Commit
·
31ba05f
1
Parent(s):
379aa41
End of training
Browse files- all_results.json +11 -11
- eval_results.json +6 -6
- runs/Feb07_08-02-43_dcad63cb3363/events.out.tfevents.1675765217.dcad63cb3363.606.2 +3 -0
- train_results.json +6 -6
- trainer_state.json +1039 -631
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": 2.
|
| 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": 5.0,
|
| 3 |
+
"eval_accuracy": 0.9741915422885572,
|
| 4 |
+
"eval_loss": 0.06838314980268478,
|
| 5 |
+
"eval_runtime": 131.8368,
|
| 6 |
+
"eval_samples_per_second": 73.181,
|
| 7 |
+
"eval_steps_per_second": 2.291,
|
| 8 |
+
"total_flos": 8.391154825876193e+18,
|
| 9 |
+
"train_loss": 0.33317647269826234,
|
| 10 |
+
"train_runtime": 8035.3313,
|
| 11 |
+
"train_samples_per_second": 42.02,
|
| 12 |
+
"train_steps_per_second": 0.328
|
| 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": 2.
|
| 8 |
}
|
|
|
|
| 1 |
{
|
| 2 |
+
"epoch": 5.0,
|
| 3 |
+
"eval_accuracy": 0.9741915422885572,
|
| 4 |
+
"eval_loss": 0.06838314980268478,
|
| 5 |
+
"eval_runtime": 131.8368,
|
| 6 |
+
"eval_samples_per_second": 73.181,
|
| 7 |
+
"eval_steps_per_second": 2.291
|
| 8 |
}
|
runs/Feb07_08-02-43_dcad63cb3363/events.out.tfevents.1675765217.dcad63cb3363.606.2
ADDED
|
@@ -0,0 +1,3 @@
|
|
|
|
|
|
|
|
|
|
|
|
|
| 1 |
+
version https://git-lfs.github.com/spec/v1
|
| 2 |
+
oid sha256:dd69521417bb4f8013807a4d9ed2212a4137fc9296960c4de72c518445b26ef1
|
| 3 |
+
size 363
|
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": 5.0,
|
| 3 |
+
"total_flos": 8.391154825876193e+18,
|
| 4 |
+
"train_loss": 0.33317647269826234,
|
| 5 |
+
"train_runtime": 8035.3313,
|
| 6 |
+
"train_samples_per_second": 42.02,
|
| 7 |
+
"train_steps_per_second": 0.328
|
| 8 |
}
|
trainer_state.json
CHANGED
|
@@ -1,1240 +1,1648 @@
|
|
| 1 |
{
|
| 2 |
-
"best_metric": 0.
|
| 3 |
-
"best_model_checkpoint": "swin-tiny-patch4-window7-224-finetuned-eurosat/checkpoint-
|
| 4 |
-
"epoch":
|
| 5 |
-
"global_step":
|
| 6 |
"is_hyper_param_search": false,
|
| 7 |
"is_local_process_zero": true,
|
| 8 |
"is_world_process_zero": true,
|
| 9 |
"log_history": [
|
| 10 |
{
|
| 11 |
"epoch": 0.02,
|
| 12 |
-
"learning_rate":
|
| 13 |
-
"loss": 1.
|
| 14 |
"step": 10
|
| 15 |
},
|
| 16 |
{
|
| 17 |
-
"epoch": 0.
|
| 18 |
-
"learning_rate":
|
| 19 |
-
"loss": 1.
|
| 20 |
"step": 20
|
| 21 |
},
|
| 22 |
{
|
| 23 |
-
"epoch": 0.
|
| 24 |
-
"learning_rate":
|
| 25 |
-
"loss": 1.
|
| 26 |
"step": 30
|
| 27 |
},
|
| 28 |
{
|
| 29 |
-
"epoch": 0.
|
| 30 |
-
"learning_rate":
|
| 31 |
-
"loss": 1.
|
| 32 |
"step": 40
|
| 33 |
},
|
| 34 |
{
|
| 35 |
-
"epoch": 0.
|
| 36 |
-
"learning_rate":
|
| 37 |
-
"loss": 0.
|
| 38 |
"step": 50
|
| 39 |
},
|
| 40 |
{
|
| 41 |
-
"epoch": 0.
|
| 42 |
-
"learning_rate": 1.
|
| 43 |
-
"loss": 0.
|
| 44 |
"step": 60
|
| 45 |
},
|
| 46 |
{
|
| 47 |
-
"epoch": 0.
|
| 48 |
-
"learning_rate": 1.
|
| 49 |
-
"loss": 0.
|
| 50 |
"step": 70
|
| 51 |
},
|
| 52 |
{
|
| 53 |
-
"epoch": 0.
|
| 54 |
-
"learning_rate":
|
| 55 |
-
"loss": 0.
|
| 56 |
"step": 80
|
| 57 |
},
|
| 58 |
{
|
| 59 |
-
"epoch": 0.
|
| 60 |
-
"learning_rate":
|
| 61 |
-
"loss": 0.
|
| 62 |
"step": 90
|
| 63 |
},
|
| 64 |
{
|
| 65 |
-
"epoch": 0.
|
| 66 |
-
"learning_rate":
|
| 67 |
-
"loss": 0.
|
| 68 |
"step": 100
|
| 69 |
},
|
| 70 |
{
|
| 71 |
-
"epoch": 0.
|
| 72 |
-
"learning_rate": 2.
|
| 73 |
-
"loss": 0.
|
| 74 |
"step": 110
|
| 75 |
},
|
| 76 |
{
|
| 77 |
-
"epoch": 0.
|
| 78 |
-
"learning_rate":
|
| 79 |
-
"loss": 0.
|
| 80 |
"step": 120
|
| 81 |
},
|
| 82 |
{
|
| 83 |
-
"epoch": 0.
|
| 84 |
-
"learning_rate":
|
| 85 |
-
"loss": 0.
|
| 86 |
"step": 130
|
| 87 |
},
|
| 88 |
{
|
| 89 |
-
"epoch": 0.
|
| 90 |
-
"learning_rate":
|
| 91 |
-
"loss": 0.
|
| 92 |
"step": 140
|
| 93 |
},
|
| 94 |
{
|
| 95 |
-
"epoch": 0.
|
| 96 |
-
"learning_rate":
|
| 97 |
-
"loss": 0.
|
| 98 |
"step": 150
|
| 99 |
},
|
| 100 |
{
|
| 101 |
-
"epoch": 0.
|
| 102 |
-
"learning_rate":
|
| 103 |
-
"loss": 0.
|
| 104 |
"step": 160
|
| 105 |
},
|
| 106 |
{
|
| 107 |
-
"epoch": 0.
|
| 108 |
-
"learning_rate":
|
| 109 |
-
"loss": 0.
|
| 110 |
"step": 170
|
| 111 |
},
|
| 112 |
{
|
| 113 |
-
"epoch": 0.
|
| 114 |
-
"learning_rate":
|
| 115 |
-
"loss": 0.
|
| 116 |
"step": 180
|
| 117 |
},
|
| 118 |
{
|
| 119 |
-
"epoch": 0.
|
| 120 |
-
"learning_rate":
|
| 121 |
-
"loss": 0.
|
| 122 |
"step": 190
|
| 123 |
},
|
| 124 |
{
|
| 125 |
-
"epoch": 0.
|
| 126 |
-
"learning_rate":
|
| 127 |
-
"loss": 0.
|
| 128 |
"step": 200
|
| 129 |
},
|
| 130 |
{
|
| 131 |
-
"epoch": 0.
|
| 132 |
-
"learning_rate":
|
| 133 |
-
"loss": 0.
|
| 134 |
"step": 210
|
| 135 |
},
|
| 136 |
{
|
| 137 |
-
"epoch": 0.
|
| 138 |
-
"learning_rate": 4.
|
| 139 |
-
"loss": 0.
|
| 140 |
"step": 220
|
| 141 |
},
|
| 142 |
{
|
| 143 |
-
"epoch": 0.
|
| 144 |
-
"learning_rate": 4.
|
| 145 |
-
"loss": 0.
|
| 146 |
"step": 230
|
| 147 |
},
|
| 148 |
{
|
| 149 |
-
"epoch": 0.
|
| 150 |
-
"learning_rate": 4.
|
| 151 |
-
"loss": 0.
|
| 152 |
"step": 240
|
| 153 |
},
|
| 154 |
{
|
| 155 |
-
"epoch": 0.
|
| 156 |
-
"learning_rate": 4.
|
| 157 |
-
"loss": 0.
|
| 158 |
"step": 250
|
| 159 |
},
|
| 160 |
{
|
| 161 |
-
"epoch": 0.
|
| 162 |
-
"learning_rate": 4.
|
| 163 |
-
"loss": 0.
|
| 164 |
"step": 260
|
| 165 |
},
|
| 166 |
{
|
| 167 |
-
"epoch": 0.
|
| 168 |
-
"learning_rate": 4.
|
| 169 |
-
"loss": 0.
|
| 170 |
"step": 270
|
| 171 |
},
|
| 172 |
{
|
| 173 |
-
"epoch": 0.
|
| 174 |
-
"learning_rate": 4.
|
| 175 |
-
"loss": 0.
|
| 176 |
"step": 280
|
| 177 |
},
|
| 178 |
{
|
| 179 |
-
"epoch": 0.
|
| 180 |
-
"learning_rate": 4.
|
| 181 |
-
"loss": 0.
|
| 182 |
"step": 290
|
| 183 |
},
|
| 184 |
{
|
| 185 |
-
"epoch": 0.
|
| 186 |
-
"learning_rate": 4.
|
| 187 |
-
"loss": 0.
|
| 188 |
"step": 300
|
| 189 |
},
|
| 190 |
{
|
| 191 |
-
"epoch": 0.
|
| 192 |
-
"learning_rate": 4.
|
| 193 |
-
"loss": 0.
|
| 194 |
"step": 310
|
| 195 |
},
|
| 196 |
{
|
| 197 |
-
"epoch": 0.
|
| 198 |
-
"learning_rate": 4.
|
| 199 |
-
"loss": 0.
|
| 200 |
"step": 320
|
| 201 |
},
|
| 202 |
{
|
| 203 |
-
"epoch": 0.
|
| 204 |
-
"learning_rate": 4.
|
| 205 |
-
"loss": 0.
|
| 206 |
"step": 330
|
| 207 |
},
|
| 208 |
{
|
| 209 |
-
"epoch": 0.
|
| 210 |
-
"learning_rate": 4.
|
| 211 |
-
"loss": 0.
|
| 212 |
"step": 340
|
| 213 |
},
|
| 214 |
{
|
| 215 |
-
"epoch": 0.
|
| 216 |
-
"learning_rate": 4.
|
| 217 |
-
"loss": 0.
|
| 218 |
"step": 350
|
| 219 |
},
|
| 220 |
{
|
| 221 |
-
"epoch": 0.
|
| 222 |
-
"learning_rate": 4.
|
| 223 |
-
"loss": 0.
|
| 224 |
"step": 360
|
| 225 |
},
|
| 226 |
{
|
| 227 |
-
"epoch": 0.
|
| 228 |
-
"learning_rate": 4.
|
| 229 |
-
"loss": 0.
|
| 230 |
"step": 370
|
| 231 |
},
|
| 232 |
{
|
| 233 |
-
"epoch": 0.
|
| 234 |
-
"learning_rate": 4.
|
| 235 |
-
"loss": 0.
|
| 236 |
"step": 380
|
| 237 |
},
|
| 238 |
{
|
| 239 |
-
"epoch": 0.
|
| 240 |
-
"learning_rate": 4.
|
| 241 |
-
"loss": 0.
|
| 242 |
"step": 390
|
| 243 |
},
|
| 244 |
{
|
| 245 |
-
"epoch": 0.
|
| 246 |
-
"learning_rate": 4.
|
| 247 |
-
"loss": 0.
|
| 248 |
"step": 400
|
| 249 |
},
|
| 250 |
{
|
| 251 |
-
"epoch": 0.
|
| 252 |
-
"learning_rate": 4.
|
| 253 |
-
"loss": 0.
|
| 254 |
"step": 410
|
| 255 |
},
|
| 256 |
{
|
| 257 |
-
"epoch": 0.
|
| 258 |
-
"learning_rate": 4.
|
| 259 |
-
"loss": 0.
|
| 260 |
"step": 420
|
| 261 |
},
|
| 262 |
{
|
| 263 |
-
"epoch": 0.
|
| 264 |
-
"learning_rate": 4.
|
| 265 |
-
"loss": 0.
|
| 266 |
"step": 430
|
| 267 |
},
|
| 268 |
{
|
| 269 |
-
"epoch": 0.
|
| 270 |
-
"learning_rate": 4.
|
| 271 |
-
"loss": 0.
|
| 272 |
"step": 440
|
| 273 |
},
|
| 274 |
{
|
| 275 |
-
"epoch": 0.
|
| 276 |
-
"learning_rate": 4.
|
| 277 |
-
"loss": 0.
|
| 278 |
"step": 450
|
| 279 |
},
|
| 280 |
{
|
| 281 |
-
"epoch": 0.
|
| 282 |
-
"learning_rate": 4.
|
| 283 |
-
"loss": 0.
|
| 284 |
"step": 460
|
| 285 |
},
|
| 286 |
{
|
| 287 |
-
"epoch": 0.
|
| 288 |
-
"learning_rate": 4.
|
| 289 |
-
"loss": 0.
|
| 290 |
"step": 470
|
| 291 |
},
|
| 292 |
{
|
| 293 |
-
"epoch": 0.
|
| 294 |
-
"learning_rate": 4.
|
| 295 |
-
"loss": 0.
|
| 296 |
"step": 480
|
| 297 |
},
|
| 298 |
{
|
| 299 |
-
"epoch": 0.
|
| 300 |
-
"learning_rate": 4.
|
| 301 |
-
"loss": 0.
|
| 302 |
"step": 490
|
| 303 |
},
|
| 304 |
{
|
| 305 |
-
"epoch": 0.
|
| 306 |
-
"learning_rate": 4.
|
| 307 |
-
"loss": 0.
|
| 308 |
"step": 500
|
| 309 |
},
|
| 310 |
{
|
| 311 |
-
"epoch": 0.
|
| 312 |
-
"learning_rate": 4.
|
| 313 |
-
"loss": 0.
|
| 314 |
"step": 510
|
| 315 |
},
|
| 316 |
{
|
| 317 |
-
"epoch": 0.
|
| 318 |
-
"learning_rate": 4.
|
| 319 |
-
"loss": 0.
|
| 320 |
"step": 520
|
| 321 |
},
|
| 322 |
{
|
| 323 |
-
"epoch": 0
|
| 324 |
-
"
|
| 325 |
-
"
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
| 326 |
"step": 530
|
| 327 |
},
|
| 328 |
{
|
| 329 |
-
"epoch":
|
| 330 |
-
"learning_rate": 4.
|
| 331 |
-
"loss": 0.
|
| 332 |
"step": 540
|
| 333 |
},
|
| 334 |
{
|
| 335 |
-
"epoch":
|
| 336 |
-
"learning_rate": 4.
|
| 337 |
-
"loss": 0.
|
| 338 |
"step": 550
|
| 339 |
},
|
| 340 |
{
|
| 341 |
-
"epoch":
|
| 342 |
-
"learning_rate":
|
| 343 |
-
"loss": 0.
|
| 344 |
"step": 560
|
| 345 |
},
|
| 346 |
{
|
| 347 |
-
"epoch":
|
| 348 |
-
"learning_rate":
|
| 349 |
-
"loss": 0.
