End of training
Browse files
all_results.json
CHANGED
|
@@ -1,8 +1,8 @@
|
|
| 1 |
{
|
| 2 |
"epoch": 3.25,
|
| 3 |
-
"eval_accuracy": 0.
|
| 4 |
-
"eval_loss": 0.
|
| 5 |
-
"eval_runtime":
|
| 6 |
-
"eval_samples_per_second": 3.
|
| 7 |
-
"eval_steps_per_second": 3.
|
| 8 |
}
|
|
|
|
| 1 |
{
|
| 2 |
"epoch": 3.25,
|
| 3 |
+
"eval_accuracy": 0.8114285714285714,
|
| 4 |
+
"eval_loss": 0.6237285733222961,
|
| 5 |
+
"eval_runtime": 52.2354,
|
| 6 |
+
"eval_samples_per_second": 3.35,
|
| 7 |
+
"eval_steps_per_second": 3.35
|
| 8 |
}
|
runs/Mar30_05-44-53_04b9fd5247a2/events.out.tfevents.1680156513.04b9fd5247a2.349.2
CHANGED
|
@@ -1,3 +1,3 @@
|
|
| 1 |
version https://git-lfs.github.com/spec/v1
|
| 2 |
-
oid sha256:
|
| 3 |
-
size
|
|
|
|
| 1 |
version https://git-lfs.github.com/spec/v1
|
| 2 |
+
oid sha256:3128ea3dbffe6a2758e0e7e312dd62d7413061e83640bffc1853870f7f6ee87a
|
| 3 |
+
size 686
|
test_results.json
CHANGED
|
@@ -1,8 +1,8 @@
|
|
| 1 |
{
|
| 2 |
"epoch": 3.25,
|
| 3 |
-
"eval_accuracy": 0.
|
| 4 |
-
"eval_loss": 0.
|
| 5 |
-
"eval_runtime":
|
| 6 |
-
"eval_samples_per_second": 3.
|
| 7 |
-
"eval_steps_per_second": 3.
|
| 8 |
}
|
|
|
|
| 1 |
{
|
| 2 |
"epoch": 3.25,
|
| 3 |
+
"eval_accuracy": 0.8114285714285714,
|
| 4 |
+
"eval_loss": 0.6237285733222961,
|
| 5 |
+
"eval_runtime": 52.2354,
|
| 6 |
+
"eval_samples_per_second": 3.35,
|
| 7 |
+
"eval_steps_per_second": 3.35
|
| 8 |
}
|
trainer_state.json
CHANGED
|
@@ -1,883 +1,955 @@
|
|
| 1 |
{
|
| 2 |
-
"best_metric": 0.
|
| 3 |
-
"best_model_checkpoint": "videomae-base-finetuned-ucf101-subset/checkpoint-
|
| 4 |
"epoch": 3.25,
|
| 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": 0.01,
|
| 12 |
-
"learning_rate": 3.
|
| 13 |
-
"loss": 2.
|
| 14 |
"step": 10
|
| 15 |
},
|
| 16 |
{
|
| 17 |
"epoch": 0.01,
|
| 18 |
-
"learning_rate":
|
| 19 |
-
"loss": 2.
|
| 20 |
"step": 20
|
| 21 |
},
|
| 22 |
{
|
| 23 |
"epoch": 0.02,
|
| 24 |
-
"learning_rate": 1.
|
| 25 |
-
"loss": 2.
|
| 26 |
"step": 30
|
| 27 |
},
|
| 28 |
{
|
| 29 |
"epoch": 0.03,
|
| 30 |
-
"learning_rate": 1.
|
| 31 |
-
"loss": 2.
|
| 32 |
"step": 40
|
| 33 |
},
|
| 34 |
{
|
| 35 |
-
"epoch": 0.
|
| 36 |
-
"learning_rate": 1.
|
| 37 |
-
"loss": 2.
|
| 38 |
"step": 50
|
| 39 |
},
|
| 40 |
{
|
| 41 |
"epoch": 0.04,
|
| 42 |
-
"learning_rate": 2.
|
| 43 |
-
"loss": 2.
