675 |
+
},
|
| 1676 |
+
{
|
| 1677 |
+
"epoch": 3.43,
|
| 1678 |
+
"learning_rate": 2.3764214434903693e-05,
|
| 1679 |
+
"loss": 0.5302,
|
| 1680 |
+
"step": 2740
|
| 1681 |
+
},
|
| 1682 |
+
{
|
| 1683 |
+
"epoch": 3.45,
|
| 1684 |
+
"learning_rate": 2.3648178231608262e-05,
|
| 1685 |
+
"loss": 0.5554,
|
| 1686 |
+
"step": 2750
|
| 1687 |
+
},
|
| 1688 |
+
{
|
| 1689 |
+
"epoch": 3.46,
|
| 1690 |
+
"learning_rate": 2.3532142028312835e-05,
|
| 1691 |
+
"loss": 0.5057,
|
| 1692 |
+
"step": 2760
|
| 1693 |
+
},
|
| 1694 |
+
{
|
| 1695 |
+
"epoch": 3.47,
|
| 1696 |
+
"learning_rate": 2.3416105825017407e-05,
|
| 1697 |
+
"loss": 0.5366,
|
| 1698 |
+
"step": 2770
|
| 1699 |
+
},
|
| 1700 |
+
{
|
| 1701 |
+
"epoch": 3.48,
|
| 1702 |
+
"learning_rate": 2.330006962172198e-05,
|
| 1703 |
+
"loss": 0.5234,
|
| 1704 |
+
"step": 2780
|
| 1705 |
+
},
|
| 1706 |
+
{
|
| 1707 |
+
"epoch": 3.5,
|
| 1708 |
+
"learning_rate": 2.318403341842655e-05,
|
| 1709 |
+
"loss": 0.5847,
|
| 1710 |
+
"step": 2790
|
| 1711 |
+
},
|
| 1712 |
+
{
|
| 1713 |
+
"epoch": 3.51,
|
| 1714 |
+
"learning_rate": 2.306799721513112e-05,
|
| 1715 |
+
"loss": 0.5515,
|
| 1716 |
+
"step": 2800
|
| 1717 |
+
},
|
| 1718 |
+
{
|
| 1719 |
+
"epoch": 3.52,
|
| 1720 |
+
"learning_rate": 2.2951961011835693e-05,
|
| 1721 |
+
"loss": 0.526,
|
| 1722 |
+
"step": 2810
|
| 1723 |
+
},
|
| 1724 |
+
{
|
| 1725 |
+
"epoch": 3.53,
|
| 1726 |
+
"learning_rate": 2.2835924808540265e-05,
|
| 1727 |
+
"loss": 0.5571,
|
| 1728 |
+
"step": 2820
|
| 1729 |
+
},
|
| 1730 |
+
{
|
| 1731 |
+
"epoch": 3.55,
|
| 1732 |
+
"learning_rate": 2.2719888605244837e-05,
|
| 1733 |
+
"loss": 0.5698,
|
| 1734 |
+
"step": 2830
|
| 1735 |
+
},
|
| 1736 |
+
{
|
| 1737 |
+
"epoch": 3.56,
|
| 1738 |
+
"learning_rate": 2.260385240194941e-05,
|
| 1739 |
+
"loss": 0.5039,
|
| 1740 |
+
"step": 2840
|
| 1741 |
+
},
|
| 1742 |
+
{
|
| 1743 |
+
"epoch": 3.57,
|
| 1744 |
+
"learning_rate": 2.248781619865398e-05,
|
| 1745 |
+
"loss": 0.5164,
|
| 1746 |
+
"step": 2850
|
| 1747 |
+
},
|
| 1748 |
+
{
|
| 1749 |
+
"epoch": 3.58,
|
| 1750 |
+
"learning_rate": 2.237177999535855e-05,
|
| 1751 |
+
"loss": 0.5741,
|
| 1752 |
+
"step": 2860
|
| 1753 |
+
},
|
| 1754 |
+
{
|
| 1755 |
+
"epoch": 3.6,
|
| 1756 |
+
"learning_rate": 2.2255743792063123e-05,
|
| 1757 |
+
"loss": 0.5243,
|
| 1758 |
+
"step": 2870
|
| 1759 |
+
},
|
| 1760 |
+
{
|
| 1761 |
+
"epoch": 3.61,
|
| 1762 |
+
"learning_rate": 2.2139707588767696e-05,
|
| 1763 |
+
"loss": 0.5214,
|
| 1764 |
+
"step": 2880
|
| 1765 |
+
},
|
| 1766 |
+
{
|
| 1767 |
+
"epoch": 3.62,
|
| 1768 |
+
"learning_rate": 2.2023671385472268e-05,
|
| 1769 |
+
"loss": 0.5029,
|
| 1770 |
+
"step": 2890
|
| 1771 |
+
},
|
| 1772 |
+
{
|
| 1773 |
+
"epoch": 3.63,
|
| 1774 |
+
"learning_rate": 2.190763518217684e-05,
|
| 1775 |
+
"loss": 0.5528,
|
| 1776 |
+
"step": 2900
|
| 1777 |
+
},
|
| 1778 |
+
{
|
| 1779 |
+
"epoch": 3.65,
|
| 1780 |
+
"learning_rate": 2.1791598978881413e-05,
|
| 1781 |
+
"loss": 0.5115,
|
| 1782 |
+
"step": 2910
|
| 1783 |
+
},
|
| 1784 |
+
{
|
| 1785 |
+
"epoch": 3.66,
|
| 1786 |
+
"learning_rate": 2.1675562775585985e-05,
|
| 1787 |
+
"loss": 0.5399,
|
| 1788 |
+
"step": 2920
|
| 1789 |
+
},
|
| 1790 |
+
{
|
| 1791 |
+
"epoch": 3.67,
|
| 1792 |
+
"learning_rate": 2.1559526572290557e-05,
|
| 1793 |
+
"loss": 0.5467,
|
| 1794 |
+
"step": 2930
|
| 1795 |
+
},
|
| 1796 |
+
{
|
| 1797 |
+
"epoch": 3.68,
|
| 1798 |
+
"learning_rate": 2.144349036899513e-05,
|
| 1799 |
+
"loss": 0.5583,
|
| 1800 |
+
"step": 2940
|
| 1801 |
+
},
|
| 1802 |
+
{
|
| 1803 |
+
"epoch": 3.7,
|
| 1804 |
+
"learning_rate": 2.1327454165699702e-05,
|
| 1805 |
+
"loss": 0.5826,
|
| 1806 |
+
"step": 2950
|
| 1807 |
+
},
|
| 1808 |
+
{
|
| 1809 |
+
"epoch": 3.71,
|
| 1810 |
+
"learning_rate": 2.121141796240427e-05,
|
| 1811 |
+
"loss": 0.5042,
|
| 1812 |
+
"step": 2960
|
| 1813 |
+
},
|
| 1814 |
+
{
|
| 1815 |
+
"epoch": 3.72,
|
| 1816 |
+
"learning_rate": 2.1095381759108843e-05,
|
| 1817 |
+
"loss": 0.4947,
|
| 1818 |
+
"step": 2970
|
| 1819 |
+
},
|
| 1820 |
+
{
|
| 1821 |
+
"epoch": 3.73,
|
| 1822 |
+
"learning_rate": 2.0979345555813415e-05,
|
| 1823 |
+
"loss": 0.5185,
|
| 1824 |
+
"step": 2980
|
| 1825 |
+
},
|
| 1826 |
+
{
|
| 1827 |
+
"epoch": 3.75,
|
| 1828 |
+
"learning_rate": 2.0863309352517988e-05,
|
| 1829 |
+
"loss": 0.539,
|
| 1830 |
+
"step": 2990
|
| 1831 |
+
},
|
| 1832 |
+
{
|
| 1833 |
+
"epoch": 3.76,
|
| 1834 |
+
"learning_rate": 2.074727314922256e-05,
|
| 1835 |
+
"loss": 0.5194,