|
| 350 |
"step": 570
|
| 351 |
},
|
| 352 |
{
|
| 353 |
-
"epoch":
|
| 354 |
-
"learning_rate":
|
| 355 |
-
"loss": 0.
|
| 356 |
"step": 580
|
| 357 |
},
|
| 358 |
{
|
| 359 |
-
"epoch":
|
| 360 |
-
"learning_rate":
|
| 361 |
-
"loss": 0.
|
| 362 |
"step": 590
|
| 363 |
},
|
| 364 |
{
|
| 365 |
-
"epoch":
|
| 366 |
-
"learning_rate":
|
| 367 |
-
"loss": 0.
|
| 368 |
"step": 600
|
| 369 |
},
|
| 370 |
{
|
| 371 |
-
"epoch":
|
| 372 |
-
"learning_rate":
|
| 373 |
-
"loss": 0.
|
| 374 |
"step": 610
|
| 375 |
},
|
| 376 |
{
|
| 377 |
-
"epoch":
|
| 378 |
-
"learning_rate":
|
| 379 |
-
"loss": 0.
|
| 380 |
"step": 620
|
| 381 |
},
|
| 382 |
{
|
| 383 |
-
"epoch":
|
| 384 |
-
"learning_rate":
|
| 385 |
-
"loss": 0.
|
| 386 |
"step": 630
|
| 387 |
},
|
| 388 |
{
|
| 389 |
-
"epoch":
|
| 390 |
-
"learning_rate":
|
| 391 |
-
"loss": 0.
|
| 392 |
"step": 640
|
| 393 |
},
|
| 394 |
{
|
| 395 |
-
"epoch":
|
| 396 |
-
"learning_rate":
|
| 397 |
-
"loss": 0.
|
| 398 |
"step": 650
|
| 399 |
},
|
| 400 |
{
|
| 401 |
-
"epoch": 1.
|
| 402 |
-
"learning_rate":
|
| 403 |
-
"loss": 0.
|
| 404 |
"step": 660
|
| 405 |
},
|
| 406 |
{
|
| 407 |
-
"epoch": 1.
|
| 408 |
-
"
|
| 409 |
-
"
|
| 410 |
-
"eval_runtime": 123.5087,
|
| 411 |
-
"eval_samples_per_second": 76.181,
|
| 412 |
-
"eval_steps_per_second": 2.388,
|
| 413 |
-
"step": 661
|
| 414 |
-
},
|
| 415 |
-
{
|
| 416 |
-
"epoch": 1.01,
|
| 417 |
-
"learning_rate": 3.679932735426009e-05,
|
| 418 |
-
"loss": 0.4176,
|
| 419 |
"step": 670
|
| 420 |
},
|
| 421 |
{
|
| 422 |
-
"epoch": 1.
|
| 423 |
-
"learning_rate":
|
| 424 |
-
"loss": 0.
|
| 425 |
"step": 680
|
| 426 |
},
|
| 427 |
{
|
| 428 |
-
"epoch": 1.
|
| 429 |
-
"learning_rate":
|
| 430 |
-
"loss": 0.
|
| 431 |
"step": 690
|
| 432 |
},
|
| 433 |
{
|
| 434 |
-
"epoch": 1.
|
| 435 |
-
"learning_rate":
|
| 436 |
-
"loss": 0.
|
| 437 |
"step": 700
|
| 438 |
},
|
| 439 |
{
|
| 440 |
-
"epoch": 1.
|
| 441 |
-
"learning_rate":
|
| 442 |
-
"loss": 0.
|
| 443 |
"step": 710
|
| 444 |
},
|
| 445 |
{
|
| 446 |
-
"epoch": 1.
|
| 447 |
-
"learning_rate":
|
| 448 |
-
"loss": 0.
|
| 449 |
"step": 720
|
| 450 |
},
|
| 451 |
{
|
| 452 |
-
"epoch": 1.
|
| 453 |
-
"learning_rate":
|
| 454 |
-
"loss": 0.
|
| 455 |
"step": 730
|
| 456 |
},
|
| 457 |
{
|
| 458 |
-
"epoch": 1.
|
| 459 |
-
"learning_rate": 3.
|
| 460 |
-
"loss": 0.
|
| 461 |
"step": 740
|
| 462 |
},
|
| 463 |
{
|
| 464 |
-
"epoch": 1.
|
| 465 |
-
"learning_rate": 3.
|
| 466 |
-
"loss": 0.
|
| 467 |
"step": 750
|
| 468 |
},
|
| 469 |
{
|
| 470 |
-
"epoch": 1.
|
| 471 |
-
"learning_rate": 3.
|
| 472 |
-
"loss": 0.
|
| 473 |
"step": 760
|
| 474 |
},
|
| 475 |
{
|
| 476 |
-
"epoch": 1.
|
| 477 |
-
"learning_rate": 3.
|
| 478 |
-
"loss": 0.
|
| 479 |
"step": 770
|
| 480 |
},
|
| 481 |
{
|
| 482 |
-
"epoch": 1.
|
| 483 |
-
"learning_rate": 3.
|
| 484 |
-
"loss": 0.
|
| 485 |
"step": 780
|
| 486 |
},
|
| 487 |
{
|
| 488 |
-
"epoch": 1.
|
| 489 |
-
"learning_rate": 3.
|
| 490 |
-
"loss": 0.
|
| 491 |
"step": 790
|
| 492 |
},
|
| 493 |
{
|
| 494 |
-
"epoch": 1.
|
| 495 |
-
"learning_rate": 3.
|
| 496 |
-
"loss": 0.
|
| 497 |
"step": 800
|
| 498 |
},
|
| 499 |
{
|
| 500 |
-
"epoch": 1.
|
| 501 |
-
"learning_rate": 3.
|
| 502 |
-
"loss": 0.
|
| 503 |
"step": 810
|
| 504 |
},
|
| 505 |
{
|
| 506 |
-
"epoch": 1.
|
| 507 |
-
"learning_rate": 3.
|
| 508 |
-
"loss": 0.
|
| 509 |
"step": 820
|
| 510 |
},
|
| 511 |
{
|
| 512 |
-
"epoch": 1.
|
| 513 |
-
"learning_rate": 3.
|
| 514 |
-
"loss": 0.
|
| 515 |
"step": 830
|
| 516 |
},
|
| 517 |
{
|
| 518 |
-
"epoch": 1.
|
| 519 |
-
"learning_rate": 3.
|
| 520 |
-
"loss": 0.
|
| 521 |
"step": 840
|
| 522 |
},
|
| 523 |
{
|
| 524 |
-
"epoch": 1.
|
| 525 |
-
"learning_rate": 3.
|
| 526 |
-
"loss": 0.
|
| 527 |
"step": 850
|
| 528 |
},
|
| 529 |
{
|
| 530 |
-
"epoch": 1.
|
| 531 |
-
"learning_rate": 3.
|
| 532 |
-
"loss": 0.
|
| 533 |
"step": 860
|
| 534 |
},
|
| 535 |
{
|
| 536 |
-
"epoch": 1.
|
| 537 |
-
"learning_rate": 3.
|
| 538 |
-
"loss": 0.
|
| 539 |
"step": 870
|
| 540 |
},
|
| 541 |
{
|
| 542 |
-
"epoch": 1.
|
| 543 |
-
"learning_rate": 3.
|
| 544 |
-
"loss": 0.
|
| 545 |
"step": 880
|
| 546 |
},
|
| 547 |
{
|
| 548 |
-
"epoch": 1.
|
| 549 |
-
"learning_rate": 3.
|
| 550 |
-
"loss": 0.
|
| 551 |
"step": 890
|
| 552 |
},
|
| 553 |
{
|
| 554 |
-
"epoch": 1.
|
| 555 |
-
"learning_rate": 3.
|
| 556 |
-
"loss": 0.
|
| 557 |
"step": 900
|
| 558 |
},
|
| 559 |
{
|
| 560 |
-
"epoch": 1.
|
| 561 |
-
"learning_rate": 3.
|
| 562 |
-
"loss": 0.
|
| 563 |
"step": 910
|
| 564 |
},
|
| 565 |
{
|
| 566 |
-
"epoch": 1.
|
| 567 |
-
"learning_rate":
|
| 568 |
-
"loss": 0.
|
| 569 |
"step": 920
|
| 570 |
},
|
| 571 |
{
|
| 572 |
-
"epoch": 1.
|
| 573 |
-
"learning_rate":
|
| 574 |
-
"loss": 0.
|
| 575 |
"step": 930
|
| 576 |
},
|
| 577 |
{
|
| 578 |
-
"epoch": 1.
|
| 579 |
-
"learning_rate":
|
| 580 |
-
"loss": 0.
|
| 581 |
"step": 940
|
| 582 |
},
|
| 583 |
{
|
| 584 |
-
"epoch": 1.
|
| 585 |
-
"learning_rate":
|
| 586 |
-
"loss": 0.
|
| 587 |
"step": 950
|
| 588 |
},
|
| 589 |
{
|
| 590 |
-
"epoch": 1.
|
| 591 |
-
"learning_rate":
|
| 592 |
-
"loss": 0.
|
| 593 |
"step": 960
|
| 594 |
},
|
| 595 |
{
|
| 596 |
-
"epoch": 1.
|
| 597 |
-
"learning_rate":
|
| 598 |
-
"loss": 0.
|
| 599 |
"step": 970
|
| 600 |
},
|
| 601 |
{
|
| 602 |
-
"epoch": 1.
|
| 603 |
-
"learning_rate":
|
| 604 |
-
"loss": 0.
|
| 605 |
"step": 980
|
| 606 |
},
|
| 607 |
{
|
| 608 |
-
"epoch": 1.
|
| 609 |
-
"learning_rate":
|
| 610 |
-
"loss": 0.
|
| 611 |
"step": 990
|
| 612 |
},
|
| 613 |
{
|
| 614 |
-
"epoch": 1.
|
| 615 |
-
"learning_rate":
|
| 616 |
-
"loss": 0.
|
| 617 |
"step": 1000
|
| 618 |
},
|
| 619 |
{
|
| 620 |
-
"epoch": 1.
|
| 621 |
-
"learning_rate":
|
| 622 |
-
"loss": 0.
|
| 623 |
"step": 1010
|
| 624 |
},
|
| 625 |
{
|
| 626 |
-
"epoch": 1.
|
| 627 |
-
"learning_rate":
|
| 628 |
-
"loss": 0.
|
| 629 |
"step": 1020
|
| 630 |
},
|
| 631 |
{
|
| 632 |
-
"epoch": 1.
|
| 633 |
-
"learning_rate":
|
| 634 |
-
"loss": 0.
|
| 635 |
"step": 1030
|
| 636 |
},
|
| 637 |
{
|
| 638 |
-
"epoch": 1.
|
| 639 |
-
"learning_rate":
|
| 640 |
-
"loss": 0.
|
| 641 |
"step": 1040
|
| 642 |
},
|
| 643 |
{
|
| 644 |
-
"epoch": 1.
|
| 645 |
-
"learning_rate":
|
| 646 |
-
"loss": 0.
|
| 647 |
"step": 1050
|
| 648 |
},
|
| 649 |
{
|
| 650 |
-
"epoch":
|
| 651 |
-
"
|
| 652 |
-
"
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
| 653 |
"step": 1060
|
| 654 |
},
|
| 655 |
{
|
| 656 |
-
"epoch":
|
| 657 |
-
"learning_rate":
|
| 658 |
-
"loss": 0.
|
| 659 |
"step": 1070
|
| 660 |
},
|
| 661 |
{
|
| 662 |
-
"epoch":
|
| 663 |
-
"learning_rate":
|
| 664 |
-
"loss": 0.
|
| 665 |
"step": 1080
|
| 666 |
},
|
| 667 |
{
|
| 668 |
-
"epoch":
|
| 669 |
-
"learning_rate":
|
| 670 |
-
"loss": 0.
|
| 671 |
"step": 1090
|
| 672 |
},
|
| 673 |
{
|
| 674 |
-
"epoch":
|
| 675 |
-
"learning_rate":
|
| 676 |
-
"loss": 0.
|
| 677 |
"step": 1100
|
| 678 |
},
|
| 679 |
{
|
| 680 |
-
"epoch":
|
| 681 |
-
"learning_rate":
|
| 682 |
-
"loss": 0.
|
| 683 |
"step": 1110
|
| 684 |
},
|
| 685 |
{
|
| 686 |
-
"epoch":
|
| 687 |
-
"learning_rate":
|
| 688 |
-
"loss": 0.
|
| 689 |
"step": 1120
|
| 690 |
},
|
| 691 |
{
|
| 692 |
-
"epoch":
|
| 693 |
-
"learning_rate":
|
| 694 |
-
"loss": 0.
|
| 695 |
"step": 1130
|
| 696 |
},
|
| 697 |
{
|
| 698 |
-
"epoch":
|
| 699 |
-
"learning_rate":
|
| 700 |
-
"loss": 0.
|
| 701 |
"step": 1140
|
| 702 |
},
|
| 703 |
{
|
| 704 |
-
"epoch":
|
| 705 |
-
"learning_rate":
|
| 706 |
-
"loss": 0.
|
| 707 |
"step": 1150
|
| 708 |
},
|
| 709 |
{
|
| 710 |
-
"epoch":
|
| 711 |
-
"learning_rate":
|
| 712 |
-
"loss": 0.
|
| 713 |
"step": 1160
|
| 714 |
},
|
| 715 |
{
|
| 716 |
-
"epoch":
|
| 717 |
-
"learning_rate":
|
| 718 |
-
"loss": 0.
|
| 719 |
"step": 1170
|
| 720 |
},
|
| 721 |
{
|
| 722 |
-
"epoch":
|
| 723 |
-
"learning_rate":
|
| 724 |
-
"loss": 0.
|
| 725 |
"step": 1180
|
| 726 |
},
|
| 727 |
{
|
| 728 |
-
"epoch":
|
| 729 |
-
"learning_rate":
|
| 730 |
-
"loss": 0.
|
| 731 |
"step": 1190
|
| 732 |
},
|
| 733 |
{
|
| 734 |
-
"epoch":
|
| 735 |
-
"learning_rate":
|
| 736 |
-
"loss": 0.
|
| 737 |
"step": 1200
|
| 738 |
},
|
| 739 |
{
|
| 740 |
-
"epoch":
|
| 741 |
-
"learning_rate":
|
| 742 |
-
"loss": 0.
|
| 743 |
"step": 1210
|
| 744 |
},
|
| 745 |
{
|
| 746 |
-
"epoch":
|
| 747 |
-
"learning_rate": 2.
|
| 748 |
-
"loss": 0.
|
| 749 |
"step": 1220
|
| 750 |
},
|
| 751 |
{
|
| 752 |
-
"epoch":
|
| 753 |
-
"learning_rate": 2.
|
| 754 |
-
"loss": 0.
|
| 755 |
"step": 1230
|
| 756 |
},
|
| 757 |
{
|
| 758 |
-
"epoch":
|
| 759 |
-
"learning_rate": 2.
|
| 760 |
-
"loss": 0.
|
| 761 |
"step": 1240
|
| 762 |
},
|
| 763 |
{
|
| 764 |
-
"epoch":
|
| 765 |
-
"learning_rate": 2.
|
| 766 |
-
"loss": 0.
|
| 767 |
"step": 1250
|
| 768 |
},
|
| 769 |
{
|
| 770 |
-
"epoch":
|
| 771 |
-
"learning_rate": 2.