|
| 44 |
"step": 60
|
| 45 |
},
|
| 46 |
{
|
| 47 |
"epoch": 0.05,
|
| 48 |
-
"learning_rate": 2.
|
| 49 |
-
"loss": 2.
|
| 50 |
"step": 70
|
| 51 |
},
|
| 52 |
{
|
| 53 |
-
"epoch": 0.
|
| 54 |
-
"learning_rate": 2.
|
| 55 |
-
"loss": 2.
|
| 56 |
"step": 80
|
| 57 |
},
|
| 58 |
{
|
| 59 |
-
"epoch": 0.
|
| 60 |
-
"learning_rate": 3.
|
| 61 |
-
"loss": 2.
|
| 62 |
"step": 90
|
| 63 |
},
|
| 64 |
{
|
| 65 |
"epoch": 0.07,
|
| 66 |
-
"learning_rate": 3.
|
| 67 |
-
"loss": 2.
|
| 68 |
"step": 100
|
| 69 |
},
|
| 70 |
{
|
| 71 |
"epoch": 0.08,
|
| 72 |
-
"learning_rate":
|
| 73 |
-
"loss": 2.
|
| 74 |
"step": 110
|
| 75 |
},
|
| 76 |
{
|
| 77 |
-
"epoch": 0.
|
| 78 |
-
"learning_rate": 4.
|
| 79 |
-
"loss": 2.
|
| 80 |
"step": 120
|
| 81 |
},
|
| 82 |
{
|
| 83 |
-
"epoch": 0.
|
| 84 |
-
"learning_rate": 4.
|
| 85 |
-
"loss": 2.
|
| 86 |
"step": 130
|
| 87 |
},
|
| 88 |
{
|
| 89 |
"epoch": 0.1,
|
| 90 |
-
"learning_rate": 4.
|
| 91 |
-
"loss": 2.
|
| 92 |
"step": 140
|
| 93 |
},
|
| 94 |
{
|
| 95 |
-
"epoch": 0.
|
| 96 |
-
"learning_rate": 4.
|
| 97 |
-
"loss": 2.
|
| 98 |
"step": 150
|
| 99 |
},
|
| 100 |
{
|
| 101 |
-
"epoch": 0.
|
| 102 |
-
"learning_rate": 4.
|
| 103 |
-
"loss": 2.
|
| 104 |
"step": 160
|
| 105 |
},
|
| 106 |
{
|
| 107 |
-
"epoch": 0.
|
| 108 |
-
"learning_rate": 4.
|
| 109 |
-
"loss":
|
| 110 |
"step": 170
|
| 111 |
},
|
| 112 |
{
|
| 113 |
-
"epoch": 0.
|
| 114 |
-
"learning_rate": 4.
|
| 115 |
-
"loss":
|
| 116 |
"step": 180
|
| 117 |
},
|
| 118 |
{
|
| 119 |
-
"epoch": 0.
|
| 120 |
-
"learning_rate": 4.
|
| 121 |
-
"loss":
|
| 122 |
"step": 190
|
| 123 |
},
|
| 124 |
{
|
| 125 |
-
"epoch": 0.
|
| 126 |
-
"learning_rate": 4.
|
| 127 |
-
"loss": 2.
|
| 128 |
"step": 200
|
| 129 |
},
|
| 130 |
{
|
| 131 |
-
"epoch": 0.
|
| 132 |
-
"learning_rate": 4.
|
| 133 |
-
"loss": 2.
|
| 134 |
"step": 210
|
| 135 |
},
|
| 136 |
{
|
| 137 |
-
"epoch": 0.
|
| 138 |
-
"learning_rate": 4.
|
| 139 |
-
"loss": 2.
|
| 140 |
"step": 220
|
| 141 |
},
|
| 142 |
{
|
| 143 |
-
"epoch": 0.
|
| 144 |
-
"learning_rate": 4.
|
| 145 |
-
"loss": 2.
|
| 146 |
"step": 230
|
| 147 |
},
|
| 148 |
{
|
| 149 |
-
"epoch": 0.
|
| 150 |
-
"learning_rate": 4.
|
| 151 |
-
"loss": 2.