|
| 1836 |
+
"step": 3000
|
| 1837 |
+
},
|
| 1838 |
+
{
|
| 1839 |
+
"epoch": 3.77,
|
| 1840 |
+
"learning_rate": 2.063123694592713e-05,
|
| 1841 |
+
"loss": 0.5049,
|
| 1842 |
+
"step": 3010
|
| 1843 |
+
},
|
| 1844 |
+
{
|
| 1845 |
+
"epoch": 3.78,
|
| 1846 |
+
"learning_rate": 2.05152007426317e-05,
|
| 1847 |
+
"loss": 0.4813,
|
| 1848 |
+
"step": 3020
|
| 1849 |
+
},
|
| 1850 |
+
{
|
| 1851 |
+
"epoch": 3.8,
|
| 1852 |
+
"learning_rate": 2.0399164539336273e-05,
|
| 1853 |
+
"loss": 0.4671,
|
| 1854 |
+
"step": 3030
|
| 1855 |
+
},
|
| 1856 |
+
{
|
| 1857 |
+
"epoch": 3.81,
|
| 1858 |
+
"learning_rate": 2.0283128336040846e-05,
|
| 1859 |
+
"loss": 0.5214,
|
| 1860 |
+
"step": 3040
|
| 1861 |
+
},
|
| 1862 |
+
{
|
| 1863 |
+
"epoch": 3.82,
|
| 1864 |
+
"learning_rate": 2.0167092132745418e-05,
|
| 1865 |
+
"loss": 0.5476,
|
| 1866 |
+
"step": 3050
|
| 1867 |
+
},
|
| 1868 |
+
{
|
| 1869 |
+
"epoch": 3.83,
|
| 1870 |
+
"learning_rate": 2.0051055929449987e-05,
|
| 1871 |
+
"loss": 0.5231,
|
| 1872 |
+
"step": 3060
|
| 1873 |
+
},
|
| 1874 |
+
{
|
| 1875 |
+
"epoch": 3.85,
|
| 1876 |
+
"learning_rate": 1.993501972615456e-05,
|
| 1877 |
+
"loss": 0.5671,
|
| 1878 |
+
"step": 3070
|
| 1879 |
+
},
|
| 1880 |
+
{
|
| 1881 |
+
"epoch": 3.86,
|
| 1882 |
+
"learning_rate": 1.981898352285913e-05,
|
| 1883 |
+
"loss": 0.5647,
|
| 1884 |
+
"step": 3080
|
| 1885 |
+
},
|
| 1886 |
+
{
|
| 1887 |
+
"epoch": 3.87,
|
| 1888 |
+
"learning_rate": 1.9702947319563704e-05,
|
| 1889 |
+
"loss": 0.5219,
|
| 1890 |
+
"step": 3090
|
| 1891 |
+
},
|
| 1892 |
+
{
|
| 1893 |
+
"epoch": 3.88,
|
| 1894 |
+
"learning_rate": 1.9586911116268276e-05,
|
| 1895 |
+
"loss": 0.5142,
|
| 1896 |
+
"step": 3100
|
| 1897 |
+
},
|
| 1898 |
+
{
|
| 1899 |
+
"epoch": 3.9,
|
| 1900 |
+
"learning_rate": 1.947087491297285e-05,
|
| 1901 |
+
"loss": 0.5069,
|
| 1902 |
+
"step": 3110
|
| 1903 |
+
},
|
| 1904 |
+
{
|
| 1905 |
+
"epoch": 3.91,
|
| 1906 |
+
"learning_rate": 1.935483870967742e-05,
|
| 1907 |
+
"loss": 0.5581,
|
| 1908 |
+
"step": 3120
|
| 1909 |
+
},
|
| 1910 |
+
{
|
| 1911 |
+
"epoch": 3.92,
|
| 1912 |
+
"learning_rate": 1.9238802506381993e-05,
|
| 1913 |
+
"loss": 0.4728,
|
| 1914 |
+
"step": 3130
|
| 1915 |
+
},
|
| 1916 |
+
{
|
| 1917 |
+
"epoch": 3.93,
|
| 1918 |
+
"learning_rate": 1.9122766303086565e-05,
|
| 1919 |
+
"loss": 0.5513,
|
| 1920 |
+
"step": 3140
|
| 1921 |
+
},
|
| 1922 |
+
{
|
| 1923 |
+
"epoch": 3.95,
|
| 1924 |
+
"learning_rate": 1.9006730099791138e-05,
|
| 1925 |
+
"loss": 0.5409,
|
| 1926 |
+
"step": 3150
|
| 1927 |
+
},
|
| 1928 |
+
{
|
| 1929 |
+
"epoch": 3.96,
|
| 1930 |
+
"learning_rate": 1.8890693896495707e-05,
|
| 1931 |
+
"loss": 0.5164,
|
| 1932 |
+
"step": 3160
|
| 1933 |
+
},
|
| 1934 |
+
{
|
| 1935 |
+
"epoch": 3.97,
|
| 1936 |
+
"learning_rate": 1.877465769320028e-05,
|
| 1937 |
+
"loss": 0.5415,
|
| 1938 |
+
"step": 3170
|
| 1939 |
+
},
|
| 1940 |
+
{
|
| 1941 |
+
"epoch": 3.98,
|
| 1942 |
+
"learning_rate": 1.865862148990485e-05,
|
| 1943 |
+
"loss": 0.52,
|
| 1944 |
+
"step": 3180
|
| 1945 |
+
},
|
| 1946 |
+
{
|
| 1947 |
+
"epoch": 4.0,
|
| 1948 |
+
"learning_rate": 1.8542585286609424e-05,
|
| 1949 |
+
"loss": 0.5242,
|
| 1950 |
+
"step": 3190
|
| 1951 |
+
},
|
| 1952 |
+
{
|
| 1953 |
+
"epoch": 4.0,
|
| 1954 |
+
"eval_accuracy": 0.7839129072681704,
|
| 1955 |
+
"eval_loss": 0.62819904088974,
|
| 1956 |
+
"eval_runtime": 245.126,
|
| 1957 |
+
"eval_samples_per_second": 104.175,
|
| 1958 |
+
"eval_steps_per_second": 3.255,
|
| 1959 |
+
"step": 3192
|
| 1960 |
+
},
|
| 1961 |
+
{
|
| 1962 |
+
"epoch": 4.01,
|
| 1963 |
+
"learning_rate": 1.8426549083313996e-05,
|
| 1964 |
+
"loss": 0.45,
|
| 1965 |
+
"step": 3200
|
| 1966 |
+
},
|
| 1967 |
+
{
|
| 1968 |
+
"epoch": 4.02,
|
| 1969 |
+
"learning_rate": 1.8310512880018565e-05,
|
| 1970 |
+
"loss": 0.4147,
|
| 1971 |
+
"step": 3210
|
| 1972 |
+
},
|
| 1973 |
+
{
|
| 1974 |
+
"epoch": 4.04,
|
| 1975 |
+
"learning_rate": 1.8194476676723137e-05,
|
| 1976 |
+
"loss": 0.415,
|
| 1977 |
+
"step": 3220
|
| 1978 |
+
},
|
| 1979 |
+
{
|
| 1980 |
+
"epoch": 4.05,
|
| 1981 |
+
"learning_rate": 1.807844047342771e-05,
|
| 1982 |
+
"loss": 0.4206,
|
| 1983 |
+
"step": 3230
|
| 1984 |
+
},
|
| 1985 |
+
{
|
| 1986 |
+
"epoch": 4.06,
|
| 1987 |
+
"learning_rate": 1.7962404270132282e-05,
|
| 1988 |
+
"loss": 0.4103,
|
| 1989 |
+
"step": 3240
|
| 1990 |
+
},
|
| 1991 |
+
{
|
| 1992 |
+
"epoch": 4.07,
|
| 1993 |
+
"learning_rate": 1.7846368066836854e-05,
|
| 1994 |
+