|
| 772 |
-
"loss": 0.
|
| 773 |
"step": 1260
|
| 774 |
},
|
| 775 |
{
|
| 776 |
-
"epoch":
|
| 777 |
-
"learning_rate":
|
| 778 |
-
"loss": 0.
|
| 779 |
"step": 1270
|
| 780 |
},
|
| 781 |
{
|
| 782 |
-
"epoch":
|
| 783 |
-
"learning_rate":
|
| 784 |
-
"loss": 0.
|
| 785 |
"step": 1280
|
| 786 |
},
|
| 787 |
{
|
| 788 |
-
"epoch":
|
| 789 |
-
"learning_rate":
|
| 790 |
-
"loss": 0.
|
| 791 |
"step": 1290
|
| 792 |
},
|
| 793 |
{
|
| 794 |
-
"epoch":
|
| 795 |
-
"learning_rate":
|
| 796 |
-
"loss": 0.
|
| 797 |
"step": 1300
|
| 798 |
},
|
| 799 |
{
|
| 800 |
-
"epoch":
|
| 801 |
-
"learning_rate":
|
| 802 |
-
"loss": 0.
|
| 803 |
"step": 1310
|
| 804 |
},
|
| 805 |
{
|
| 806 |
-
"epoch": 2.
|
| 807 |
-
"learning_rate":
|
| 808 |
-
"loss": 0.
|
| 809 |
"step": 1320
|
| 810 |
},
|
| 811 |
{
|
| 812 |
-
"epoch": 2.
|
| 813 |
-
"
|
| 814 |
-
"
|
| 815 |
-
"eval_runtime": 121.8002,
|
| 816 |
-
"eval_samples_per_second": 77.249,
|
| 817 |
-
"eval_steps_per_second": 2.422,
|
| 818 |
-
"step": 1322
|
| 819 |
-
},
|
| 820 |
-
{
|
| 821 |
-
"epoch": 2.01,
|
| 822 |
-
"learning_rate": 1.8301569506726456e-05,
|
| 823 |
-
"loss": 0.3002,
|
| 824 |
"step": 1330
|
| 825 |
},
|
| 826 |
{
|
| 827 |
-
"epoch": 2.
|
| 828 |
-
"learning_rate":
|
| 829 |
-
"loss": 0.
|
| 830 |
"step": 1340
|
| 831 |
},
|
| 832 |
{
|
| 833 |
-
"epoch": 2.
|
| 834 |
-
"learning_rate":
|
| 835 |
-
"loss": 0.
|
| 836 |
"step": 1350
|
| 837 |
},
|
| 838 |
{
|
| 839 |
-
"epoch": 2.
|
| 840 |
-
"learning_rate":
|
| 841 |
-
"loss": 0.
|
| 842 |
"step": 1360
|
| 843 |
},
|
| 844 |
{
|
| 845 |
-
"epoch": 2.
|
| 846 |
-
"learning_rate":
|
| 847 |
-
"loss": 0.
|
| 848 |
"step": 1370
|
| 849 |
},
|
| 850 |
{
|
| 851 |
-
"epoch": 2.
|
| 852 |
-
"learning_rate":
|
| 853 |
-
"loss": 0.
|
| 854 |
"step": 1380
|
| 855 |
},
|
| 856 |
{
|
| 857 |
-
"epoch": 2.
|
| 858 |
-
"learning_rate":
|
| 859 |
-
"loss": 0.
|
| 860 |
"step": 1390
|
| 861 |
},
|
| 862 |
{
|
| 863 |
-
"epoch": 2.
|
| 864 |
-
"learning_rate":
|
| 865 |
-
"loss": 0.
|
| 866 |
"step": 1400
|
| 867 |
},
|
| 868 |
{
|
| 869 |
-
"epoch": 2.
|
| 870 |
-
"learning_rate":
|
| 871 |
-
"loss": 0.
|
| 872 |
"step": 1410
|
| 873 |
},
|
| 874 |
{
|
| 875 |
-
"epoch": 2.
|
| 876 |
-
"learning_rate":
|
| 877 |
-
"loss": 0.
|
| 878 |
"step": 1420
|
| 879 |
},
|
| 880 |
{
|
| 881 |
-
"epoch": 2.
|
| 882 |
-
"learning_rate":
|
| 883 |
-
"loss": 0.
|
| 884 |
"step": 1430
|
| 885 |
},
|
| 886 |
{
|
| 887 |
-
"epoch": 2.
|
| 888 |
-
"learning_rate":
|
| 889 |
-
"loss": 0.
|
| 890 |
"step": 1440
|
| 891 |
},
|
| 892 |
{
|
| 893 |
-
"epoch": 2.
|
| 894 |
-
"learning_rate":
|
| 895 |
-
"loss": 0.
|
| 896 |
"step": 1450
|
| 897 |
},
|
| 898 |
{
|
| 899 |
-
"epoch": 2.
|
| 900 |
-
"learning_rate":
|
| 901 |
-
"loss": 0.
|
| 902 |
"step": 1460
|
| 903 |
},
|
| 904 |
{
|
| 905 |
-
"epoch": 2.
|
| 906 |
-
"learning_rate":
|
| 907 |
-
"loss": 0.
|
| 908 |
"step": 1470
|
| 909 |
},
|
| 910 |
{
|
| 911 |
-
"epoch": 2.
|
| 912 |
-
"learning_rate":
|
| 913 |
-
"loss": 0.
|
| 914 |
"step": 1480
|
| 915 |
},
|
| 916 |
{
|
| 917 |
-
"epoch": 2.
|
| 918 |
-
"learning_rate":
|
| 919 |
-
"loss": 0.
|
| 920 |
"step": 1490
|
| 921 |
},
|
| 922 |
{
|
| 923 |
-
"epoch": 2.
|
| 924 |
-
"learning_rate":
|
| 925 |
-
"loss": 0.
|
| 926 |
"step": 1500
|
| 927 |
},
|
| 928 |
{
|
| 929 |
-
"epoch": 2.
|
| 930 |
-
"learning_rate":
|
| 931 |
-
"loss": 0.
|
| 932 |
"step": 1510
|
| 933 |
},
|
| 934 |
{
|
| 935 |
-
"epoch": 2.
|
| 936 |
-
"learning_rate":
|
| 937 |
-
"loss": 0.
|
| 938 |
"step": 1520
|
| 939 |
},
|
| 940 |
{
|
| 941 |
-
"epoch": 2.
|
| 942 |
-
"learning_rate":
|
| 943 |
-
"loss": 0.
|
| 944 |
"step": 1530
|
| 945 |
},
|
| 946 |
{
|
| 947 |
-
"epoch": 2.
|
| 948 |
-
"learning_rate":
|
| 949 |
-
"loss": 0.
|
| 950 |
"step": 1540
|
| 951 |
},
|
| 952 |
{
|
| 953 |
-
"epoch": 2.
|
| 954 |
-
"learning_rate":
|
| 955 |
-
"loss": 0.
|
| 956 |
"step": 1550
|
| 957 |
},
|
| 958 |
{
|
| 959 |
-
"epoch": 2.
|
| 960 |
-
"learning_rate":
|
| 961 |
-
"loss": 0.
|
| 962 |
"step": 1560
|
| 963 |
},
|
| 964 |
{
|
| 965 |
-
"epoch": 2.
|
| 966 |
-
"learning_rate":
|
| 967 |
-
"loss": 0.
|
| 968 |
"step": 1570
|
| 969 |
},
|
| 970 |
{
|
| 971 |
-
"epoch":
|
| 972 |
-
"learning_rate":
|
| 973 |
-
"loss": 0.
|
| 974 |
"step": 1580
|
| 975 |
},
|
| 976 |
{
|
| 977 |
-
"epoch":
|
| 978 |
-
"
|
| 979 |
-
"
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
| 980 |
"step": 1590
|
| 981 |
},
|
| 982 |
{
|
| 983 |
-
"epoch":
|
| 984 |
-
"learning_rate":
|
| 985 |
-
"loss": 0.
|
| 986 |
"step": 1600
|
| 987 |
},
|
| 988 |
{
|
| 989 |
-
"epoch":
|
| 990 |
-
"learning_rate":
|
| 991 |
-
"loss": 0.
|
| 992 |
"step": 1610
|
| 993 |
},
|
| 994 |
{
|
| 995 |
-
"epoch":
|
| 996 |
-
"learning_rate":
|
| 997 |
-
"loss": 0.
|
| 998 |
"step": 1620
|
| 999 |
},
|
| 1000 |
{
|
| 1001 |
-
"epoch":
|
| 1002 |
-
"learning_rate":
|
| 1003 |
-
"loss": 0.
|
| 1004 |
"step": 1630
|
| 1005 |
},
|
| 1006 |
{
|
| 1007 |
-
"epoch":
|
| 1008 |
-
"learning_rate":
|
| 1009 |
-
"loss": 0.
|
| 1010 |
"step": 1640
|
| 1011 |
},
|
| 1012 |
{
|
| 1013 |
-
"epoch":
|
| 1014 |
-
"learning_rate":
|
| 1015 |
-
"loss": 0.
|
| 1016 |
"step": 1650
|
| 1017 |
},
|
| 1018 |
{
|
| 1019 |
-
"epoch":
|
| 1020 |
-
"learning_rate":
|
| 1021 |
-
"loss": 0.
|
| 1022 |
"step": 1660
|
| 1023 |
},
|
| 1024 |
{
|
| 1025 |
-
"epoch":
|
| 1026 |
-
"learning_rate":
|
| 1027 |
-
"loss": 0.
|
| 1028 |
"step": 1670
|
| 1029 |
},
|
| 1030 |
{
|
| 1031 |
-
"epoch":
|
| 1032 |
-
"learning_rate":
|
| 1033 |
-
"loss": 0.
|
| 1034 |
"step": 1680
|
| 1035 |
},
|
| 1036 |
{
|
| 1037 |
-
"epoch":
|
| 1038 |
-
"learning_rate":
|
| 1039 |
-
"loss": 0.
|
| 1040 |
"step": 1690
|
| 1041 |
},
|
| 1042 |
{
|
| 1043 |
-
"epoch":
|
| 1044 |
-
"learning_rate":
|
| 1045 |
-
"loss": 0.
|
| 1046 |
"step": 1700
|
| 1047 |
},
|
| 1048 |
{
|
| 1049 |
-
"epoch":
|
| 1050 |
-
"learning_rate":
|
| 1051 |
-
"loss": 0.
|
| 1052 |
"step": 1710
|
| 1053 |
},
|
| 1054 |
{
|
| 1055 |
-
"epoch":
|
| 1056 |
-
"learning_rate":
|
| 1057 |
-
"loss": 0.
|
| 1058 |
"step": 1720
|
| 1059 |
},
|
| 1060 |
{
|
| 1061 |
-
"epoch":
|
| 1062 |
-
"learning_rate":
|
| 1063 |
-
"loss": 0.
|
| 1064 |
"step": 1730
|
| 1065 |
},
|
| 1066 |
{
|
| 1067 |
-
"epoch":
|
| 1068 |
-
"learning_rate":
|
| 1069 |
-
"loss": 0.
|
| 1070 |
"step": 1740
|
| 1071 |
},
|
| 1072 |
{
|
| 1073 |
-
"epoch":
|
| 1074 |
-
"learning_rate":
|
| 1075 |
"loss": 0.2409,
|
| 1076 |
"step": 1750
|
| 1077 |
},
|
| 1078 |
{
|
| 1079 |
-
"epoch":
|
| 1080 |
-
"learning_rate":
|
| 1081 |
-
"loss": 0.
|
| 1082 |
"step": 1760
|
| 1083 |
},
|
| 1084 |
{
|
| 1085 |
-
"epoch":
|
| 1086 |
-
"learning_rate":
|
| 1087 |
-
"loss": 0.
|
| 1088 |
"step": 1770
|
| 1089 |
},
|
| 1090 |
{
|
| 1091 |
-
"epoch":
|
| 1092 |
-
"learning_rate":
|
| 1093 |
-
"loss": 0.
|
| 1094 |
"step": 1780
|
| 1095 |
},
|
| 1096 |
{
|
| 1097 |
-
"epoch":
|
| 1098 |
-
"learning_rate":
|
| 1099 |
-
"loss": 0.
|
| 1100 |
"step": 1790
|
| 1101 |
},
|
| 1102 |
{
|
| 1103 |
-
"epoch":
|
| 1104 |
-
"learning_rate":
|
| 1105 |
-
"loss": 0.
|
| 1106 |
"step": 1800
|
| 1107 |
},
|
| 1108 |
{
|
| 1109 |
-
"epoch":
|
| 1110 |
-
"learning_rate":
|
| 1111 |
-
"loss": 0.
|
| 1112 |
"step": 1810
|
| 1113 |
},
|
| 1114 |
{
|
| 1115 |
-
"epoch":
|
| 1116 |
-
"learning_rate":
|
| 1117 |
-
"loss": 0.
|
| 1118 |
"step": 1820
|
| 1119 |
},
|
| 1120 |
{
|
| 1121 |
-
"epoch":
|
| 1122 |
-
"learning_rate":
|
| 1123 |
-
"loss": 0.
|
| 1124 |
"step": 1830
|
| 1125 |
},
|
| 1126 |
{
|
| 1127 |
-
"epoch":
|
| 1128 |
-
"learning_rate":
|
| 1129 |
-
"loss": 0.
|
| 1130 |
"step": 1840
|
| 1131 |
},
|
| 1132 |
{
|
| 1133 |
-
"epoch":
|
| 1134 |
-
"learning_rate":
|
| 1135 |
-
"loss": 0.
|
| 1136 |
"step": 1850
|
| 1137 |
},
|
| 1138 |
{
|
| 1139 |
-
"epoch":
|
| 1140 |
-
"learning_rate":
|
| 1141 |
-
"loss": 0.
|
| 1142 |
"step": 1860
|
| 1143 |
},
|
| 1144 |
{
|
| 1145 |
-
"epoch":
|
| 1146 |
-
"learning_rate":
|
| 1147 |
-
"loss": 0.
|
| 1148 |
"step": 1870
|
| 1149 |
},
|
| 1150 |
{
|
| 1151 |
-
"epoch":
|
| 1152 |
-
"learning_rate":
|
| 1153 |
-
"loss": 0.
|
| 1154 |
"step": 1880
|
| 1155 |
},
|
| 1156 |
{
|
| 1157 |
-
"epoch":
|
| 1158 |
-
"learning_rate":
|
| 1159 |
-
"loss": 0.
|
| 1160 |
"step": 1890
|
| 1161 |
},
|
| 1162 |
{
|
| 1163 |
-
"epoch":
|
| 1164 |
-
"learning_rate":
|
| 1165 |
-
"loss": 0.
|
| 1166 |
"step": 1900
|
| 1167 |
},
|
| 1168 |
{
|
| 1169 |
-
"epoch":
|
| 1170 |
-
"learning_rate":
|
| 1171 |
-
"loss": 0.
|
| 1172 |
"step": 1910
|
| 1173 |
},
|
| 1174 |
{
|
| 1175 |
-
"epoch":
|
| 1176 |
-
"learning_rate": 1.
|
| 1177 |
-
"loss": 0.
|
| 1178 |
"step": 1920
|
| 1179 |
},
|
| 1180 |
{
|
| 1181 |
-
"epoch":
|
| 1182 |
-
"learning_rate": 1.
|
| 1183 |
-
"loss": 0.