|
| 152 |
"step": 240
|
| 153 |
},
|
| 154 |
{
|
| 155 |
-
"epoch": 0.
|
| 156 |
-
"learning_rate": 4.
|
| 157 |
-
"loss":
|
| 158 |
"step": 250
|
| 159 |
},
|
| 160 |
{
|
| 161 |
-
"epoch": 0.
|
| 162 |
-
"learning_rate": 4.
|
| 163 |
-
"loss":
|
| 164 |
"step": 260
|
| 165 |
},
|
| 166 |
{
|
| 167 |
-
"epoch": 0.
|
| 168 |
-
"learning_rate": 4.
|
| 169 |
-
"loss":
|
| 170 |
"step": 270
|
| 171 |
},
|
| 172 |
{
|
| 173 |
-
"epoch": 0.
|
| 174 |
-
"learning_rate": 4.
|
| 175 |
-
"loss":
|
| 176 |
"step": 280
|
| 177 |
},
|
| 178 |
{
|
| 179 |
-
"epoch": 0.
|
| 180 |
-
"learning_rate": 4.
|
| 181 |
-
"loss": 2.
|
| 182 |
"step": 290
|
| 183 |
},
|
| 184 |
{
|
| 185 |
-
"epoch": 0.
|
| 186 |
-
"learning_rate": 4.
|
| 187 |
-
"loss":
|
| 188 |
"step": 300
|
| 189 |
},
|
| 190 |
{
|
| 191 |
-
"epoch": 0.
|
| 192 |
-
"learning_rate": 4.
|
| 193 |
-
"loss":
|
| 194 |
"step": 310
|
| 195 |
},
|
| 196 |
{
|
| 197 |
-
"epoch": 0.
|
| 198 |
-
"learning_rate": 4.
|
| 199 |
-
"loss":
|
| 200 |
"step": 320
|
| 201 |
},
|
| 202 |
{
|
| 203 |
-
"epoch": 0.
|
| 204 |
-
"learning_rate": 4.
|
| 205 |
-
"loss":
|
| 206 |
"step": 330
|
| 207 |
},
|
| 208 |
{
|
| 209 |
-
"epoch": 0.
|
| 210 |
-
"
|
| 211 |
-
"
|
| 212 |
-
"eval_runtime": 21.5381,
|
| 213 |
-
"eval_samples_per_second": 3.436,
|
| 214 |
-
"eval_steps_per_second": 3.436,
|
| 215 |
-
"step": 335
|
| 216 |
-
},
|
| 217 |
-
{
|
| 218 |
-
"epoch": 1.0,
|
| 219 |
-
"learning_rate": 4.145936981757877e-05,
|
| 220 |
-
"loss": 2.0567,
|
| 221 |
"step": 340
|
| 222 |
},
|
| 223 |
{
|
| 224 |
-
"epoch":
|
| 225 |
-
"learning_rate": 4.
|
| 226 |
-
"loss":
|
| 227 |
"step": 350
|
| 228 |
},
|
| 229 |
{
|
| 230 |
-
"epoch":
|
| 231 |
-
"learning_rate": 4.
|
| 232 |
-
"loss":
|
| 233 |
"step": 360
|
| 234 |
},
|
| 235 |
{
|
| 236 |
-
"epoch":
|
| 237 |
-
"
|
| 238 |
-
"
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
| 239 |
"step": 370
|
| 240 |
},
|
| 241 |
{
|
| 242 |
-
"epoch": 1.
|
| 243 |
-
"learning_rate":
|
| 244 |
-
"loss": 1.
|
| 245 |
"step": 380
|
| 246 |
},
|
| 247 |
{
|
| 248 |
-
"epoch": 1.
|
| 249 |
-
"learning_rate":
|
| 250 |
-
"loss": 1.
|
| 251 |
"step": 390
|
| 252 |
},
|
| 253 |
{
|
| 254 |
-
"epoch": 1.
|
| 255 |
-
"learning_rate":
|
| 256 |
-
"loss": 1.
|
| 257 |
"step": 400
|
| 258 |
},
|
| 259 |
{
|
| 260 |
-
"epoch": 1.
|
| 261 |
-
"learning_rate": 3.