"loss": 0.4269,
|
| 1995 |
+
"step": 3250
|
| 1996 |
+
},
|
| 1997 |
+
{
|
| 1998 |
+
"epoch": 4.09,
|
| 1999 |
+
"learning_rate": 1.7730331863541423e-05,
|
| 2000 |
+
"loss": 0.4498,
|
| 2001 |
+
"step": 3260
|
| 2002 |
+
},
|
| 2003 |
+
{
|
| 2004 |
+
"epoch": 4.1,
|
| 2005 |
+
"learning_rate": 1.7614295660245995e-05,
|
| 2006 |
+
"loss": 0.4054,
|
| 2007 |
+
"step": 3270
|
| 2008 |
+
},
|
| 2009 |
+
{
|
| 2010 |
+
"epoch": 4.11,
|
| 2011 |
+
"learning_rate": 1.7498259456950568e-05,
|
| 2012 |
+
"loss": 0.4487,
|
| 2013 |
+
"step": 3280
|
| 2014 |
+
},
|
| 2015 |
+
{
|
| 2016 |
+
"epoch": 4.12,
|
| 2017 |
+
"learning_rate": 1.738222325365514e-05,
|
| 2018 |
+
"loss": 0.4052,
|
| 2019 |
+
"step": 3290
|
| 2020 |
+
},
|
| 2021 |
+
{
|
| 2022 |
+
"epoch": 4.14,
|
| 2023 |
+
"learning_rate": 1.7266187050359716e-05,
|
| 2024 |
+
"loss": 0.4106,
|
| 2025 |
+
"step": 3300
|
| 2026 |
+
},
|
| 2027 |
+
{
|
| 2028 |
+
"epoch": 4.15,
|
| 2029 |
+
"learning_rate": 1.7150150847064285e-05,
|
| 2030 |
+
"loss": 0.4004,
|
| 2031 |
+
"step": 3310
|
| 2032 |
+
},
|
| 2033 |
+
{
|
| 2034 |
+
"epoch": 4.16,
|
| 2035 |
+
"learning_rate": 1.7034114643768857e-05,
|
| 2036 |
+
"loss": 0.433,
|
| 2037 |
+
"step": 3320
|
| 2038 |
+
},
|
| 2039 |
+
{
|
| 2040 |
+
"epoch": 4.17,
|
| 2041 |
+
"learning_rate": 1.691807844047343e-05,
|
| 2042 |
+
"loss": 0.3979,
|
| 2043 |
+
"step": 3330
|
| 2044 |
+
},
|
| 2045 |
+
{
|
| 2046 |
+
"epoch": 4.19,
|
| 2047 |
+
"learning_rate": 1.6802042237178e-05,
|
| 2048 |
+
"loss": 0.4445,
|
| 2049 |
+
"step": 3340
|
| 2050 |
+
},
|
| 2051 |
+
{
|
| 2052 |
+
"epoch": 4.2,
|
| 2053 |
+
"learning_rate": 1.6686006033882574e-05,
|
| 2054 |
+
"loss": 0.4357,
|
| 2055 |
+
"step": 3350
|
| 2056 |
+
},
|
| 2057 |
+
{
|
| 2058 |
+
"epoch": 4.21,
|
| 2059 |
+
"learning_rate": 1.6569969830587146e-05,
|
| 2060 |
+
"loss": 0.4279,
|
| 2061 |
+
"step": 3360
|
| 2062 |
+
},
|
| 2063 |
+
{
|
| 2064 |
+
"epoch": 4.22,
|
| 2065 |
+
"learning_rate": 1.6453933627291715e-05,
|
| 2066 |
+
"loss": 0.4089,
|
| 2067 |
+
"step": 3370
|
| 2068 |
+
},
|
| 2069 |
+
{
|
| 2070 |
+
"epoch": 4.24,
|
| 2071 |
+
"learning_rate": 1.6337897423996287e-05,
|
| 2072 |
+
"loss": 0.4039,
|
| 2073 |
+
"step": 3380
|
| 2074 |
+
},
|
| 2075 |
+
{
|
| 2076 |
+
"epoch": 4.25,
|
| 2077 |
+
"learning_rate": 1.622186122070086e-05,
|
| 2078 |
+
"loss": 0.4216,
|
| 2079 |
+
"step": 3390
|
| 2080 |
+
},
|
| 2081 |
+
{
|
| 2082 |
+
"epoch": 4.26,
|
| 2083 |
+
"learning_rate": 1.6105825017405432e-05,
|
| 2084 |
+
"loss": 0.4535,
|
| 2085 |
+
"step": 3400
|
| 2086 |
+
},
|
| 2087 |
+
{
|
| 2088 |
+
"epoch": 4.27,
|
| 2089 |
+
"learning_rate": 1.5989788814110004e-05,
|
| 2090 |
+
"loss": 0.4646,
|
| 2091 |
+
"step": 3410
|
| 2092 |
+
},
|
| 2093 |
+
{
|
| 2094 |
+
"epoch": 4.29,
|
| 2095 |
+
"learning_rate": 1.5873752610814573e-05,
|
| 2096 |
+
"loss": 0.4324,
|
| 2097 |
+
"step": 3420
|
| 2098 |
+
},
|
| 2099 |
+
{
|
| 2100 |
+
"epoch": 4.3,
|
| 2101 |
+
"learning_rate": 1.5757716407519145e-05,
|
| 2102 |
+
"loss": 0.4402,
|
| 2103 |
+
"step": 3430
|
| 2104 |
+
},
|
| 2105 |
+
{
|
| 2106 |
+
"epoch": 4.31,
|
| 2107 |
+
"learning_rate": 1.5641680204223718e-05,
|
| 2108 |
+
"loss": 0.4123,
|
| 2109 |
+
"step": 3440
|
| 2110 |
+
},
|
| 2111 |
+
{
|
| 2112 |
+
"epoch": 4.32,
|
| 2113 |
+
"learning_rate": 1.552564400092829e-05,
|
| 2114 |
+
"loss": 0.4742,
|
| 2115 |
+
"step": 3450
|
| 2116 |
+
},
|
| 2117 |
+
{
|
| 2118 |
+
"epoch": 4.34,
|
| 2119 |
+
"learning_rate": 1.5409607797632862e-05,
|
| 2120 |
+
"loss": 0.4177,
|
| 2121 |
+
"step": 3460
|
| 2122 |
+
},
|
| 2123 |
+
{
|
| 2124 |
+
"epoch": 4.35,
|
| 2125 |
+
"learning_rate": 1.5293571594337435e-05,
|
| 2126 |
+
"loss": 0.4215,
|
| 2127 |
+
"step": 3470
|
| 2128 |
+
},
|
| 2129 |
+
{
|
| 2130 |
+
"epoch": 4.36,
|
| 2131 |
+
"learning_rate": 1.5177535391042005e-05,
|
| 2132 |
+
"loss": 0.3837,
|
| 2133 |
+
"step": 3480
|
| 2134 |
+
},
|
| 2135 |
+
{
|
| 2136 |
+
"epoch": 4.37,
|
| 2137 |
+
"learning_rate": 1.5061499187746578e-05,
|
| 2138 |
+
"loss": 0.4275,
|
| 2139 |
+
"step": 3490
|
| 2140 |
+
},
|
| 2141 |
+
{
|
| 2142 |
+
"epoch": 4.39,
|
| 2143 |
+
"learning_rate": 1.494546298445115e-05,
|
| 2144 |
+
"loss": 0.4526,
|
| 2145 |
+
"step": 3500
|
| 2146 |
+
},
|
| 2147 |
+
{
|
| 2148 |
+
"epoch": 4.4,
|
| 2149 |
+
"learning_rate": 1.4829426781155722e-05,
|
| 2150 |
+
"loss": 0.3635,
|
| 2151 |
+
"step": 3510
|
| 2152 |
+
},
|
| 2153 |
+
{
|
| 2154 |
+
"epoch": 4.41,
|
| 2155 |
+