|
| 1184 |
"step": 1930
|
| 1185 |
},
|
| 1186 |
{
|
| 1187 |
-
"epoch":
|
| 1188 |
-
"learning_rate": 1.
|
| 1189 |
-
"loss": 0.
|
| 1190 |
"step": 1940
|
| 1191 |
},
|
| 1192 |
{
|
| 1193 |
-
"epoch":
|
| 1194 |
-
"learning_rate":
|
| 1195 |
-
"loss": 0.
|
| 1196 |
"step": 1950
|
| 1197 |
},
|
| 1198 |
{
|
| 1199 |
-
"epoch":
|
| 1200 |
-
"learning_rate":
|
| 1201 |
-
"loss": 0.
|
| 1202 |
"step": 1960
|
| 1203 |
},
|
| 1204 |
{
|
| 1205 |
-
"epoch":
|
| 1206 |
-
"learning_rate":
|
| 1207 |
-
"loss": 0.
|
| 1208 |
"step": 1970
|
| 1209 |
},
|
| 1210 |
{
|
| 1211 |
-
"epoch":
|
| 1212 |
-
"learning_rate":
|
| 1213 |
-
"loss": 0.
|
| 1214 |
"step": 1980
|
| 1215 |
},
|
| 1216 |
{
|
| 1217 |
-
"epoch": 3.
|
| 1218 |
-
"
|
| 1219 |
-
"
|
| 1220 |
-
"
|
| 1221 |
-
"eval_samples_per_second": 75.387,
|
| 1222 |
-
"eval_steps_per_second": 2.364,
|
| 1223 |
-
"step": 1983
|
| 1224 |
},
|
| 1225 |
{
|
| 1226 |
-
"epoch": 3.
|
| 1227 |
-
"
|
| 1228 |
-
"
|
| 1229 |
-
"
|
| 1230 |
-
|
| 1231 |
-
|
| 1232 |
-
"
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
| 1233 |
}
|
| 1234 |
],
|
| 1235 |
-
"max_steps":
|
| 1236 |
-
"num_train_epochs":
|
| 1237 |
-
"total_flos":
|
| 1238 |
"trial_name": null,
|
| 1239 |
"trial_params": null
|
| 1240 |
}
|
|
|
|
| 1 |
{
|
| 2 |
+
"best_metric": 0.9741915422885572,
|
| 3 |
+
"best_model_checkpoint": "swin-tiny-patch4-window7-224-finetuned-eurosat/checkpoint-2635",
|
| 4 |
+
"epoch": 4.99857887257224,
|
| 5 |
+
"global_step": 2635,
|
| 6 |
"is_hyper_param_search": false,
|
| 7 |
"is_local_process_zero": true,
|
| 8 |
"is_world_process_zero": true,
|
| 9 |
"log_history": [
|
| 10 |
{
|
| 11 |
"epoch": 0.02,
|
| 12 |
+
"learning_rate": 1.8939393939393941e-06,
|
| 13 |
+
"loss": 1.4619,
|
| 14 |
"step": 10
|
| 15 |
},
|
| 16 |
{
|
| 17 |
+
"epoch": 0.04,
|
| 18 |
+
"learning_rate": 3.7878787878787882e-06,
|
| 19 |
+
"loss": 1.3888,
|
| 20 |
"step": 20
|
| 21 |
},
|
| 22 |
{
|
| 23 |
+
"epoch": 0.06,
|
| 24 |
+
"learning_rate": 5.681818181818182e-06,
|
| 25 |
+
"loss": 1.2545,
|
| 26 |
"step": 30
|
| 27 |
},
|
| 28 |
{
|
| 29 |
+
"epoch": 0.08,
|
| 30 |
+
"learning_rate": 7.5757575757575764e-06,
|
| 31 |
+
"loss": 1.1428,
|
| 32 |
"step": 40
|
| 33 |
},
|
| 34 |
{
|
| 35 |
+
"epoch": 0.09,
|
| 36 |
+
"learning_rate": 9.46969696969697e-06,
|
| 37 |
+
"loss": 0.9846,
|
| 38 |
"step": 50
|
| 39 |
},
|
| 40 |
{
|
| 41 |
+
"epoch": 0.11,
|
| 42 |
+
"learning_rate": 1.1363636363636365e-05,
|
| 43 |
+
"loss": 0.8566,
|
| 44 |
"step": 60
|
| 45 |
},
|
| 46 |
{
|
| 47 |
+
"epoch": 0.13,
|
| 48 |
+
"learning_rate": 1.3257575757575758e-05,
|
| 49 |
+
"loss": 0.8184,
|
| 50 |
"step": 70
|
| 51 |
},
|
| 52 |
{
|
| 53 |
+
"epoch": 0.15,
|
| 54 |
+
"learning_rate": 1.5151515151515153e-05,
|
| 55 |
+
"loss": 0.8055,
|
| 56 |
"step": 80
|
| 57 |
},
|
| 58 |
{
|
| 59 |
+
"epoch": 0.17,
|
| 60 |
+
"learning_rate": 1.7045454545454546e-05,
|
| 61 |
+
"loss": 0.7846,
|
| 62 |
"step": 90
|
| 63 |
},
|
| 64 |
{
|
| 65 |
+
"epoch": 0.19,
|
| 66 |
+
"learning_rate": 1.893939393939394e-05,
|
| 67 |
+
"loss": 0.7953,
|
| 68 |
"step": 100
|
| 69 |
},
|
| 70 |
{
|
| 71 |
+
"epoch": 0.21,
|
| 72 |
+
"learning_rate": 2.0833333333333336e-05,
|
| 73 |
+
"loss": 0.7205,
|
| 74 |
"step": 110
|
| 75 |
},
|
| 76 |
{
|
| 77 |
+
"epoch": 0.23,
|
| 78 |
+
"learning_rate": 2.272727272727273e-05,
|
| 79 |
+
"loss": 0.699,
|
| 80 |
"step": 120
|
| 81 |
},
|
| 82 |
{
|
| 83 |
+
"epoch": 0.25,
|
| 84 |
+
"learning_rate": 2.4621212121212123e-05,
|
| 85 |
+
"loss": 0.698,
|
| 86 |
"step": 130
|
| 87 |
},
|
| 88 |
{
|
| 89 |
+
"epoch": 0.27,
|
| 90 |
+
"learning_rate": 2.6515151515151516e-05,
|
| 91 |
+
"loss": 0.7095,
|
| 92 |
"step": 140
|
| 93 |
},
|
| 94 |
{
|
| 95 |
+
"epoch": 0.28,
|
| 96 |
+
"learning_rate": 2.8409090909090912e-05,
|
| 97 |
+
"loss": 0.6753,
|
| 98 |
"step": 150
|
| 99 |
},
|
| 100 |
{
|
| 101 |
+
"epoch": 0.3,
|
| 102 |
+
"learning_rate": 3.0303030303030306e-05,
|
| 103 |
+
"loss": 0.6655,
|
| 104 |
"step": 160
|
| 105 |
},
|
| 106 |
{
|
| 107 |
+
"epoch": 0.32,
|
| 108 |
+
"learning_rate": 3.2196969696969696e-05,
|
| 109 |
+
"loss": 0.6625,
|
| 110 |
"step": 170
|
| 111 |
},
|
| 112 |
{
|
| 113 |
+
"epoch": 0.34,
|
| 114 |
+
"learning_rate": 3.409090909090909e-05,
|
| 115 |
+
"loss": 0.6881,
|
| 116 |
"step": 180
|
| 117 |
},
|
| 118 |
{
|
| 119 |
+
"epoch": 0.36,
|
| 120 |
+
"learning_rate": 3.598484848484849e-05,
|
| 121 |
+
"loss": 0.6698,
|
| 122 |
"step": 190
|
| 123 |
},
|
| 124 |
{
|
| 125 |
+
"epoch": 0.38,
|
| 126 |
+
"learning_rate": 3.787878787878788e-05,
|
| 127 |
+
"loss": 0.6548,
|
| 128 |
"step": 200
|
| 129 |
},
|
| 130 |
{
|
| 131 |
+
"epoch": 0.4,
|
| 132 |
+
"learning_rate": 3.9772727272727275e-05,
|
| 133 |
+
"loss": 0.6373,
|
| 134 |
"step": 210
|
| 135 |
},
|
| 136 |
{
|
| 137 |
+
"epoch": 0.42,
|
| 138 |
+
"learning_rate": 4.166666666666667e-05,
|
| 139 |
+
"loss": 0.6437,
|
| 140 |
"step": 220
|
| 141 |
},
|
| 142 |
{
|
| 143 |
+
"epoch": 0.44,
|
| 144 |
+
"learning_rate": 4.356060606060606e-05,
|
| 145 |
+
"loss": 0.6308,
|
| 146 |
"step": 230
|
| 147 |
},
|
| 148 |
{
|
| 149 |
+
"epoch": 0.45,
|
| 150 |
+
"learning_rate": 4.545454545454546e-05,
|
| 151 |
+
"loss": 0.6119,
|
| 152 |
"step": 240
|
| 153 |
},
|
| 154 |
{
|
| 155 |
+
"epoch": 0.47,
|
| 156 |
+
"learning_rate": 4.7348484848484855e-05,
|
| 157 |
+
"loss": 0.5998,
|
| 158 |
"step": 250
|
| 159 |
},
|
| 160 |
{
|
| 161 |
+
"epoch": 0.49,
|
| 162 |
+
"learning_rate": 4.9242424242424245e-05,
|
| 163 |
+
"loss": 0.6188,
|
| 164 |
"step": 260
|
| 165 |
},
|
| 166 |
{
|
| 167 |
+
"epoch": 0.51,
|
| 168 |
+
"learning_rate": 4.987347110923661e-05,
|
| 169 |
+
"loss": 0.5837,
|
| 170 |
"step": 270
|
| 171 |
},
|
| 172 |
{
|
| 173 |
+
"epoch": 0.53,
|
| 174 |
+
"learning_rate": 4.966258962463096e-05,
|
| 175 |
+
"loss": 0.585,
|
| 176 |
"step": 280
|
| 177 |
},
|
| 178 |
{
|
| 179 |
+
"epoch": 0.55,
|
| 180 |
+
"learning_rate": 4.945170814002531e-05,
|
| 181 |
+
"loss": 0.6001,
|
| 182 |
"step": 290
|
| 183 |
},
|
| 184 |
{
|
| 185 |
+
"epoch": 0.57,
|
| 186 |
+
"learning_rate": 4.924082665541966e-05,
|
| 187 |
+
"loss": 0.5664,
|
| 188 |
"step": 300
|
| 189 |
},
|
| 190 |
{
|
| 191 |
+
"epoch": 0.59,
|
| 192 |
+
"learning_rate": 4.9029945170814005e-05,
|
| 193 |
+
"loss": 0.5603,
|
| 194 |
"step": 310
|
| 195 |
},
|
| 196 |
{
|
| 197 |
+
"epoch": 0.61,
|
| 198 |
+
"learning_rate": 4.8819063686208354e-05,
|
| 199 |
+
"loss": 0.601,
|
| 200 |
"step": 320
|
| 201 |
},
|
| 202 |
{
|
| 203 |
+
"epoch": 0.63,
|
| 204 |
+
"learning_rate": 4.86081822016027e-05,
|
| 205 |
+
"loss": 0.5489,
|
| 206 |
"step": 330
|
| 207 |
},
|
| 208 |
{
|
| 209 |
+
"epoch": 0.64,
|
| 210 |
+
"learning_rate": 4.839730071699705e-05,
|
| 211 |
+
"loss": 0.5525,
|
| 212 |
"step": 340
|
| 213 |
},
|
| 214 |
{
|
| 215 |
+
"epoch": 0.66,
|
| 216 |
+
"learning_rate": 4.81864192323914e-05,
|
| 217 |
+
"loss": 0.5451,
|
| 218 |
"step": 350
|
| 219 |
},
|
| 220 |
{
|
| 221 |
+
"epoch": 0.68,
|
| 222 |
+
"learning_rate": 4.797553774778575e-05,
|
| 223 |
+
"loss": 0.5515,
|
| 224 |
"step": 360
|
| 225 |
},
|
| 226 |
{
|
| 227 |
+
"epoch": 0.7,
|
| 228 |
+
"learning_rate": 4.7764656263180096e-05,
|
| 229 |
+
"loss": 0.4817,
|
| 230 |
"step": 370
|
| 231 |
},
|
| 232 |
{
|
| 233 |
+
"epoch": 0.72,
|
| 234 |
+
"learning_rate": 4.7553774778574444e-05,
|
| 235 |
+
"loss": 0.5398,
|
| 236 |
"step": 380
|
| 237 |
},
|
| 238 |
{
|
| 239 |
+
"epoch": 0.74,
|
| 240 |
+
"learning_rate": 4.734289329396879e-05,
|
| 241 |
+
"loss": 0.4987,
|
| 242 |
"step": 390
|
| 243 |
},
|
| 244 |
{
|
| 245 |
+
"epoch": 0.76,
|
| 246 |
+
"learning_rate": 4.713201180936314e-05,
|
| 247 |
+
"loss": 0.5202,
|
| 248 |
"step": 400
|
| 249 |
},
|
| 250 |
{
|
| 251 |
+
"epoch": 0.78,
|
| 252 |
+
"learning_rate": 4.692113032475749e-05,
|
| 253 |
+
"loss": 0.5143,
|
| 254 |
"step": 410
|
| 255 |
},
|
| 256 |
{
|
| 257 |
+
"epoch": 0.8,
|
| 258 |
+
"learning_rate": 4.671024884015184e-05,
|
| 259 |
+
"loss": 0.4946,
|
| 260 |
"step": 420
|
| 261 |
},
|
| 262 |
{
|
| 263 |
+
"epoch": 0.81,
|
| 264 |
+
"learning_rate": 4.6499367355546186e-05,
|
| 265 |
+
"loss": 0.5009,
|
| 266 |
"step": 430
|
| 267 |
},
|
| 268 |
{
|
| 269 |
+
"epoch": 0.83,
|
| 270 |
+
"learning_rate": 4.6288485870940535e-05,
|
| 271 |
+
"loss": 0.5015,
|
| 272 |
"step": 440
|
| 273 |
},
|
| 274 |
{
|
| 275 |
+
"epoch": 0.85,
|
| 276 |
+
"learning_rate": 4.607760438633488e-05,
|
| 277 |
+
"loss": 0.4837,
|
| 278 |
"step": 450
|
| 279 |
},
|
| 280 |
{
|
| 281 |
+
"epoch": 0.87,
|
| 282 |
+
"learning_rate": 4.586672290172923e-05,
|
| 283 |
+
"loss": 0.4814,
|
| 284 |
"step": 460
|
| 285 |
},
|
| 286 |
{
|
| 287 |
+
"epoch": 0.89,
|
| 288 |
+
"learning_rate": 4.565584141712358e-05,
|
| 289 |
+
"loss": 0.4827,
|
| 290 |
"step": 470
|
| 291 |
},
|
| 292 |
{
|
| 293 |
+
"epoch": 0.91,
|
| 294 |
+
"learning_rate": 4.544495993251793e-05,
|
| 295 |
+
"loss": 0.5304,
|
| 296 |
"step": 480
|
| 297 |
},
|
| 298 |
{
|
| 299 |
+
"epoch": 0.93,
|
| 300 |
+
"learning_rate": 4.523407844791228e-05,
|
| 301 |
+
"loss": 0.4592,
|
| 302 |
"step": 490
|
| 303 |
},
|
| 304 |
{
|
| 305 |
+
"epoch": 0.95,
|
| 306 |
+
"learning_rate": 4.5023196963306626e-05,
|
| 307 |
+
"loss": 0.4759,
|
| 308 |
"step": 500
|
| 309 |
},
|
| 310 |
{
|
| 311 |
+
"epoch": 0.97,
|
| 312 |
+
"learning_rate": 4.4812315478700974e-05,
|
| 313 |
+
"loss": 0.4431,
|
| 314 |
"step": 510
|
| 315 |
},
|
| 316 |
{
|
| 317 |
+
"epoch": 0.99,
|
| 318 |
+
"learning_rate": 4.460143399409532e-05,
|
| 319 |
+
"loss": 0.4378,
|
| 320 |
"step": 520
|
| 321 |
},
|
| 322 |
{
|
| 323 |
+
"epoch": 1.0,
|
| 324 |
+
"eval_accuracy": 0.8507462686567164,
|
| 325 |
+
"eval_loss": 0.36681434512138367,
|
| 326 |
+
"eval_runtime": 136.5295,
|
| 327 |
+
"eval_samples_per_second": 70.666,
|
| 328 |
+
"eval_steps_per_second": 2.212,
|
| 329 |
+
"step": 527
|
| 330 |
+
},
|
| 331 |
+
{
|
| 332 |
+
"epoch": 1.01,
|
| 333 |
+
"learning_rate": 4.439055250948967e-05,
|
| 334 |
+
"loss": 0.486,
|
| 335 |
"step": 530
|
| 336 |
},
|
| 337 |
{
|
| 338 |
+
"epoch": 1.02,
|
| 339 |
+
"learning_rate": 4.417967102488402e-05,
|
| 340 |
+
"loss": 0.4533,
|
| 341 |
"step": 540
|
| 342 |
},
|
| 343 |
{
|
| 344 |
+
"epoch": 1.04,
|
| 345 |
+
"learning_rate": 4.396878954027837e-05,
|
| 346 |
+
"loss": 0.4119,
|
| 347 |
"step": 550
|
| 348 |
},
|
| 349 |
{
|
| 350 |
+
"epoch": 1.06,
|
| 351 |
+
"learning_rate": 4.3757908055672716e-05,
|
| 352 |
+
"loss": 0.4279,
|
| 353 |
"step": 560
|
| 354 |
},
|
| 355 |
{
|
| 356 |
+
"epoch": 1.08,
|
| 357 |
+
"learning_rate": 4.3547026571067065e-05,
|
| 358 |
+
"loss": 0.4389,
|
| 359 |
"step": 570
|
| 360 |
},
|
| 361 |
{
|
| 362 |
+
"epoch": 1.1,
|
| 363 |
+
"learning_rate": 4.333614508646141e-05,
|
| 364 |
+
"loss": 0.4344,
|
| 365 |
"step": 580
|
| 366 |
},
|
| 367 |
{
|
| 368 |
+
"epoch": 1.12,
|
| 369 |
+
"learning_rate": 4.312526360185576e-05,
|
| 370 |
+
"loss": 0.3977,
|
| 371 |
"step": 590
|
| 372 |
},
|
| 373 |
{
|
| 374 |
+
"epoch": 1.14,
|
| 375 |
+
"learning_rate": 4.291438211725011e-05,
|
| 376 |
+
"loss": 0.413,
|
| 377 |
"step": 600
|
| 378 |
},
|
| 379 |
{
|
| 380 |
+
"epoch": 1.16,
|
| 381 |
+
"learning_rate": 4.270350063264446e-05,
|
| 382 |
+
"loss": 0.4302,
|
| 383 |
"step": 610
|
| 384 |
},
|
| 385 |
{
|
| 386 |
+
"epoch": 1.18,
|
| 387 |
+
"learning_rate": 4.249261914803881e-05,
|
| 388 |
+
"loss": 0.4229,
|
| 389 |
"step": 620
|
| 390 |
},
|
| 391 |
{
|
| 392 |
+
"epoch": 1.2,
|
| 393 |
+
"learning_rate": 4.2281737663433155e-05,
|
| 394 |
+
"loss": 0.4015,
|
| 395 |
"step": 630
|
| 396 |
},
|
| 397 |
{
|
| 398 |
+
"epoch": 1.21,
|
| 399 |
+
"learning_rate": 4.20708561788275e-05,
|
| 400 |
+
"loss": 0.3929,
|
| 401 |
"step": 640
|
| 402 |
},
|
| 403 |
{
|
| 404 |
+
"epoch": 1.23,
|
| 405 |
+
"learning_rate": 4.1859974694221845e-05,
|
| 406 |
+
"loss": 0.4133,
|
| 407 |
"step": 650
|
| 408 |
},
|
| 409 |
{
|
| 410 |
+
"epoch": 1.25,
|
| 411 |
+
"learning_rate": 4.1649093209616194e-05,
|
| 412 |
+
"loss": 0.384,
|
| 413 |
"step": 660
|
| 414 |
},
|
| 415 |
{
|
| 416 |
+
"epoch": 1.27,
|
| 417 |
+
"learning_rate": 4.143821172501054e-05,
|
| 418 |
+
"loss": 0.4043,
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
| 419 |
"step": 670
|
| 420 |
},
|
| 421 |
{
|
| 422 |
+
"epoch": 1.29,
|
| 423 |
+
"learning_rate": 4.122733024040489e-05,
|
| 424 |
+
"loss": 0.3696,
|
| 425 |
"step": 680
|
| 426 |
},
|
| 427 |
{
|
| 428 |
+
"epoch": 1.31,
|
| 429 |
+
"learning_rate": 4.101644875579924e-05,
|
| 430 |
+
"loss": 0.3942,
|
| 431 |
"step": 690
|
| 432 |
},
|
| 433 |
{
|
| 434 |
+
"epoch": 1.33,
|
| 435 |
+
"learning_rate": 4.080556727119359e-05,
|
| 436 |
+
"loss": 0.3915,
|
| 437 |
"step": 700
|
| 438 |
},
|
| 439 |
{
|
| 440 |
+
"epoch": 1.35,
|
| 441 |
+
"learning_rate": 4.0594685786587936e-05,
|
| 442 |
+
"loss": 0.3431,
|
| 443 |
"step": 710
|
| 444 |
},
|
| 445 |
{
|
| 446 |
+
"epoch": 1.37,
|
| 447 |
+
"learning_rate": 4.0383804301982284e-05,
|
| 448 |
+
"loss": 0.352,
|
| 449 |
"step": 720
|
| 450 |
},
|
| 451 |
{
|
| 452 |
+
"epoch": 1.38,
|
| 453 |
+
"learning_rate": 4.017292281737663e-05,
|
| 454 |
+
"loss": 0.3876,
|
| 455 |
"step": 730
|
| 456 |
},
|
| 457 |
{
|
| 458 |
+
"epoch": 1.4,
|
| 459 |
+
"learning_rate": 3.996204133277098e-05,
|
| 460 |
+
"loss": 0.3589,
|
| 461 |
"step": 740
|
| 462 |
},
|
| 463 |
{
|
| 464 |
+
"epoch": 1.42,
|
| 465 |
+
"learning_rate": 3.975115984816533e-05,
|
| 466 |
+
"loss": 0.3519,
|
| 467 |
"step": 750
|
| 468 |
},
|
| 469 |
{
|
| 470 |
+
"epoch": 1.44,
|
| 471 |
+
"learning_rate": 3.954027836355968e-05,
|
| 472 |
+
"loss": 0.3908,
|
| 473 |
"step": 760
|
| 474 |
},
|
| 475 |
{
|
| 476 |
+
"epoch": 1.46,
|
| 477 |
+
"learning_rate": 3.9329396878954027e-05,
|
| 478 |
+
"loss": 0.3677,
|
| 479 |
"step": 770
|
| 480 |
},
|
| 481 |
{
|
| 482 |
+
"epoch": 1.48,
|
| 483 |
+
"learning_rate": 3.9118515394348375e-05,
|
| 484 |
+
"loss": 0.3549,
|
| 485 |
"step": 780
|
| 486 |
},
|
| 487 |
{
|
| 488 |
+
"epoch": 1.5,
|
| 489 |
+
"learning_rate": 3.8907633909742723e-05,
|
| 490 |
+
"loss": 0.3453,
|
| 491 |
"step": 790
|
| 492 |
},
|
| 493 |
{
|
| 494 |
+
"epoch": 1.52,
|
| 495 |
+
"learning_rate": 3.869675242513707e-05,
|
| 496 |
+
"loss": 0.3366,
|
| 497 |
"step": 800
|
| 498 |
},
|
| 499 |
{
|
| 500 |
+
"epoch": 1.54,
|
| 501 |
+
"learning_rate": 3.848587094053142e-05,
|
| 502 |
+
"loss": 0.3553,
|
| 503 |
"step": 810
|
| 504 |
},
|
| 505 |
{
|
| 506 |
+
"epoch": 1.56,
|
| 507 |
+
"learning_rate": 3.827498945592577e-05,
|
| 508 |
+
"loss": 0.3224,
|
| 509 |
"step": 820
|
| 510 |
},
|
| 511 |
{
|
| 512 |
+
"epoch": 1.57,
|
| 513 |
+
"learning_rate": 3.806410797132012e-05,
|
| 514 |
+
"loss": 0.345,
|
| 515 |
"step": 830
|
| 516 |
},
|
| 517 |
{
|
| 518 |
+
"epoch": 1.59,
|
| 519 |
+
"learning_rate": 3.7853226486714466e-05,
|
| 520 |
+
"loss": 0.3298,
|
| 521 |
"step": 840
|
| 522 |
},
|
| 523 |
{
|
| 524 |
+
"epoch": 1.61,
|
| 525 |
+
"learning_rate": 3.7642345002108814e-05,
|
| 526 |
+
"loss": 0.3417,
|
| 527 |
"step": 850
|
| 528 |
},
|
| 529 |
{
|
| 530 |
+
"epoch": 1.63,
|
| 531 |
+
"learning_rate": 3.743146351750316e-05,
|
| 532 |
+
"loss": 0.3832,
|
| 533 |
"step": 860
|
| 534 |
},
|
| 535 |
{
|
| 536 |
+
"epoch": 1.65,
|
| 537 |
+
"learning_rate": 3.722058203289751e-05,
|
| 538 |
+
"loss": 0.3343,
|
| 539 |
"step": 870
|
| 540 |
},
|
| 541 |
{
|
| 542 |
+
"epoch": 1.67,
|
| 543 |
+
"learning_rate": 3.700970054829186e-05,
|
| 544 |
+
"loss": 0.3112,
|
| 545 |
"step": 880
|
| 546 |
},
|
| 547 |
{
|
| 548 |
+
"epoch": 1.69,
|
| 549 |
+
"learning_rate": 3.679881906368621e-05,
|
| 550 |
+
"loss": 0.3355,
|
| 551 |
"step": 890
|
| 552 |
},
|
| 553 |
{
|
| 554 |
+
"epoch": 1.71,
|
| 555 |
+
"learning_rate": 3.658793757908056e-05,
|
| 556 |
+
"loss": 0.3223,
|
| 557 |
"step": 900
|
| 558 |
},
|
| 559 |
{
|
| 560 |
+
"epoch": 1.73,
|
| 561 |
+
"learning_rate": 3.637705609447491e-05,
|
| 562 |
+
"loss": 0.3229,
|
| 563 |
"step": 910
|
| 564 |
},
|
| 565 |
{
|
| 566 |
+
"epoch": 1.74,
|
| 567 |
+
"learning_rate": 3.616617460986926e-05,
|
| 568 |
+
"loss": 0.3395,
|
| 569 |
"step": 920
|
| 570 |
},
|
| 571 |
{
|
| 572 |
+
"epoch": 1.76,
|
| 573 |
+
"learning_rate": 3.595529312526361e-05,
|
| 574 |
+
"loss": 0.3092,
|
| 575 |
"step": 930
|
| 576 |
},
|
| 577 |
{
|
| 578 |
+
"epoch": 1.78,
|
| 579 |
+
"learning_rate": 3.574441164065796e-05,
|
| 580 |
+
"loss": 0.3006,
|
| 581 |
"step": 940
|
| 582 |
},
|
| 583 |
{
|
| 584 |
+
"epoch": 1.8,
|
| 585 |
+
"learning_rate": 3.5533530156052305e-05,
|
| 586 |
+
"loss": 0.3557,
|
| 587 |
"step": 950
|
| 588 |
},
|
| 589 |
{
|
| 590 |
+
"epoch": 1.82,
|
| 591 |
+
"learning_rate": 3.5322648671446654e-05,
|
| 592 |
+
"loss": 0.325,
|
| 593 |
"step": 960
|
| 594 |
},
|
| 595 |
{
|
| 596 |
+
"epoch": 1.84,
|
| 597 |
+
"learning_rate": 3.5111767186841e-05,
|
| 598 |
+
"loss": 0.3039,
|
| 599 |
"step": 970
|
| 600 |
},
|
| 601 |
{
|
| 602 |
+
"epoch": 1.86,
|
| 603 |
+
"learning_rate": 3.490088570223535e-05,
|
| 604 |
+
"loss": 0.3506,
|
| 605 |
"step": 980
|
| 606 |
},
|
| 607 |
{
|
| 608 |
+
"epoch": 1.88,
|
| 609 |
+
"learning_rate": 3.46900042176297e-05,
|
| 610 |
+
"loss": 0.3268,
|
| 611 |
"step": 990
|
| 612 |
},
|
| 613 |
{
|
| 614 |
+
"epoch": 1.9,
|
| 615 |
+
"learning_rate": 3.447912273302405e-05,
|
| 616 |
+
"loss": 0.3007,
|
| 617 |
"step": 1000
|
| 618 |
},
|
| 619 |
{
|
| 620 |
+
"epoch": 1.92,
|
| 621 |
+
"learning_rate": 3.426824124841839e-05,
|
| 622 |
+
"loss": 0.3202,
|
| 623 |
"step": 1010
|
| 624 |
},
|
| 625 |
{
|
| 626 |
+
"epoch": 1.93,
|
| 627 |
+
"learning_rate": 3.405735976381274e-05,
|
| 628 |
+
"loss": 0.3118,
|
| 629 |
"step": 1020
|
| 630 |
},
|
| 631 |
{
|
| 632 |
+
"epoch": 1.95,
|
| 633 |
+
"learning_rate": 3.3846478279207086e-05,
|
| 634 |
+
"loss": 0.3278,
|
| 635 |
"step": 1030
|
| 636 |
},
|
| 637 |
{
|
| 638 |
+
"epoch": 1.97,
|
| 639 |
+
"learning_rate": 3.3635596794601434e-05,
|
| 640 |
+
"loss": 0.292,
|
| 641 |
"step": 1040
|
| 642 |
},
|
| 643 |
{
|
| 644 |
+
"epoch": 1.99,
|
| 645 |
+
"learning_rate": 3.342471530999578e-05,
|
| 646 |
+
"loss": 0.3133,
|
| 647 |
"step": 1050
|
| 648 |
},
|
| 649 |
{
|
| 650 |
+
"epoch": 2.0,
|
| 651 |
+
"eval_accuracy": 0.9429933665008292,
|
| 652 |