"learning_rate": 1.4713390577860293e-05,
|
| 2156 |
+
"loss": 0.4344,
|
| 2157 |
+
"step": 3520
|
| 2158 |
+
},
|
| 2159 |
+
{
|
| 2160 |
+
"epoch": 4.42,
|
| 2161 |
+
"learning_rate": 1.4597354374564865e-05,
|
| 2162 |
+
"loss": 0.4282,
|
| 2163 |
+
"step": 3530
|
| 2164 |
+
},
|
| 2165 |
+
{
|
| 2166 |
+
"epoch": 4.44,
|
| 2167 |
+
"learning_rate": 1.4481318171269437e-05,
|
| 2168 |
+
"loss": 0.4838,
|
| 2169 |
+
"step": 3540
|
| 2170 |
+
},
|
| 2171 |
+
{
|
| 2172 |
+
"epoch": 4.45,
|
| 2173 |
+
"learning_rate": 1.436528196797401e-05,
|
| 2174 |
+
"loss": 0.3926,
|
| 2175 |
+
"step": 3550
|
| 2176 |
+
},
|
| 2177 |
+
{
|
| 2178 |
+
"epoch": 4.46,
|
| 2179 |
+
"learning_rate": 1.4249245764678582e-05,
|
| 2180 |
+
"loss": 0.4181,
|
| 2181 |
+
"step": 3560
|
| 2182 |
+
},
|
| 2183 |
+
{
|
| 2184 |
+
"epoch": 4.47,
|
| 2185 |
+
"learning_rate": 1.4133209561383151e-05,
|
| 2186 |
+
"loss": 0.4614,
|
| 2187 |
+
"step": 3570
|
| 2188 |
+
},
|
| 2189 |
+
{
|
| 2190 |
+
"epoch": 4.49,
|
| 2191 |
+
"learning_rate": 1.4017173358087723e-05,
|
| 2192 |
+
"loss": 0.4338,
|
| 2193 |
+
"step": 3580
|
| 2194 |
+
},
|
| 2195 |
+
{
|
| 2196 |
+
"epoch": 4.5,
|
| 2197 |
+
"learning_rate": 1.3901137154792296e-05,
|
| 2198 |
+
"loss": 0.4306,
|
| 2199 |
+
"step": 3590
|
| 2200 |
+
},
|
| 2201 |
+
{
|
| 2202 |
+
"epoch": 4.51,
|
| 2203 |
+
"learning_rate": 1.3785100951496868e-05,
|
| 2204 |
+
"loss": 0.4294,
|
| 2205 |
+
"step": 3600
|
| 2206 |
+
},
|
| 2207 |
+
{
|
| 2208 |
+
"epoch": 4.52,
|
| 2209 |
+
"learning_rate": 1.366906474820144e-05,
|
| 2210 |
+
"loss": 0.4147,
|
| 2211 |
+
"step": 3610
|
| 2212 |
+
},
|
| 2213 |
+
{
|
| 2214 |
+
"epoch": 4.54,
|
| 2215 |
+
"learning_rate": 1.3553028544906011e-05,
|
| 2216 |
+
"loss": 0.4619,
|
| 2217 |
+
"step": 3620
|
| 2218 |
+
},
|
| 2219 |
+
{
|
| 2220 |
+
"epoch": 4.55,
|
| 2221 |
+
"learning_rate": 1.3436992341610583e-05,
|
| 2222 |
+
"loss": 0.4146,
|
| 2223 |
+
"step": 3630
|
| 2224 |
+
},
|
| 2225 |
+
{
|
| 2226 |
+
"epoch": 4.56,
|
| 2227 |
+
"learning_rate": 1.3320956138315155e-05,
|
| 2228 |
+
"loss": 0.4718,
|
| 2229 |
+
"step": 3640
|
| 2230 |
+
},
|
| 2231 |
+
{
|
| 2232 |
+
"epoch": 4.57,
|
| 2233 |
+
"learning_rate": 1.3204919935019728e-05,
|
| 2234 |
+
"loss": 0.3824,
|
| 2235 |
+
"step": 3650
|
| 2236 |
+
},
|
| 2237 |
+
{
|
| 2238 |
+
"epoch": 4.59,
|
| 2239 |
+
"learning_rate": 1.30888837317243e-05,
|
| 2240 |
+
"loss": 0.4173,
|
| 2241 |
+
"step": 3660
|
| 2242 |
+
},
|
| 2243 |
+
{
|
| 2244 |
+
"epoch": 4.6,
|
| 2245 |
+
"learning_rate": 1.2972847528428869e-05,
|
| 2246 |
+
"loss": 0.4221,
|
| 2247 |
+
"step": 3670
|
| 2248 |
+
},
|
| 2249 |
+
{
|
| 2250 |
+
"epoch": 4.61,
|
| 2251 |
+
"learning_rate": 1.2856811325133441e-05,
|
| 2252 |
+
"loss": 0.4027,
|
| 2253 |
+
"step": 3680
|
| 2254 |
+
},
|
| 2255 |
+
{
|
| 2256 |
+
"epoch": 4.62,
|
| 2257 |
+
"learning_rate": 1.2740775121838014e-05,
|
| 2258 |
+
"loss": 0.4,
|
| 2259 |
+
"step": 3690
|
| 2260 |
+
},
|
| 2261 |
+
{
|
| 2262 |
+
"epoch": 4.64,
|
| 2263 |
+
"learning_rate": 1.2624738918542586e-05,
|
| 2264 |
+
"loss": 0.3792,
|
| 2265 |
+
"step": 3700
|
| 2266 |
+
},
|
| 2267 |
+
{
|
| 2268 |
+
"epoch": 4.65,
|
| 2269 |
+
"learning_rate": 1.2508702715247158e-05,
|
| 2270 |
+
"loss": 0.4146,
|
| 2271 |
+
"step": 3710
|
| 2272 |
+
},
|
| 2273 |
+
{
|
| 2274 |
+
"epoch": 4.66,
|
| 2275 |
+
"learning_rate": 1.2392666511951729e-05,
|
| 2276 |
+
"loss": 0.4145,
|
| 2277 |
+
"step": 3720
|
| 2278 |
+
},
|
| 2279 |
+
{
|
| 2280 |
+
"epoch": 4.67,
|
| 2281 |
+
"learning_rate": 1.2276630308656301e-05,
|
| 2282 |
+
"loss": 0.4651,
|
| 2283 |
+
"step": 3730
|
| 2284 |
+
},
|
| 2285 |
+
{
|
| 2286 |
+
"epoch": 4.69,
|
| 2287 |
+
"learning_rate": 1.2160594105360874e-05,
|
| 2288 |
+
"loss": 0.4459,
|
| 2289 |
+
"step": 3740
|
| 2290 |
+
},
|
| 2291 |
+
{
|
| 2292 |
+
"epoch": 4.7,
|
| 2293 |
+
"learning_rate": 1.2044557902065446e-05,
|
| 2294 |
+
"loss": 0.4244,
|
| 2295 |
+
"step": 3750
|
| 2296 |
+
},
|
| 2297 |
+
{
|
| 2298 |
+
"epoch": 4.71,
|
| 2299 |
+
"learning_rate": 1.1928521698770016e-05,
|
| 2300 |
+
"loss": 0.476,
|
| 2301 |
+
"step": 3760
|
| 2302 |
+
},
|
| 2303 |
+
{
|
| 2304 |
+
"epoch": 4.72,
|
| 2305 |
+
"learning_rate": 1.1812485495474589e-05,
|
| 2306 |
+
"loss": 0.4,
|
| 2307 |
+
"step": 3770
|
| 2308 |
+
},
|
| 2309 |
+
{
|
| 2310 |
+
"epoch": 4.74,
|
| 2311 |
+
"learning_rate": 1.1696449292179161e-05,
|
| 2312 |
+
"loss": 0.4452,
|
| 2313 |
+
"step": 3780
|
| 2314 |
+
},
|
| 2315 |
+
{
|
| 2316 |
+
"epoch": 4.75,
|
| 2317 |
+
"learning_rate": 1.1580413088883732e-05,
|
| 2318 |
+
"loss": 0.3968,
|
| 2319 |
+
"step": 3790
|
| 2320 |
+
},
|
| 2321 |
+
{
|
| 2322 |
+
"epoch": 4.76,
|
| 2323 |
+
"learning_rate": 1.1464376885588304e-05,
|
| 2324 |
+
"loss": 0.4265,
|
| 2325 |
+
"step": 3800
|
| 2326 |
+
},
|
| 2327 |
+
{
|
| 2328 |
+
"epoch": 4.77,
|
| 2329 |
+
"learning_rate": 1.1348340682292875e-05,
|
| 2330 |
+
"loss": 0.414,
|
| 2331 |
+
"step": 3810
|
| 2332 |
+
},
|
| 2333 |
+
{
|
| 2334 |
+
"epoch": 4.79,
|
| 2335 |
+
"learning_rate": 1.1232304478997447e-05,
|
| 2336 |
+
"loss": 0.4283,
|
| 2337 |
+
"step": 3820
|
| 2338 |
+
},
|
| 2339 |
+
{
|
| 2340 |
+
"epoch": 4.8,
|
| 2341 |
+
"learning_rate": 1.1116268275702021e-05,
|
| 2342 |
+
"loss": 0.4335,
|
| 2343 |
+
"step": 3830
|
| 2344 |
+
},
|
| 2345 |
+
{
|
| 2346 |
+
"epoch": 4.81,
|
| 2347 |
+
"learning_rate": 1.1000232072406592e-05,
|
| 2348 |
+
"loss": 0.4735,
|
| 2349 |
+
"step": 3840
|
| 2350 |
+
},
|
| 2351 |
+
{
|
| 2352 |
+
"epoch": 4.82,
|
| 2353 |
+
"learning_rate": 1.0884195869111164e-05,
|
| 2354 |
+
"loss": 0.4222,
|
| 2355 |
+
"step": 3850
|
| 2356 |
+
},
|
| 2357 |
+
{
|
| 2358 |
+
"epoch": 4.84,
|
| 2359 |
+
"learning_rate": 1.0768159665815734e-05,
|
| 2360 |
+
"loss": 0.4328,
|
| 2361 |
+
"step": 3860
|
| 2362 |
+
},
|
| 2363 |
+
{
|
| 2364 |
+
"epoch": 4.85,
|
| 2365 |
+
"learning_rate": 1.0652123462520307e-05,
|
| 2366 |
+
"loss": 0.4262,
|
| 2367 |
+
"step": 3870
|
| 2368 |
+
},
|
| 2369 |
+
{
|
| 2370 |
+
"epoch": 4.86,
|
| 2371 |
+
"learning_rate": 1.0536087259224879e-05,
|
| 2372 |
+
"loss": 0.4021,
|
| 2373 |
+
"step": 3880
|
| 2374 |
+
},
|
| 2375 |
+
{
|
| 2376 |
+
"epoch": 4.87,
|
| 2377 |
+
"learning_rate": 1.042005105592945e-05,
|
| 2378 |
+
"loss": 0.4266,
|
| 2379 |
+
"step": 3890
|
| 2380 |
+
},
|
| 2381 |
+
{
|
| 2382 |
+
"epoch": 4.89,
|
| 2383 |
+
"learning_rate": 1.0304014852634022e-05,
|
| 2384 |
+
"loss": 0.4273,
|
| 2385 |
+
"step": 3900
|
| 2386 |
+
},
|
| 2387 |
+
{
|
| 2388 |
+
"epoch": 4.9,
|
| 2389 |
+
"learning_rate": 1.0187978649338594e-05,
|
| 2390 |
+
"loss": 0.4609,
|
| 2391 |
+
"step": 3910
|
| 2392 |
+
},
|
| 2393 |
+
{
|
| 2394 |
+
"epoch": 4.91,
|
| 2395 |
+
"learning_rate": 1.0071942446043167e-05,
|
| 2396 |
+
"loss": 0.4008,
|
| 2397 |
+
"step": 3920
|
| 2398 |
+
},
|
| 2399 |
+
{
|
| 2400 |
+
"epoch": 4.92,
|
| 2401 |
+
"learning_rate": 9.955906242747739e-06,
|
| 2402 |
+
"loss": 0.3794,
|
| 2403 |
+
"step": 3930
|
| 2404 |
+
},
|
| 2405 |
+
{
|
| 2406 |
+
"epoch": 4.94,
|
| 2407 |
+
"learning_rate": 9.83987003945231e-06,
|
| 2408 |
+
"loss": 0.3763,
|
| 2409 |
+
"step": 3940
|
| 2410 |
+
},
|
| 2411 |
+
{
|
| 2412 |
+
"epoch": 4.95,
|
| 2413 |
+
"learning_rate": 9.723833836156882e-06,
|
| 2414 |
+
"loss": 0.4246,
|
| 2415 |
+
"step": 3950
|
| 2416 |
+
},
|
| 2417 |
+
{
|
| 2418 |
+
"epoch": 4.96,
|
| 2419 |
+
"learning_rate": 9.607797632861454e-06,
|
| 2420 |
+
"loss": 0.4331,
|
| 2421 |
+
"step": 3960
|
| 2422 |
+
},
|
| 2423 |
+
{
|
| 2424 |
+
"epoch": 4.97,
|
| 2425 |
+
"learning_rate": 9.491761429566025e-06,
|
| 2426 |
+
"loss": 0.4279,
|
| 2427 |
+
"step": 3970
|
| 2428 |
+
},
|
| 2429 |
+
{
|
| 2430 |
+
"epoch": 4.99,
|
| 2431 |
+
"learning_rate": 9.375725226270597e-06,
|
| 2432 |
+
"loss": 0.4245,
|
| 2433 |
+
"step": 3980
|
| 2434 |
+
},
|
| 2435 |
+
{
|
| 2436 |
+
"epoch": 5.0,
|
| 2437 |
+
"learning_rate": 9.259689022975168e-06,
|
| 2438 |
+
"loss": 0.4321,
|
| 2439 |
+
"step": 3990
|
| 2440 |
+
},
|
| 2441 |
+
{
|
| 2442 |
+
"epoch": 5.0,
|
| 2443 |
+
"eval_accuracy": 0.7823073308270677,
|
| 2444 |
+
"eval_loss": 0.6423132419586182,
|
| 2445 |
+
"eval_runtime": 243.4032,
|
| 2446 |
+
"eval_samples_per_second": 104.912,
|
| 2447 |
+
"eval_steps_per_second": 3.279,
|
| 2448 |
+
"step": 3990
|
| 2449 |
+
},
|
| 2450 |
+
{
|
| 2451 |
+
"epoch": 5.01,
|
| 2452 |
+
"learning_rate": 9.14365281967974e-06,
|
| 2453 |
+
"loss": 0.3513,
|
| 2454 |
+
"step": 4000
|
| 2455 |
+
},
|
| 2456 |
+
{
|
| 2457 |
+
"epoch": 5.03,
|
| 2458 |
+
"learning_rate": 9.027616616384312e-06,
|
| 2459 |
+
"loss": 0.337,
|
| 2460 |
+
"step": 4010
|
| 2461 |
+
},
|
| 2462 |
+
{
|
| 2463 |
+
"epoch": 5.04,
|
| 2464 |
+
"learning_rate": 8.911580413088885e-06,
|
| 2465 |
+
"loss": 0.3227,
|
| 2466 |
+
"step": 4020
|
| 2467 |
+
},
|
| 2468 |
+
{
|
| 2469 |
+
"epoch": 5.05,
|
| 2470 |
+
"learning_rate": 8.795544209793457e-06,
|
| 2471 |
+
"loss": 0.3443,
|
| 2472 |
+
"step": 4030
|
| 2473 |
+
},
|
| 2474 |
+
{
|