+
"eval_loss": 0.15862515568733215,
|
| 653 |
+
"eval_runtime": 141.8451,
|
| 654 |
+
"eval_samples_per_second": 68.018,
|
| 655 |
+
"eval_steps_per_second": 2.129,
|
| 656 |
+
"step": 1054
|
| 657 |
+
},
|
| 658 |
+
{
|
| 659 |
+
"epoch": 2.01,
|
| 660 |
+
"learning_rate": 3.321383382539013e-05,
|
| 661 |
+
"loss": 0.3435,
|
| 662 |
"step": 1060
|
| 663 |
},
|
| 664 |
{
|
| 665 |
+
"epoch": 2.03,
|
| 666 |
+
"learning_rate": 3.300295234078448e-05,
|
| 667 |
+
"loss": 0.31,
|
| 668 |
"step": 1070
|
| 669 |
},
|
| 670 |
{
|
| 671 |
+
"epoch": 2.05,
|
| 672 |
+
"learning_rate": 3.279207085617883e-05,
|
| 673 |
+
"loss": 0.3035,
|
| 674 |
"step": 1080
|
| 675 |
},
|
| 676 |
{
|
| 677 |
+
"epoch": 2.07,
|
| 678 |
+
"learning_rate": 3.2581189371573177e-05,
|
| 679 |
+
"loss": 0.3138,
|
| 680 |
"step": 1090
|
| 681 |
},
|
| 682 |
{
|
| 683 |
+
"epoch": 2.09,
|
| 684 |
+
"learning_rate": 3.2370307886967525e-05,
|
| 685 |
+
"loss": 0.309,
|
| 686 |
"step": 1100
|
| 687 |
},
|
| 688 |
{
|
| 689 |
+
"epoch": 2.11,
|
| 690 |
+
"learning_rate": 3.2159426402361873e-05,
|
| 691 |
+
"loss": 0.3034,
|
| 692 |
"step": 1110
|
| 693 |
},
|
| 694 |
{
|
| 695 |
+
"epoch": 2.13,
|
| 696 |
+
"learning_rate": 3.194854491775622e-05,
|
| 697 |
+
"loss": 0.2805,
|
| 698 |
"step": 1120
|
| 699 |
},
|
| 700 |
{
|
| 701 |
+
"epoch": 2.14,
|
| 702 |
+
"learning_rate": 3.173766343315057e-05,
|
| 703 |
+
"loss": 0.2889,
|
| 704 |
"step": 1130
|
| 705 |
},
|
| 706 |
{
|
| 707 |
+
"epoch": 2.16,
|
| 708 |
+
"learning_rate": 3.152678194854492e-05,
|
| 709 |
+
"loss": 0.2864,
|
| 710 |
"step": 1140
|
| 711 |
},
|
| 712 |
{
|
| 713 |
+
"epoch": 2.18,
|
| 714 |
+
"learning_rate": 3.131590046393927e-05,
|
| 715 |
+
"loss": 0.2723,
|
| 716 |
"step": 1150
|
| 717 |
},
|
| 718 |
{
|
| 719 |
+
"epoch": 2.2,
|
| 720 |
+
"learning_rate": 3.1105018979333616e-05,
|
| 721 |
+
"loss": 0.2763,
|
| 722 |
"step": 1160
|
| 723 |
},
|
| 724 |
{
|
| 725 |
+
"epoch": 2.22,
|
| 726 |
+
"learning_rate": 3.0894137494727964e-05,
|
| 727 |
+
"loss": 0.2561,
|
| 728 |
"step": 1170
|
| 729 |
},
|
| 730 |
{
|
| 731 |
+
"epoch": 2.24,
|
| 732 |
+
"learning_rate": 3.068325601012231e-05,
|
| 733 |
+
"loss": 0.2732,
|
| 734 |
"step": 1180
|
| 735 |
},
|
| 736 |
{
|
| 737 |
+
"epoch": 2.26,
|
| 738 |
+
"learning_rate": 3.047237452551666e-05,
|
| 739 |
+
"loss": 0.2458,
|
| 740 |
"step": 1190
|
| 741 |
},
|
| 742 |
{
|
| 743 |
+
"epoch": 2.28,
|
| 744 |
+
"learning_rate": 3.026149304091101e-05,
|
| 745 |
+
"loss": 0.2632,
|
| 746 |
"step": 1200
|
| 747 |
},
|
| 748 |
{
|
| 749 |
+
"epoch": 2.3,
|
| 750 |
+
"learning_rate": 3.0050611556305358e-05,
|
| 751 |
+
"loss": 0.2889,
|
| 752 |
"step": 1210
|
| 753 |
},
|
| 754 |
{
|
| 755 |
+
"epoch": 2.31,
|
| 756 |
+
"learning_rate": 2.9839730071699706e-05,
|
| 757 |
+
"loss": 0.2406,
|
| 758 |
"step": 1220
|
| 759 |
},
|
| 760 |
{
|
| 761 |
+
"epoch": 2.33,
|
| 762 |
+
"learning_rate": 2.9628848587094055e-05,
|
| 763 |
+
"loss": 0.2566,
|
| 764 |
"step": 1230
|
| 765 |
},
|
| 766 |
{
|
| 767 |
+
"epoch": 2.35,
|
| 768 |
+
"learning_rate": 2.9417967102488403e-05,
|
| 769 |
+
"loss": 0.2576,
|
| 770 |
"step": 1240
|
| 771 |
},
|
| 772 |
{
|
| 773 |
+
"epoch": 2.37,
|
| 774 |
+
"learning_rate": 2.920708561788275e-05,
|
| 775 |
+
"loss": 0.268,
|
| 776 |
"step": 1250
|
| 777 |
},
|
| 778 |
{
|
| 779 |
+
"epoch": 2.39,
|
| 780 |
+
"learning_rate": 2.89962041332771e-05,
|
| 781 |
+
"loss": 0.2556,
|
| 782 |
"step": 1260
|
| 783 |
},
|
| 784 |
{
|
| 785 |
+
"epoch": 2.41,
|
| 786 |
+
"learning_rate": 2.878532264867145e-05,
|
| 787 |
+
"loss": 0.2577,
|
| 788 |
"step": 1270
|
| 789 |
},
|
| 790 |
{
|
| 791 |
+
"epoch": 2.43,
|
| 792 |
+
"learning_rate": 2.8574441164065797e-05,
|
| 793 |
+
"loss": 0.2507,
|
| 794 |
"step": 1280
|
| 795 |
},
|
| 796 |
{
|
| 797 |
+
"epoch": 2.45,
|
| 798 |
+
"learning_rate": 2.8363559679460145e-05,
|
| 799 |
+
"loss": 0.2271,
|
| 800 |
"step": 1290
|
| 801 |
},
|
| 802 |
{
|
| 803 |
+
"epoch": 2.47,
|
| 804 |
+
"learning_rate": 2.8152678194854494e-05,
|
| 805 |
+
"loss": 0.2792,
|
| 806 |
"step": 1300
|
| 807 |
},
|
| 808 |
{
|
| 809 |
+
"epoch": 2.49,
|
| 810 |
+
"learning_rate": 2.7941796710248842e-05,
|
| 811 |
+
"loss": 0.2755,
|
| 812 |
"step": 1310
|
| 813 |
},
|
| 814 |
{
|
| 815 |
+
"epoch": 2.5,
|
| 816 |
+
"learning_rate": 2.773091522564319e-05,
|
| 817 |
+
"loss": 0.2456,
|
| 818 |
"step": 1320
|
| 819 |
},
|
| 820 |
{
|
| 821 |
+
"epoch": 2.52,
|
| 822 |
+
"learning_rate": 2.752003374103754e-05,
|
| 823 |
+
"loss": 0.2463,
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
| 824 |
"step": 1330
|
| 825 |
},
|
| 826 |
{
|
| 827 |
+
"epoch": 2.54,
|
| 828 |
+
"learning_rate": 2.7309152256431884e-05,
|
| 829 |
+
"loss": 0.2569,
|
| 830 |
"step": 1340
|
| 831 |
},
|
| 832 |
{
|
| 833 |
+
"epoch": 2.56,
|
| 834 |
+
"learning_rate": 2.7098270771826233e-05,
|
| 835 |
+
"loss": 0.2609,
|
| 836 |
"step": 1350
|
| 837 |
},
|
| 838 |
{
|
| 839 |
+
"epoch": 2.58,
|
| 840 |
+
"learning_rate": 2.688738928722058e-05,
|
| 841 |
+
"loss": 0.2437,
|
| 842 |
"step": 1360
|
| 843 |
},
|
| 844 |
{
|
| 845 |
+
"epoch": 2.6,
|
| 846 |
+
"learning_rate": 2.667650780261493e-05,
|
| 847 |
+
"loss": 0.2491,
|
| 848 |
"step": 1370
|
| 849 |
},
|
| 850 |
{
|
| 851 |
+
"epoch": 2.62,
|
| 852 |
+
"learning_rate": 2.6465626318009278e-05,
|
| 853 |
+
"loss": 0.1923,
|
| 854 |
"step": 1380
|
| 855 |
},
|
| 856 |
{
|
| 857 |
+
"epoch": 2.64,
|
| 858 |
+
"learning_rate": 2.6254744833403626e-05,
|
| 859 |
+
"loss": 0.2435,
|
| 860 |
"step": 1390
|
| 861 |
},
|
| 862 |
{
|
| 863 |
+
"epoch": 2.66,
|
| 864 |
+
"learning_rate": 2.6043863348797975e-05,
|
| 865 |
+
"loss": 0.2382,
|
| 866 |
"step": 1400
|
| 867 |
},
|
| 868 |
{
|
| 869 |
+
"epoch": 2.67,
|
| 870 |
+
"learning_rate": 2.5832981864192323e-05,
|
| 871 |
+
"loss": 0.2316,
|
| 872 |
"step": 1410
|
| 873 |
},
|
| 874 |
{
|
| 875 |
+
"epoch": 2.69,
|
| 876 |
+
"learning_rate": 2.562210037958667e-05,
|
| 877 |
+
"loss": 0.2294,
|
| 878 |
"step": 1420
|
| 879 |
},
|
| 880 |
{
|
| 881 |
+
"epoch": 2.71,
|
| 882 |
+
"learning_rate": 2.541121889498102e-05,
|
| 883 |
+
"loss": 0.2187,
|
| 884 |
"step": 1430
|
| 885 |
},
|
| 886 |
{
|
| 887 |
+
"epoch": 2.73,
|
| 888 |
+
"learning_rate": 2.520033741037537e-05,
|
| 889 |
+
"loss": 0.2595,
|
| 890 |
"step": 1440
|
| 891 |
},
|
| 892 |
{
|
| 893 |
+
"epoch": 2.75,
|
| 894 |
+
"learning_rate": 2.498945592576972e-05,
|
| 895 |
+
"loss": 0.2401,
|
| 896 |
"step": 1450
|
| 897 |
},
|
| 898 |
{
|
| 899 |
+
"epoch": 2.77,
|
| 900 |
+
"learning_rate": 2.477857444116407e-05,
|
| 901 |
+
"loss": 0.2344,
|
| 902 |
"step": 1460
|
| 903 |
},
|
| 904 |
{
|
| 905 |
+
"epoch": 2.79,
|
| 906 |
+
"learning_rate": 2.4567692956558417e-05,
|
| 907 |
+
"loss": 0.2334,
|
| 908 |
"step": 1470
|
| 909 |
},
|
| 910 |
{
|
| 911 |
+
"epoch": 2.81,
|
| 912 |
+
"learning_rate": 2.4356811471952766e-05,
|
| 913 |
+
"loss": 0.2697,
|
| 914 |
"step": 1480
|
| 915 |
},
|
| 916 |
{
|
| 917 |
+
"epoch": 2.83,
|
| 918 |
+
"learning_rate": 2.4145929987347114e-05,
|
| 919 |
+
"loss": 0.2407,
|
| 920 |
"step": 1490
|
| 921 |
},
|
| 922 |
{
|
| 923 |
+
"epoch": 2.85,
|
| 924 |
+
"learning_rate": 2.3935048502741463e-05,
|
| 925 |
+
"loss": 0.2434,
|
| 926 |
"step": 1500
|
| 927 |
},
|
| 928 |
{
|
| 929 |
+
"epoch": 2.86,
|
| 930 |
+
"learning_rate": 2.372416701813581e-05,
|
| 931 |
+
"loss": 0.2781,
|
| 932 |
"step": 1510
|
| 933 |
},
|
| 934 |
{
|
| 935 |
+
"epoch": 2.88,
|
| 936 |
+
"learning_rate": 2.351328553353016e-05,
|
| 937 |
+
"loss": 0.2377,
|
| 938 |
"step": 1520
|
| 939 |
},
|
| 940 |
{
|
| 941 |
+
"epoch": 2.9,
|
| 942 |
+
"learning_rate": 2.3302404048924504e-05,
|
| 943 |
+
"loss": 0.2478,
|
| 944 |
"step": 1530
|
| 945 |
},
|
| 946 |
{
|
| 947 |
+
"epoch": 2.92,
|
| 948 |
+
"learning_rate": 2.3091522564318853e-05,
|
| 949 |
+
"loss": 0.2558,
|
| 950 |
"step": 1540
|
| 951 |
},
|
| 952 |
{
|
| 953 |
+
"epoch": 2.94,
|
| 954 |
+
"learning_rate": 2.28806410797132e-05,
|
| 955 |
+
"loss": 0.2378,
|
| 956 |
"step": 1550
|
| 957 |
},
|
| 958 |
{
|
| 959 |
+
"epoch": 2.96,
|
| 960 |
+
"learning_rate": 2.266975959510755e-05,
|
| 961 |
+
"loss": 0.2303,
|
| 962 |
"step": 1560
|
| 963 |
},
|
| 964 |
{
|
| 965 |
+
"epoch": 2.98,
|
| 966 |
+
"learning_rate": 2.2458878110501898e-05,
|
| 967 |
+
"loss": 0.271,
|
| 968 |
"step": 1570
|
| 969 |
},
|
| 970 |
{
|
| 971 |
+
"epoch": 3.0,
|
| 972 |
+
"learning_rate": 2.2247996625896247e-05,
|
| 973 |
+
"loss": 0.2065,
|
| 974 |
"step": 1580
|
| 975 |
},
|
| 976 |
{
|
| 977 |
+
"epoch": 3.0,
|
| 978 |
+
"eval_accuracy": 0.9607172470978441,
|
| 979 |
+
"eval_loss": 0.1048618033528328,
|
| 980 |
+
"eval_runtime": 131.6181,
|
| 981 |
+
"eval_samples_per_second": 73.303,
|
| 982 |
+
"eval_steps_per_second": 2.295,
|
| 983 |
+
"step": 1581
|
| 984 |
+
},
|
| 985 |
+
{
|
| 986 |
+
"epoch": 3.02,
|
| 987 |
+
"learning_rate": 2.2037115141290595e-05,
|
| 988 |
+
"loss": 0.211,
|
| 989 |
"step": 1590
|
| 990 |
},
|
| 991 |
{
|
| 992 |
+
"epoch": 3.04,
|
| 993 |
+
"learning_rate": 2.1826233656684943e-05,
|
| 994 |
+
"loss": 0.2139,
|
| 995 |