| 2475 |
+
"epoch": 5.06,
|
| 2476 |
+
"learning_rate": 8.679508006498028e-06,
|
| 2477 |
+
"loss": 0.3443,
|
| 2478 |
+
"step": 4040
|
| 2479 |
+
},
|
| 2480 |
+
{
|
| 2481 |
+
"epoch": 5.08,
|
| 2482 |
+
"learning_rate": 8.5634718032026e-06,
|
| 2483 |
+
"loss": 0.3033,
|
| 2484 |
+
"step": 4050
|
| 2485 |
+
},
|
| 2486 |
+
{
|
| 2487 |
+
"epoch": 5.09,
|
| 2488 |
+
"learning_rate": 8.447435599907172e-06,
|
| 2489 |
+
"loss": 0.3805,
|
| 2490 |
+
"step": 4060
|
| 2491 |
+
},
|
| 2492 |
+
{
|
| 2493 |
+
"epoch": 5.1,
|
| 2494 |
+
"learning_rate": 8.331399396611743e-06,
|
| 2495 |
+
"loss": 0.2883,
|
| 2496 |
+
"step": 4070
|
| 2497 |
+
},
|
| 2498 |
+
{
|
| 2499 |
+
"epoch": 5.11,
|
| 2500 |
+
"learning_rate": 8.215363193316315e-06,
|
| 2501 |
+
"loss": 0.3702,
|
| 2502 |
+
"step": 4080
|
| 2503 |
+
},
|
| 2504 |
+
{
|
| 2505 |
+
"epoch": 5.13,
|
| 2506 |
+
"learning_rate": 8.099326990020886e-06,
|
| 2507 |
+
"loss": 0.3468,
|
| 2508 |
+
"step": 4090
|
| 2509 |
+
},
|
| 2510 |
+
{
|
| 2511 |
+
"epoch": 5.14,
|
| 2512 |
+
"learning_rate": 7.983290786725458e-06,
|
| 2513 |
+
"loss": 0.3937,
|
| 2514 |
+
"step": 4100
|
| 2515 |
+
},
|
| 2516 |
+
{
|
| 2517 |
+
"epoch": 5.15,
|
| 2518 |
+
"learning_rate": 7.86725458343003e-06,
|
| 2519 |
+
"loss": 0.4071,
|
| 2520 |
+
"step": 4110
|
| 2521 |
+
},
|
| 2522 |
+
{
|
| 2523 |
+
"epoch": 5.16,
|
| 2524 |
+
"learning_rate": 7.751218380134603e-06,
|
| 2525 |
+
"loss": 0.3627,
|
| 2526 |
+
"step": 4120
|
| 2527 |
+
},
|
| 2528 |
+
{
|
| 2529 |
+
"epoch": 5.18,
|
| 2530 |
+
"learning_rate": 7.635182176839175e-06,
|
| 2531 |
+
"loss": 0.397,
|
| 2532 |
+
"step": 4130
|
| 2533 |
+
},
|
| 2534 |
+
{
|
| 2535 |
+
"epoch": 5.19,
|
| 2536 |
+
"learning_rate": 7.519145973543746e-06,
|
| 2537 |
+
"loss": 0.338,
|
| 2538 |
+
"step": 4140
|
| 2539 |
+
},
|
| 2540 |
+
{
|
| 2541 |
+
"epoch": 5.2,
|
| 2542 |
+
"learning_rate": 7.403109770248318e-06,
|
| 2543 |
+
"loss": 0.3559,
|
| 2544 |
+
"step": 4150
|
| 2545 |
+
},
|
| 2546 |
+
{
|
| 2547 |
+
"epoch": 5.21,
|
| 2548 |
+
"learning_rate": 7.28707356695289e-06,
|
| 2549 |
+
"loss": 0.3805,
|
| 2550 |
+
"step": 4160
|
| 2551 |
+
},
|
| 2552 |
+
{
|
| 2553 |
+
"epoch": 5.23,
|
| 2554 |
+
"learning_rate": 7.171037363657461e-06,
|
| 2555 |
+
"loss": 0.3679,
|
| 2556 |
+
"step": 4170
|
| 2557 |
+
},
|
| 2558 |
+
{
|
| 2559 |
+
"epoch": 5.24,
|
| 2560 |
+
"learning_rate": 7.055001160362033e-06,
|
| 2561 |
+
"loss": 0.3594,
|
| 2562 |
+
"step": 4180
|
| 2563 |
+
},
|
| 2564 |
+
{
|
| 2565 |
+
"epoch": 5.25,
|
| 2566 |
+
"learning_rate": 6.938964957066605e-06,
|
| 2567 |
+
"loss": 0.3311,
|
| 2568 |
+
"step": 4190
|
| 2569 |
+
},
|
| 2570 |
+
{
|
| 2571 |
+
"epoch": 5.26,
|
| 2572 |
+
"learning_rate": 6.822928753771177e-06,
|
| 2573 |
+
"loss": 0.3348,
|
| 2574 |
+
"step": 4200
|
| 2575 |
+
},
|
| 2576 |
+
{
|
| 2577 |
+
"epoch": 5.28,
|
| 2578 |
+
"learning_rate": 6.706892550475749e-06,
|
| 2579 |
+
"loss": 0.3552,
|
| 2580 |
+
"step": 4210
|
| 2581 |
+
},
|
| 2582 |
+
{
|
| 2583 |
+
"epoch": 5.29,
|
| 2584 |
+
"learning_rate": 6.59085634718032e-06,
|
| 2585 |
+
"loss": 0.34,
|
| 2586 |
+
"step": 4220
|
| 2587 |
+
},
|
| 2588 |
+
{
|
| 2589 |
+
"epoch": 5.3,
|
| 2590 |
+
"learning_rate": 6.474820143884892e-06,
|
| 2591 |
+
"loss": 0.3524,
|
| 2592 |
+
"step": 4230
|
| 2593 |
+
},
|
| 2594 |
+
{
|
| 2595 |
+
"epoch": 5.31,
|
| 2596 |
+
"learning_rate": 6.358783940589464e-06,
|
| 2597 |
+
"loss": 0.3401,
|
| 2598 |
+
"step": 4240
|
| 2599 |
+
},
|
| 2600 |
+
{
|
| 2601 |
+
"epoch": 5.33,
|
| 2602 |
+
"learning_rate": 6.242747737294036e-06,
|
| 2603 |
+
"loss": 0.3467,
|
| 2604 |
+
"step": 4250
|
| 2605 |
+
},
|
| 2606 |
+
{
|
| 2607 |
+
"epoch": 5.34,
|
| 2608 |
+
"learning_rate": 6.126711533998608e-06,
|
| 2609 |
+
"loss": 0.3115,
|
| 2610 |
+
"step": 4260
|
| 2611 |
+
},
|
| 2612 |
+
{
|
| 2613 |
+
"epoch": 5.35,
|
| 2614 |
+
"learning_rate": 6.01067533070318e-06,
|
| 2615 |
+
"loss": 0.344,
|
| 2616 |
+
"step": 4270
|
| 2617 |
+
},
|
| 2618 |
+
{
|
| 2619 |
+
"epoch": 5.36,
|
| 2620 |
+
"learning_rate": 5.894639127407751e-06,
|
| 2621 |
+
"loss": 0.3388,
|
| 2622 |
+
"step": 4280
|
| 2623 |
+
},
|
| 2624 |
+
{
|
| 2625 |
+
"epoch": 5.38,
|
| 2626 |
+
"learning_rate": 5.778602924112323e-06,
|
| 2627 |
+
"loss": 0.3483,
|
| 2628 |
+
"step": 4290
|
| 2629 |
+
},
|
| 2630 |
+
{
|
| 2631 |
+
"epoch": 5.39,
|
| 2632 |
+
"learning_rate": 5.662566720816896e-06,
|
| 2633 |
+
"loss": 0.355,
|
| 2634 |
+
"step": 4300
|
| 2635 |
+
},
|
| 2636 |