"step": 1600
|
| 996 |
},
|
| 997 |
{
|
| 998 |
+
"epoch": 3.05,
|
| 999 |
+
"learning_rate": 2.1615352172079292e-05,
|
| 1000 |
+
"loss": 0.226,
|
| 1001 |
"step": 1610
|
| 1002 |
},
|
| 1003 |
{
|
| 1004 |
+
"epoch": 3.07,
|
| 1005 |
+
"learning_rate": 2.140447068747364e-05,
|
| 1006 |
+
"loss": 0.2295,
|
| 1007 |
"step": 1620
|
| 1008 |
},
|
| 1009 |
{
|
| 1010 |
+
"epoch": 3.09,
|
| 1011 |
+
"learning_rate": 2.119358920286799e-05,
|
| 1012 |
+
"loss": 0.2284,
|
| 1013 |
"step": 1630
|
| 1014 |
},
|
| 1015 |
{
|
| 1016 |
+
"epoch": 3.11,
|
| 1017 |
+
"learning_rate": 2.0982707718262337e-05,
|
| 1018 |
+
"loss": 0.2181,
|
| 1019 |
"step": 1640
|
| 1020 |
},
|
| 1021 |
{
|
| 1022 |
+
"epoch": 3.13,
|
| 1023 |
+
"learning_rate": 2.0771826233656686e-05,
|
| 1024 |
+
"loss": 0.2174,
|
| 1025 |
"step": 1650
|
| 1026 |
},
|
| 1027 |
{
|
| 1028 |
+
"epoch": 3.15,
|
| 1029 |
+
"learning_rate": 2.0560944749051034e-05,
|
| 1030 |
+
"loss": 0.2301,
|
| 1031 |
"step": 1660
|
| 1032 |
},
|
| 1033 |
{
|
| 1034 |
+
"epoch": 3.17,
|
| 1035 |
+
"learning_rate": 2.0350063264445383e-05,
|
| 1036 |
+
"loss": 0.2078,
|
| 1037 |
"step": 1670
|
| 1038 |
},
|
| 1039 |
{
|
| 1040 |
+
"epoch": 3.19,
|
| 1041 |
+
"learning_rate": 2.013918177983973e-05,
|
| 1042 |
+
"loss": 0.2338,
|
| 1043 |
"step": 1680
|
| 1044 |
},
|
| 1045 |
{
|
| 1046 |
+
"epoch": 3.21,
|
| 1047 |
+
"learning_rate": 1.992830029523408e-05,
|
| 1048 |
+
"loss": 0.235,
|
| 1049 |
"step": 1690
|
| 1050 |
},
|
| 1051 |
{
|
| 1052 |
+
"epoch": 3.23,
|
| 1053 |
+
"learning_rate": 1.9717418810628428e-05,
|
| 1054 |
+
"loss": 0.2152,
|
| 1055 |
"step": 1700
|
| 1056 |
},
|
| 1057 |
{
|
| 1058 |
+
"epoch": 3.24,
|
| 1059 |
+
"learning_rate": 1.9506537326022776e-05,
|
| 1060 |
+
"loss": 0.2154,
|
| 1061 |
"step": 1710
|
| 1062 |
},
|
| 1063 |
{
|
| 1064 |
+
"epoch": 3.26,
|
| 1065 |
+
"learning_rate": 1.9295655841417125e-05,
|
| 1066 |
+
"loss": 0.2146,
|
| 1067 |
"step": 1720
|
| 1068 |
},
|
| 1069 |
{
|
| 1070 |
+
"epoch": 3.28,
|
| 1071 |
+
"learning_rate": 1.9084774356811473e-05,
|
| 1072 |
+
"loss": 0.2279,
|
| 1073 |
"step": 1730
|
| 1074 |
},
|
| 1075 |
{
|
| 1076 |
+
"epoch": 3.3,
|
| 1077 |
+
"learning_rate": 1.887389287220582e-05,
|
| 1078 |
+
"loss": 0.2176,
|
| 1079 |
"step": 1740
|
| 1080 |
},
|
| 1081 |
{
|
| 1082 |
+
"epoch": 3.32,
|
| 1083 |
+
"learning_rate": 1.8663011387600167e-05,
|
| 1084 |
"loss": 0.2409,
|
| 1085 |
"step": 1750
|
| 1086 |
},
|
| 1087 |
{
|
| 1088 |
+
"epoch": 3.34,
|
| 1089 |
+
"learning_rate": 1.8452129902994515e-05,
|
| 1090 |
+
"loss": 0.2294,
|
| 1091 |
"step": 1760
|
| 1092 |
},
|
| 1093 |
{
|
| 1094 |
+
"epoch": 3.36,
|
| 1095 |
+
"learning_rate": 1.8241248418388867e-05,
|
| 1096 |
+
"loss": 0.2118,
|
| 1097 |
"step": 1770
|
| 1098 |
},
|
| 1099 |
{
|
| 1100 |
+
"epoch": 3.38,
|
| 1101 |
+
"learning_rate": 1.8030366933783215e-05,
|
| 1102 |
+
"loss": 0.2141,
|
| 1103 |
"step": 1780
|
| 1104 |
},
|
| 1105 |
{
|
| 1106 |
+
"epoch": 3.4,
|
| 1107 |
+
"learning_rate": 1.7819485449177564e-05,
|
| 1108 |
+
"loss": 0.1921,
|
| 1109 |
"step": 1790
|
| 1110 |
},
|
| 1111 |
{
|
| 1112 |
+
"epoch": 3.41,
|
| 1113 |
+
"learning_rate": 1.7608603964571912e-05,
|
| 1114 |
+
"loss": 0.184,
|
| 1115 |
"step": 1800
|
| 1116 |
},
|
| 1117 |
{
|
| 1118 |
+
"epoch": 3.43,
|
| 1119 |
+
"learning_rate": 1.739772247996626e-05,
|
| 1120 |
+
"loss": 0.2109,
|
| 1121 |
"step": 1810
|
| 1122 |
},
|
| 1123 |
{
|
| 1124 |
+
"epoch": 3.45,
|
| 1125 |
+
"learning_rate": 1.718684099536061e-05,
|
| 1126 |
+
"loss": 0.2119,
|
| 1127 |
"step": 1820
|
| 1128 |
},
|
| 1129 |
{
|
| 1130 |
+
"epoch": 3.47,
|
| 1131 |
+
"learning_rate": 1.6975959510754958e-05,
|
| 1132 |
+
"loss": 0.2294,
|
| 1133 |
"step": 1830
|
| 1134 |
},
|
| 1135 |
{
|
| 1136 |
+
"epoch": 3.49,
|
| 1137 |
+
"learning_rate": 1.6765078026149306e-05,
|
| 1138 |
+
"loss": 0.234,
|
| 1139 |
"step": 1840
|
| 1140 |
},
|
| 1141 |
{
|
| 1142 |
+
"epoch": 3.51,
|
| 1143 |
+
"learning_rate": 1.6554196541543654e-05,
|
| 1144 |
+
"loss": 0.1865,
|
| 1145 |
"step": 1850
|
| 1146 |
},
|
| 1147 |
{
|
| 1148 |
+
"epoch": 3.53,
|
| 1149 |
+
"learning_rate": 1.6343315056938003e-05,
|
| 1150 |
+
"loss": 0.22,
|
| 1151 |
"step": 1860
|
| 1152 |
},
|
| 1153 |
{
|
| 1154 |
+
"epoch": 3.55,
|
| 1155 |
+
"learning_rate": 1.613243357233235e-05,
|
| 1156 |
+
"loss": 0.2034,
|
| 1157 |
"step": 1870
|
| 1158 |
},
|
| 1159 |
{
|
| 1160 |
+
"epoch": 3.57,
|
| 1161 |
+
"learning_rate": 1.59215520877267e-05,
|
| 1162 |
+
"loss": 0.2088,
|
| 1163 |
"step": 1880
|
| 1164 |
},
|
| 1165 |
{
|
| 1166 |
+
"epoch": 3.59,
|
| 1167 |
+
"learning_rate": 1.5710670603121048e-05,
|
| 1168 |
+
"loss": 0.1922,
|
| 1169 |
"step": 1890
|
| 1170 |
},
|
| 1171 |
{
|
| 1172 |
+
"epoch": 3.6,
|
| 1173 |
+
"learning_rate": 1.5499789118515397e-05,
|
| 1174 |
+
"loss": 0.2146,
|
| 1175 |
"step": 1900
|
| 1176 |
},
|
| 1177 |
{
|
| 1178 |
+
"epoch": 3.62,
|
| 1179 |
+
"learning_rate": 1.5288907633909745e-05,
|
| 1180 |
+
"loss": 0.2198,
|
| 1181 |
"step": 1910
|
| 1182 |
},
|
| 1183 |
{
|
| 1184 |
+
"epoch": 3.64,
|
| 1185 |
+
"learning_rate": 1.5078026149304092e-05,
|
| 1186 |
+
"loss": 0.2008,
|
| 1187 |
"step": 1920
|
| 1188 |
},
|
| 1189 |
{
|
| 1190 |
+
"epoch": 3.66,
|
| 1191 |
+
"learning_rate": 1.486714466469844e-05,
|
| 1192 |
+
"loss": 0.204,
|
| 1193 |
"step": 1930
|
| 1194 |
},
|
| 1195 |
{
|
| 1196 |
+
"epoch": 3.68,
|
| 1197 |
+
"learning_rate": 1.4656263180092789e-05,
|
| 1198 |
+
"loss": 0.2053,
|
| 1199 |
"step": 1940
|
| 1200 |
},
|
| 1201 |
{
|
| 1202 |
+
"epoch": 3.7,
|
| 1203 |
+
"learning_rate": 1.4445381695487137e-05,
|
| 1204 |
+
"loss": 0.2194,
|
| 1205 |
"step": 1950
|
| 1206 |
},
|
| 1207 |
{
|
| 1208 |
+
"epoch": 3.72,
|
| 1209 |
+
"learning_rate": 1.4234500210881486e-05,
|
| 1210 |
+
"loss": 0.1869,
|
| 1211 |
"step": 1960
|
| 1212 |
},
|
| 1213 |
{
|
| 1214 |
+
"epoch": 3.74,
|
| 1215 |
+
"learning_rate": 1.4023618726275834e-05,
|
| 1216 |
+
"loss": 0.2008,
|
| 1217 |
"step": 1970
|
| 1218 |
},
|
| 1219 |
{
|
| 1220 |
+
"epoch": 3.76,
|
| 1221 |
+
"learning_rate": 1.3812737241670182e-05,
|
| 1222 |
+
"loss": 0.2011,
|
| 1223 |
"step": 1980
|
| 1224 |
},
|
| 1225 |
{
|
| 1226 |
+
"epoch": 3.77,
|
| 1227 |
+
"learning_rate": 1.360185575706453e-05,
|
| 1228 |
+
"loss": 0.1867,
|
| 1229 |
+
"step": 1990
|
|
|
|
|
|
|
|
|
|
| 1230 |
},
|
| 1231 |
{
|
| 1232 |
+
"epoch": 3.79,
|
| 1233 |
+
"learning_rate": 1.3390974272458878e-05,
|
| 1234 |
+
"loss": 0.1892,
|
| 1235 |
+
"step": 2000
|
| 1236 |
+
},
|
| 1237 |
+
{
|
| 1238 |
+
"epoch": 3.81,
|
| 1239 |
+
"learning_rate": 1.3180092787853226e-05,
|
| 1240 |
+
"loss": 0.1646,
|
| 1241 |
+
"step": 2010
|
| 1242 |
+
},
|
| 1243 |
+
{
|
| 1244 |
+
"epoch": 3.83,
|
| 1245 |
+
"learning_rate": 1.2969211303247574e-05,
|
| 1246 |
+
"loss": 0.2096,
|
| 1247 |
+
"step": 2020
|
| 1248 |
+
},
|
| 1249 |
+
{
|
| 1250 |
+
"epoch": 3.85,
|
| 1251 |
+
"learning_rate": 1.2758329818641923e-05,
|
| 1252 |
+
"loss": 0.1834,
|
| 1253 |
+
"step": 2030
|
| 1254 |
+
},
|
| 1255 |
+
{
|
| 1256 |
+
"epoch": 3.87,
|
| 1257 |
+
"learning_rate": 1.2547448334036271e-05,
|
| 1258 |
+
"loss": 0.1909,
|
| 1259 |
+
"step": 2040
|
| 1260 |
+
},
|
| 1261 |
+
{
|
| 1262 |
+
"epoch": 3.89,
|
| 1263 |
+
"learning_rate": 1.2336566849430621e-05,
|
| 1264 |
+
"loss": 0.1867,
|
| 1265 |
+
"step": 2050
|
| 1266 |
+
},
|
| 1267 |
+
{
|
| 1268 |
+
"epoch": 3.91,
|
| 1269 |
+
"learning_rate": 1.212568536482497e-05,
|
| 1270 |
+
"loss": 0.197,
|
| 1271 |
+
"step": 2060
|
| 1272 |
+
},
|
| 1273 |
+
{
|
| 1274 |
+
"epoch": 3.93,
|
| 1275 |
+
"learning_rate": 1.1914803880219318e-05,
|
| 1276 |
+
"loss": 0.2021,
|
| 1277 |
+
"step": 2070
|
| 1278 |
+
},
|
| 1279 |
+
{
|
| 1280 |
+
"epoch": 3.95,
|
| 1281 |
+
"learning_rate": 1.1703922395613665e-05,
|
| 1282 |
+
"loss": 0.1906,
|
| 1283 |
+
"step": 2080
|
| 1284 |
+
},
|
| 1285 |
+
{
|
| 1286 |
+
"epoch": 3.96,
|
| 1287 |
+
"learning_rate": 1.1493040911008014e-05,
|
| 1288 |
+
"loss": 0.1851,
|
| 1289 |
+
"step": 2090
|
| 1290 |
+
},
|
| 1291 |
+
{
|
| 1292 |
+
"epoch": 3.98,
|
| 1293 |
+
"learning_rate": 1.1282159426402362e-05,
|
| 1294 |
+
"loss": 0.22,
|
| 1295 |
+
"step": 2100
|
| 1296 |
+
},
|
| 1297 |
+
{
|
| 1298 |
+
"epoch": 4.0,
|
| 1299 |
+
"eval_accuracy": 0.9701492537313433,
|
| 1300 |
+
"eval_loss": 0.08384203910827637,
|
| 1301 |
+
"eval_runtime": 131.303,
|
| 1302 |
+
"eval_samples_per_second": 73.479,
|
| 1303 |
+
"eval_steps_per_second": 2.3,
|
| 1304 |
+
"step": 2108
|
| 1305 |
+
},
|
| 1306 |
+
{
|
| 1307 |
+
"epoch": 4.0,
|
| 1308 |
+
"learning_rate": 1.107127794179671e-05,
|
| 1309 |
+
"loss": 0.2109,
|
| 1310 |
+
"step": 2110
|
| 1311 |
+
},
|
| 1312 |
+
{
|
| 1313 |
+
"epoch": 4.02,
|
| 1314 |
+
"learning_rate": 1.0860396457191059e-05,
|
| 1315 |
+
"loss": 0.1934,
|
| 1316 |
+
"step": 2120
|
| 1317 |
+
},
|
| 1318 |
+
{
|
| 1319 |
+
"epoch": 4.04,
|
| 1320 |
+
"learning_rate": 1.0649514972585407e-05,
|
| 1321 |
+
"loss": 0.1984,
|
| 1322 |
+
"step": 2130
|
| 1323 |
+
},
|
| 1324 |
+
{
|
| 1325 |
+
"epoch": 4.06,
|
| 1326 |
+
"learning_rate": 1.0438633487979756e-05,