+
{
|
| 2637 |
+
"epoch": 5.4,
|
| 2638 |
+
"learning_rate": 5.546530517521467e-06,
|
| 2639 |
+
"loss": 0.3636,
|
| 2640 |
+
"step": 4310
|
| 2641 |
+
},
|
| 2642 |
+
{
|
| 2643 |
+
"epoch": 5.41,
|
| 2644 |
+
"learning_rate": 5.430494314226039e-06,
|
| 2645 |
+
"loss": 0.3675,
|
| 2646 |
+
"step": 4320
|
| 2647 |
+
},
|
| 2648 |
+
{
|
| 2649 |
+
"epoch": 5.43,
|
| 2650 |
+
"learning_rate": 5.31445811093061e-06,
|
| 2651 |
+
"loss": 0.3346,
|
| 2652 |
+
"step": 4330
|
| 2653 |
+
},
|
| 2654 |
+
{
|
| 2655 |
+
"epoch": 5.44,
|
| 2656 |
+
"learning_rate": 5.198421907635182e-06,
|
| 2657 |
+
"loss": 0.3191,
|
| 2658 |
+
"step": 4340
|
| 2659 |
+
},
|
| 2660 |
+
{
|
| 2661 |
+
"epoch": 5.45,
|
| 2662 |
+
"learning_rate": 5.082385704339755e-06,
|
| 2663 |
+
"loss": 0.3095,
|
| 2664 |
+
"step": 4350
|
| 2665 |
+
},
|
| 2666 |
+
{
|
| 2667 |
+
"epoch": 5.46,
|
| 2668 |
+
"learning_rate": 4.966349501044326e-06,
|
| 2669 |
+
"loss": 0.357,
|
| 2670 |
+
"step": 4360
|
| 2671 |
+
},
|
| 2672 |
+
{
|
| 2673 |
+
"epoch": 5.48,
|
| 2674 |
+
"learning_rate": 4.850313297748898e-06,
|
| 2675 |
+
"loss": 0.3123,
|
| 2676 |
+
"step": 4370
|
| 2677 |
+
},
|
| 2678 |
+
{
|
| 2679 |
+
"epoch": 5.49,
|
| 2680 |
+
"learning_rate": 4.734277094453469e-06,
|
| 2681 |
+
"loss": 0.347,
|
| 2682 |
+
"step": 4380
|
| 2683 |
+
},
|
| 2684 |
+
{
|
| 2685 |
+
"epoch": 5.5,
|
| 2686 |
+
"learning_rate": 4.618240891158042e-06,
|
| 2687 |
+
"loss": 0.3338,
|
| 2688 |
+
"step": 4390
|
| 2689 |
+
},
|
| 2690 |
+
{
|
| 2691 |
+
"epoch": 5.51,
|
| 2692 |
+
"learning_rate": 4.502204687862614e-06,
|
| 2693 |
+
"loss": 0.3698,
|
| 2694 |
+
"step": 4400
|
| 2695 |
+
},
|
| 2696 |
+
{
|
| 2697 |
+
"epoch": 5.53,
|
| 2698 |
+
"learning_rate": 4.386168484567185e-06,
|
| 2699 |
+
"loss": 0.3654,
|
| 2700 |
+
"step": 4410
|
| 2701 |
+
},
|
| 2702 |
+
{
|
| 2703 |
+
"epoch": 5.54,
|
| 2704 |
+
"learning_rate": 4.270132281271757e-06,
|
| 2705 |
+
"loss": 0.3458,
|
| 2706 |
+
"step": 4420
|
| 2707 |
+
},
|
| 2708 |
+
{
|
| 2709 |
+
"epoch": 5.55,
|
| 2710 |
+
"learning_rate": 4.154096077976329e-06,
|
| 2711 |
+
"loss": 0.3096,
|
| 2712 |
+
"step": 4430
|
| 2713 |
+
},
|
| 2714 |
+
{
|
| 2715 |
+
"epoch": 5.56,
|
| 2716 |
+
"learning_rate": 4.038059874680901e-06,
|
| 2717 |
+
"loss": 0.3391,
|
| 2718 |
+
"step": 4440
|
| 2719 |
+
},
|
| 2720 |
+
{
|
| 2721 |
+
"epoch": 5.58,
|
| 2722 |
+
"learning_rate": 3.922023671385473e-06,
|
| 2723 |
+
"loss": 0.3683,
|
| 2724 |
+
"step": 4450
|
| 2725 |
+
},
|
| 2726 |
+
{
|
| 2727 |
+
"epoch": 5.59,
|
| 2728 |
+
"learning_rate": 3.805987468090044e-06,
|
| 2729 |
+
"loss": 0.3855,
|
| 2730 |
+
"step": 4460
|
| 2731 |
+
},
|
| 2732 |
+
{
|
| 2733 |
+
"epoch": 5.6,
|
| 2734 |
+
"learning_rate": 3.689951264794616e-06,
|
| 2735 |
+
"loss": 0.353,
|
| 2736 |
+
"step": 4470
|
| 2737 |
+
},
|
| 2738 |
+
{
|
| 2739 |
+
"epoch": 5.61,
|
| 2740 |
+
"learning_rate": 3.5739150614991875e-06,
|
| 2741 |
+
"loss": 0.3303,
|
| 2742 |
+
"step": 4480
|
| 2743 |
+
},
|
| 2744 |
+
{
|
| 2745 |
+
"epoch": 5.63,
|
| 2746 |
+
"learning_rate": 3.45787885820376e-06,
|
| 2747 |
+
"loss": 0.2987,
|
| 2748 |
+
"step": 4490
|
| 2749 |
+
},
|
| 2750 |
+
{
|
| 2751 |
+
"epoch": 5.64,
|
| 2752 |
+
"learning_rate": 3.3418426549083317e-06,
|
| 2753 |
+
"loss": 0.3412,
|
| 2754 |
+
"step": 4500
|
| 2755 |
+
},
|
| 2756 |
+
{
|
| 2757 |
+
"epoch": 5.65,
|
| 2758 |
+
"learning_rate": 3.225806451612903e-06,
|
| 2759 |
+
"loss": 0.3066,
|
| 2760 |
+
"step": 4510
|
| 2761 |
+
},
|
| 2762 |
+
{
|
| 2763 |
+
"epoch": 5.66,
|
| 2764 |
+
"learning_rate": 3.1097702483174755e-06,
|
| 2765 |
+
"loss": 0.3252,
|
| 2766 |
+
"step": 4520
|
| 2767 |
+
},
|
| 2768 |
+
{
|
| 2769 |
+
"epoch": 5.68,
|
| 2770 |
+
"learning_rate": 2.993734045022047e-06,
|
| 2771 |
+
"loss": 0.3243,
|
| 2772 |
+
"step": 4530
|
| 2773 |
+
},
|
| 2774 |
+
{
|
| 2775 |
+
"epoch": 5.69,
|
| 2776 |
+
"learning_rate": 2.877697841726619e-06,
|
| 2777 |
+
"loss": 0.3888,
|
| 2778 |
+
"step": 4540
|
| 2779 |
+
},
|
| 2780 |
+
{
|
| 2781 |
+
"epoch": 5.7,
|
| 2782 |
+
"learning_rate": 2.7616616384311907e-06,
|
| 2783 |
+
"loss": 0.3635,
|
| 2784 |
+
"step": 4550
|
| 2785 |
+
},
|
| 2786 |
+
{
|
| 2787 |
+
"epoch": 5.71,
|
| 2788 |
+
"learning_rate": 2.645625435135762e-06,
|
| 2789 |
+
"loss": 0.2875,
|
| 2790 |
+
"step": 4560
|
| 2791 |
+
},
|
| 2792 |
+
{
|
| 2793 |
+
"epoch": 5.73,
|
| 2794 |
+
"learning_rate": 2.5295892318403345e-06,
|
| 2795 |
+
"loss": 0.3443,
|
| 2796 |
+
"step": 4570
|
| 2797 |
+
},
|
| 2798 |
+
{
|
| 2799 |
+
"epoch": 5.74,
|
| 2800 |
+
"learning_rate": 2.413553028544906e-06,
|
| 2801 |
+
"loss": 0.3334,
|
| 2802 |
+
"step": 4580
|
| 2803 |
+
},
|
| 2804 |
+
{
|
| 2805 |
+
"epoch": 5.75,
|
| 2806 |
+
"learning_rate": 2.297516825249478e-06,
|
| 2807 |
+
"loss": 0.337,
|
| 2808 |
+
"step": 4590
|
| 2809 |
+
},
|
| 2810 |
+
{
|
| 2811 |
+
"epoch": 5.76,
|
| 2812 |
+
"learning_rate": 2.1814806219540497e-06,
|
| 2813 |
+
"loss": 0.3204,
|
| 2814 |
+
"step": 4600
|
| 2815 |
+
},
|
| 2816 |
+
{
|
| 2817 |
+
"epoch": 5.78,
|
| 2818 |
+
"learning_rate": 2.0654444186586216e-06,
|
| 2819 |
+
"loss": 0.318,
|
| 2820 |
+
"step": 4610
|
| 2821 |
+
},
|
| 2822 |
+
{
|
| 2823 |
+
"epoch": 5.79,
|
| 2824 |
+
"learning_rate": 1.9494082153631935e-06,
|
| 2825 |
+
"loss": 0.3113,
|
| 2826 |
+
"step": 4620
|
| 2827 |
+
},
|
| 2828 |
+
{
|
| 2829 |
+
"epoch": 5.8,
|
| 2830 |
+
"learning_rate": 1.833372012067765e-06,
|
| 2831 |
+
"loss": 0.3623,
|
| 2832 |
+
"step": 4630
|
| 2833 |
+
},
|
| 2834 |
+
{
|
| 2835 |
+
"epoch": 5.81,
|
| 2836 |
+
"learning_rate": 1.717335808772337e-06,
|
| 2837 |
+
"loss": 0.3489,
|
| 2838 |
+
"step": 4640
|
| 2839 |
+
},
|
| 2840 |
+
{
|
| 2841 |
+
"epoch": 5.83,
|
| 2842 |
+
"learning_rate": 1.6012996054769087e-06,
|
| 2843 |
+
"loss": 0.3477,
|
| 2844 |
+
"step": 4650
|
| 2845 |
+
},
|
| 2846 |
+
{
|
| 2847 |
+
"epoch": 5.84,
|
| 2848 |
+
"learning_rate": 1.4852634021814808e-06,
|
| 2849 |
+
"loss": 0.3124,
|
| 2850 |
+
"step": 4660
|
| 2851 |
+
},
|
| 2852 |
+
{
|
| 2853 |
+
"epoch": 5.85,
|
| 2854 |
+
"learning_rate": 1.3692271988860525e-06,
|
| 2855 |
+
"loss": 0.3658,
|
| 2856 |
+
"step": 4670
|
| 2857 |
+
},
|
| 2858 |
+
{
|
| 2859 |
+
"epoch": 5.86,
|
| 2860 |
+
"learning_rate": 1.2531909955906242e-06,
|
| 2861 |
+
"loss": 0.3652,
|
| 2862 |
+
"step": 4680
|
| 2863 |
+
},
|
| 2864 |
+
{
|
| 2865 |
+
"epoch": 5.88,
|
| 2866 |
+
"learning_rate": 1.137154792295196e-06,
|
| 2867 |
+
"loss": 0.3415,
|
| 2868 |
+
"step": 4690
|
| 2869 |
+
},
|
| 2870 |
+
{
|
| 2871 |
+
"epoch": 5.89,
|
| 2872 |
+
"learning_rate": 1.021118588999768e-06,
|
| 2873 |
+
"loss": 0.3479,
|
| 2874 |
+
"step": 4700
|
| 2875 |
+
},
|
| 2876 |
+
{
|
| 2877 |
+
"epoch": 5.9,
|
| 2878 |
+
"learning_rate": 9.050823857043398e-07,
|
| 2879 |
+
"loss": 0.3746,
|
| 2880 |
+
"step": 4710
|
| 2881 |
+
},
|
| 2882 |
+
{
|
| 2883 |
+
"epoch": 5.91,
|
| 2884 |
+
"learning_rate": 7.890461824089116e-07,
|
| 2885 |
+
"loss": 0.3311,
|
| 2886 |
+
"step": 4720
|
| 2887 |
+
},
|
| 2888 |
+
{
|
| 2889 |
+
"epoch": 5.93,
|
| 2890 |
+
"learning_rate": 6.730099791134834e-07,
|
| 2891 |
+
"loss": 0.3282,
|
| 2892 |
+
"step": 4730
|
| 2893 |
+
},
|
| 2894 |
+
{
|
| 2895 |
+
"epoch": 5.94,
|
| 2896 |
+
"learning_rate": 5.569737758180553e-07,
|
| 2897 |
+
"loss": 0.3269,
|
| 2898 |
+
"step": 4740
|
| 2899 |
+
},
|
| 2900 |
+
{
|
| 2901 |
+
"epoch": 5.95,
|
| 2902 |
+
"learning_rate": 4.4093757252262707e-07,
|
| 2903 |
+
"loss": 0.3035,
|
| 2904 |
+
"step": 4750
|
| 2905 |
+
},
|
| 2906 |
+
{
|
| 2907 |
+
"epoch": 5.96,
|
| 2908 |
+
"learning_rate": 3.249013692271989e-07,
|
| 2909 |
+
"loss": 0.3208,
|
| 2910 |
+
"step": 4760
|
| 2911 |
+
},
|
| 2912 |
+
{
|
| 2913 |
+
"epoch": 5.98,
|
| 2914 |
+
"learning_rate": 2.0886516593177073e-07,
|
| 2915 |
+
"loss": 0.3553,
|
| 2916 |
+
"step": 4770
|
| 2917 |
+
},
|
| 2918 |
+
{
|
| 2919 |
+
"epoch": 5.99,
|
| 2920 |
+
"learning_rate": 9.282896263634255e-08,
|
| 2921 |
+
"loss": 0.3129,
|
| 2922 |
+
"step": 4780
|
| 2923 |
+
},
|
| 2924 |
+
{
|
| 2925 |
+
"epoch": 6.0,
|
| 2926 |
+
"eval_accuracy": 0.7837562656641605,
|
| 2927 |
+
"eval_loss": 0.6533188819885254,
|
| 2928 |
+
"eval_runtime": 249.7809,
|
| 2929 |
+
"eval_samples_per_second": 102.234,
|
| 2930 |
+
"eval_steps_per_second": 3.195,
|
| 2931 |
+
"step": 4788
|
| 2932 |
+
},
|
| 2933 |
+
{
|
| 2934 |
+
"epoch": 6.0,
|
| 2935 |
+
"step": 4788,
|
| 2936 |
+
"total_flos": 4.749417900430118e+19,
|
| 2937 |
+
"train_loss": 0.321556950521748,
|
| 2938 |
+
"train_runtime": 10205.5984,
|
| 2939 |
+
"train_samples_per_second": 60.052,
|
| 2940 |
+
"train_steps_per_second": 0.469
|
| 2941 |
}
|
| 2942 |
],
|
| 2943 |
"logging_steps": 10,
|
| 2944 |
+
"max_steps": 4788,
|
| 2945 |
"num_input_tokens_seen": 0,
|
| 2946 |
+
"num_train_epochs": 6,
|
| 2947 |
"save_steps": 500,
|
| 2948 |
+
"total_flos": 4.749417900430118e+19,
|
| 2949 |
"train_batch_size": 32,
|
| 2950 |
"trial_name": null,
|
| 2951 |
"trial_params": null
|