|
| 1327 |
+
"loss": 0.2011,
|
| 1328 |
+
"step": 2140
|
| 1329 |
+
},
|
| 1330 |
+
{
|
| 1331 |
+
"epoch": 4.08,
|
| 1332 |
+
"learning_rate": 1.0227752003374104e-05,
|
| 1333 |
+
"loss": 0.1784,
|
| 1334 |
+
"step": 2150
|
| 1335 |
+
},
|
| 1336 |
+
{
|
| 1337 |
+
"epoch": 4.1,
|
| 1338 |
+
"learning_rate": 1.0016870518768453e-05,
|
| 1339 |
+
"loss": 0.2025,
|
| 1340 |
+
"step": 2160
|
| 1341 |
+
},
|
| 1342 |
+
{
|
| 1343 |
+
"epoch": 4.12,
|
| 1344 |
+
"learning_rate": 9.805989034162801e-06,
|
| 1345 |
+
"loss": 0.1766,
|
| 1346 |
+
"step": 2170
|
| 1347 |
+
},
|
| 1348 |
+
{
|
| 1349 |
+
"epoch": 4.14,
|
| 1350 |
+
"learning_rate": 9.59510754955715e-06,
|
| 1351 |
+
"loss": 0.1866,
|
| 1352 |
+
"step": 2180
|
| 1353 |
+
},
|
| 1354 |
+
{
|
| 1355 |
+
"epoch": 4.16,
|
| 1356 |
+
"learning_rate": 9.384226064951498e-06,
|
| 1357 |
+
"loss": 0.1985,
|
| 1358 |
+
"step": 2190
|
| 1359 |
+
},
|
| 1360 |
+
{
|
| 1361 |
+
"epoch": 4.17,
|
| 1362 |
+
"learning_rate": 9.173344580345846e-06,
|
| 1363 |
+
"loss": 0.174,
|
| 1364 |
+
"step": 2200
|
| 1365 |
+
},
|
| 1366 |
+
{
|
| 1367 |
+
"epoch": 4.19,
|
| 1368 |
+
"learning_rate": 8.962463095740195e-06,
|
| 1369 |
+
"loss": 0.2045,
|
| 1370 |
+
"step": 2210
|
| 1371 |
+
},
|
| 1372 |
+
{
|
| 1373 |
+
"epoch": 4.21,
|
| 1374 |
+
"learning_rate": 8.751581611134543e-06,
|
| 1375 |
+
"loss": 0.1847,
|
| 1376 |
+
"step": 2220
|
| 1377 |
+
},
|
| 1378 |
+
{
|
| 1379 |
+
"epoch": 4.23,
|
| 1380 |
+
"learning_rate": 8.540700126528892e-06,
|
| 1381 |
+
"loss": 0.1785,
|
| 1382 |
+
"step": 2230
|
| 1383 |
+
},
|
| 1384 |
+
{
|
| 1385 |
+
"epoch": 4.25,
|
| 1386 |
+
"learning_rate": 8.32981864192324e-06,
|
| 1387 |
+
"loss": 0.1421,
|
| 1388 |
+
"step": 2240
|
| 1389 |
+
},
|
| 1390 |
+
{
|
| 1391 |
+
"epoch": 4.27,
|
| 1392 |
+
"learning_rate": 8.118937157317589e-06,
|
| 1393 |
+
"loss": 0.162,
|
| 1394 |
+
"step": 2250
|
| 1395 |
+
},
|
| 1396 |
+
{
|
| 1397 |
+
"epoch": 4.29,
|
| 1398 |
+
"learning_rate": 7.908055672711937e-06,
|
| 1399 |
+
"loss": 0.1989,
|
| 1400 |
+
"step": 2260
|
| 1401 |
+
},
|
| 1402 |
+
{
|
| 1403 |
+
"epoch": 4.31,
|
| 1404 |
+
"learning_rate": 7.697174188106285e-06,
|
| 1405 |
+
"loss": 0.1748,
|
| 1406 |
+
"step": 2270
|
| 1407 |
+
},
|
| 1408 |
+
{
|
| 1409 |
+
"epoch": 4.33,
|
| 1410 |
+
"learning_rate": 7.486292703500633e-06,
|
| 1411 |
+
"loss": 0.1702,
|
| 1412 |
+
"step": 2280
|
| 1413 |
+
},
|
| 1414 |
+
{
|
| 1415 |
+
"epoch": 4.34,
|
| 1416 |
+
"learning_rate": 7.2754112188949814e-06,
|
| 1417 |
+
"loss": 0.1721,
|
| 1418 |
+
"step": 2290
|
| 1419 |
+
},
|
| 1420 |
+
{
|
| 1421 |
+
"epoch": 4.36,
|
| 1422 |
+
"learning_rate": 7.06452973428933e-06,
|
| 1423 |
+
"loss": 0.1754,
|
| 1424 |
+
"step": 2300
|
| 1425 |
+
},
|
| 1426 |
+
{
|
| 1427 |
+
"epoch": 4.38,
|
| 1428 |
+
"learning_rate": 6.853648249683678e-06,
|
| 1429 |
+
"loss": 0.175,
|
| 1430 |
+
"step": 2310
|
| 1431 |
+
},
|
| 1432 |
+
{
|
| 1433 |
+
"epoch": 4.4,
|
| 1434 |
+
"learning_rate": 6.642766765078026e-06,
|
| 1435 |
+
"loss": 0.1906,
|
| 1436 |
+
"step": 2320
|
| 1437 |
+
},
|
| 1438 |
+
{
|
| 1439 |
+
"epoch": 4.42,
|
| 1440 |
+
"learning_rate": 6.431885280472374e-06,
|
| 1441 |
+
"loss": 0.1789,
|
| 1442 |
+
"step": 2330
|
| 1443 |
+
},
|
| 1444 |
+
{
|
| 1445 |
+
"epoch": 4.44,
|
| 1446 |
+
"learning_rate": 6.221003795866724e-06,
|
| 1447 |
+
"loss": 0.1607,
|
| 1448 |
+
"step": 2340
|
| 1449 |
+
},
|
| 1450 |
+
{
|
| 1451 |
+
"epoch": 4.46,
|
| 1452 |
+
"learning_rate": 6.010122311261072e-06,
|
| 1453 |
+
"loss": 0.1747,
|
| 1454 |
+
"step": 2350
|
| 1455 |
+
},
|
| 1456 |
+
{
|
| 1457 |
+
"epoch": 4.48,
|
| 1458 |
+
"learning_rate": 5.79924082665542e-06,
|
| 1459 |
+
"loss": 0.1611,
|
| 1460 |
+
"step": 2360
|
| 1461 |
+
},
|
| 1462 |
+
{
|
| 1463 |
+
"epoch": 4.5,
|
| 1464 |
+
"learning_rate": 5.588359342049768e-06,
|
| 1465 |
+
"loss": 0.1611,
|
| 1466 |
+
"step": 2370
|
| 1467 |
+
},
|
| 1468 |
+
{
|
| 1469 |
+
"epoch": 4.52,
|
| 1470 |
+
"learning_rate": 5.3774778574441165e-06,
|
| 1471 |
+
"loss": 0.1674,
|
| 1472 |
+
"step": 2380
|
| 1473 |
+
},
|
| 1474 |
+
{
|
| 1475 |
+
"epoch": 4.53,
|
| 1476 |
+
"learning_rate": 5.166596372838465e-06,
|
| 1477 |
+
"loss": 0.1797,
|
| 1478 |
+
"step": 2390
|
| 1479 |
+
},
|
| 1480 |
+
{
|
| 1481 |
+
"epoch": 4.55,
|
| 1482 |
+
"learning_rate": 4.955714888232813e-06,
|
| 1483 |
+
"loss": 0.2051,
|
| 1484 |
+
"step": 2400
|
| 1485 |
+
},
|
| 1486 |
+
{
|
| 1487 |
+
"epoch": 4.57,
|
| 1488 |
+
"learning_rate": 4.744833403627162e-06,
|
| 1489 |
+
"loss": 0.171,
|
| 1490 |
+
"step": 2410
|
| 1491 |
+
},
|
| 1492 |
+
{
|
| 1493 |
+
"epoch": 4.59,
|
| 1494 |
+
"learning_rate": 4.53395191902151e-06,
|
| 1495 |
+
"loss": 0.1622,
|
| 1496 |
+
"step": 2420
|
| 1497 |
+
},
|
| 1498 |
+
{
|
| 1499 |
+
"epoch": 4.61,
|
| 1500 |
+
"learning_rate": 4.323070434415859e-06,
|
| 1501 |
+
"loss": 0.1891,
|
| 1502 |
+
"step": 2430
|
| 1503 |
+
},
|
| 1504 |
+
{
|
| 1505 |
+
"epoch": 4.63,
|
| 1506 |
+
"learning_rate": 4.112188949810207e-06,
|
| 1507 |
+
"loss": 0.1884,
|
| 1508 |
+
"step": 2440
|
| 1509 |
+
},
|
| 1510 |
+
{
|
| 1511 |
+
"epoch": 4.65,
|
| 1512 |
+
"learning_rate": 3.901307465204556e-06,
|
| 1513 |
+
"loss": 0.1946,
|
| 1514 |
+
"step": 2450
|
| 1515 |
+
},
|
| 1516 |
+
{
|
| 1517 |
+
"epoch": 4.67,
|
| 1518 |
+
"learning_rate": 3.6904259805989036e-06,
|
| 1519 |
+
"loss": 0.1769,
|
| 1520 |
+
"step": 2460
|
| 1521 |
+
},
|
| 1522 |
+
{
|
| 1523 |
+
"epoch": 4.69,
|
| 1524 |
+
"learning_rate": 3.479544495993252e-06,
|
| 1525 |
+
"loss": 0.19,
|
| 1526 |
+
"step": 2470
|
| 1527 |
+
},
|
| 1528 |
+
{
|
| 1529 |
+
"epoch": 4.7,
|
| 1530 |
+
"learning_rate": 3.2686630113876e-06,
|
| 1531 |
+
"loss": 0.1741,
|
| 1532 |
+
"step": 2480
|
| 1533 |
+
},
|
| 1534 |
+
{
|
| 1535 |
+
"epoch": 4.72,
|
| 1536 |
+
"learning_rate": 3.057781526781949e-06,
|
| 1537 |
+
"loss": 0.1697,
|
| 1538 |
+
"step": 2490
|
| 1539 |
+
},
|
| 1540 |
+
{
|
| 1541 |
+
"epoch": 4.74,
|
| 1542 |
+
"learning_rate": 2.846900042176297e-06,
|
| 1543 |
+
"loss": 0.1901,
|
| 1544 |
+
"step": 2500
|
| 1545 |
+
},
|
| 1546 |
+
{
|
| 1547 |
+
"epoch": 4.76,
|
| 1548 |
+
"learning_rate": 2.6360185575706454e-06,
|
| 1549 |
+
"loss": 0.1794,
|
| 1550 |
+
"step": 2510
|
| 1551 |
+
},
|
| 1552 |
+
{
|
| 1553 |
+
"epoch": 4.78,
|
| 1554 |
+
"learning_rate": 2.425137072964994e-06,
|
| 1555 |
+
"loss": 0.1514,
|
| 1556 |
+
"step": 2520
|
| 1557 |
+
},
|
| 1558 |
+
{
|
| 1559 |
+
"epoch": 4.8,
|
| 1560 |
+
"learning_rate": 2.2142555883593423e-06,
|
| 1561 |
+
"loss": 0.1841,
|
| 1562 |
+
"step": 2530
|
| 1563 |
+
},
|
| 1564 |
+
{
|
| 1565 |
+
"epoch": 4.82,
|
| 1566 |
+
"learning_rate": 2.0033741037536907e-06,
|
| 1567 |
+
"loss": 0.1676,
|
| 1568 |
+
"step": 2540
|
| 1569 |
+
},
|
| 1570 |
+
{
|
| 1571 |
+
"epoch": 4.84,
|
| 1572 |
+
"learning_rate": 1.792492619148039e-06,
|
| 1573 |
+
"loss": 0.1723,
|
| 1574 |
+
"step": 2550
|
| 1575 |
+
},
|
| 1576 |
+
{
|
| 1577 |
+
"epoch": 4.86,
|
| 1578 |
+
"learning_rate": 1.5816111345423871e-06,
|
| 1579 |
+
"loss": 0.186,
|
| 1580 |
+
"step": 2560
|
| 1581 |
+
},
|
| 1582 |
+
{
|
| 1583 |
+
"epoch": 4.88,
|
| 1584 |
+
"learning_rate": 1.3707296499367356e-06,
|
| 1585 |
+
"loss": 0.1669,
|
| 1586 |
+
"step": 2570
|
| 1587 |
+
},
|
| 1588 |
+
{
|
| 1589 |
+
"epoch": 4.89,
|
| 1590 |
+
"learning_rate": 1.1598481653310838e-06,
|
| 1591 |
+
"loss": 0.1801,
|
| 1592 |
+
"step": 2580
|
| 1593 |
+
},
|
| 1594 |
+
{
|
| 1595 |
+
"epoch": 4.91,
|
| 1596 |
+
"learning_rate": 9.489666807254323e-07,
|
| 1597 |
+
"loss": 0.1469,
|
| 1598 |
+
"step": 2590
|
| 1599 |
+
},
|
| 1600 |
+
{
|
| 1601 |
+
"epoch": 4.93,
|
| 1602 |
+
"learning_rate": 7.380851961197808e-07,
|
| 1603 |
+
"loss": 0.1642,
|
| 1604 |
+
"step": 2600
|
| 1605 |
+
},
|
| 1606 |
+
{
|
| 1607 |
+
"epoch": 4.95,
|
| 1608 |
+
"learning_rate": 5.272037115141291e-07,
|
| 1609 |
+
"loss": 0.1691,
|
| 1610 |
+
"step": 2610
|
| 1611 |
+
},
|
| 1612 |
+
{
|
| 1613 |
+
"epoch": 4.97,
|
| 1614 |
+
"learning_rate": 3.1632222690847745e-07,
|
| 1615 |
+
"loss": 0.1644,
|
| 1616 |
+
"step": 2620
|
| 1617 |
+
},
|
| 1618 |
+
{
|
| 1619 |
+
"epoch": 4.99,
|
| 1620 |
+
"learning_rate": 1.0544074230282581e-07,
|
| 1621 |
+
"loss": 0.173,
|
| 1622 |
+
"step": 2630
|
| 1623 |
+
},
|
| 1624 |
+
{
|
| 1625 |
+
"epoch": 5.0,
|
| 1626 |
+
"eval_accuracy": 0.9741915422885572,
|
| 1627 |
+
"eval_loss": 0.06838314980268478,
|
| 1628 |
+
"eval_runtime": 133.6073,
|
| 1629 |
+
"eval_samples_per_second": 72.212,
|
| 1630 |
+
"eval_steps_per_second": 2.26,
|
| 1631 |
+
"step": 2635
|
| 1632 |
+
},
|
| 1633 |
+
{
|
| 1634 |
+
"epoch": 5.0,
|
| 1635 |
+
"step": 2635,
|
| 1636 |
+
"total_flos": 8.391154825876193e+18,
|
| 1637 |
+
"train_loss": 0.33317647269826234,
|
| 1638 |
+
"train_runtime": 8035.3313,
|
| 1639 |
+
"train_samples_per_second": 42.02,
|
| 1640 |
+
"train_steps_per_second": 0.328
|
| 1641 |
}
|
| 1642 |
],
|
| 1643 |
+
"max_steps": 2635,
|
| 1644 |
+
"num_train_epochs": 5,
|
| 1645 |
+
"total_flos": 8.391154825876193e+18,
|
| 1646 |
"trial_name": null,
|
| 1647 |
"trial_params": null
|
| 1648 |
}
|