|
| 262 |
-
"loss":
|
| 263 |
"step": 410
|
| 264 |
},
|
| 265 |
{
|
| 266 |
-
"epoch": 1.
|
| 267 |
-
"learning_rate": 3.
|
| 268 |
-
"loss":
|
| 269 |
"step": 420
|
| 270 |
},
|
| 271 |
{
|
| 272 |
-
"epoch": 1.
|
| 273 |
-
"learning_rate": 3.
|
| 274 |
-
"loss":
|
| 275 |
"step": 430
|
| 276 |
},
|
| 277 |
{
|
| 278 |
-
"epoch": 1.
|
| 279 |
-
"learning_rate": 3.
|
| 280 |
-
"loss":
|
| 281 |
"step": 440
|
| 282 |
},
|
| 283 |
{
|
| 284 |
-
"epoch": 1.
|
| 285 |
-
"learning_rate": 3.
|
| 286 |
-
"loss": 1.
|
| 287 |
"step": 450
|
| 288 |
},
|
| 289 |
{
|
| 290 |
-
"epoch": 1.
|
| 291 |
-
"learning_rate": 3.
|
| 292 |
-
"loss":
|
| 293 |
"step": 460
|
| 294 |
},
|
| 295 |
{
|
| 296 |
-
"epoch": 1.
|
| 297 |
-
"learning_rate": 3.
|
| 298 |
-
"loss":
|
| 299 |
"step": 470
|
| 300 |
},
|
| 301 |
{
|
| 302 |
-
"epoch": 1.
|
| 303 |
-
"learning_rate": 3.
|
| 304 |
-
"loss":
|
| 305 |
"step": 480
|
| 306 |
},
|
| 307 |
{
|
| 308 |
-
"epoch": 1.
|
| 309 |
-
"learning_rate": 3.
|
| 310 |
-
"loss": 1.
|
| 311 |
"step": 490
|
| 312 |
},
|
| 313 |
{
|
| 314 |
-
"epoch": 1.
|
| 315 |
-
"learning_rate": 3.
|
| 316 |
-
"loss":
|
| 317 |
"step": 500
|
| 318 |
},
|
| 319 |
{
|
| 320 |
-
"epoch": 1.
|
| 321 |
-
"learning_rate": 3.
|
| 322 |
-
"loss": 1.
|
| 323 |
"step": 510
|
| 324 |
},
|
| 325 |
{
|
| 326 |
-
"epoch": 1.
|
| 327 |
-
"learning_rate": 3.
|
| 328 |
-
"loss":
|
| 329 |
"step": 520
|
| 330 |
},
|
| 331 |
{
|
| 332 |
-
"epoch": 1.
|
| 333 |
-
"learning_rate": 3.
|
| 334 |
-
"loss":
|
| 335 |
"step": 530
|
| 336 |
},
|
| 337 |
{
|
| 338 |
-
"epoch": 1.
|
| 339 |
-
"learning_rate": 3.
|
| 340 |
-
"loss":
|
| 341 |
"step": 540
|
| 342 |
},
|
| 343 |
{
|
| 344 |
-
"epoch": 1.
|
| 345 |
-
"learning_rate": 3.
|
| 346 |
-
"loss": 1.
|
| 347 |
"step": 550
|
| 348 |
},
|
| 349 |
{
|
| 350 |
-
"epoch": 1.
|
| 351 |
-
"learning_rate": 3.
|
| 352 |
-
"loss":
|
| 353 |
"step": 560
|
| 354 |
},
|
| 355 |
{
|
| 356 |
-
"epoch": 1.
|
| 357 |
-
"learning_rate": 3.
|
| 358 |
-
"loss": 1.
|
| 359 |
"step": 570
|
| 360 |
},
|
| 361 |
{
|
| 362 |
-
"epoch": 1.
|
| 363 |
-
"learning_rate": 3.
|
| 364 |
-
"loss": 1.
|
| 365 |
"step": 580
|
| 366 |
},
|
| 367 |
{
|
| 368 |
-
"epoch": 1.
|
| 369 |
-
"learning_rate": 3.
|
| 370 |
-
"loss": 1.
|
| 371 |
"step": 590
|
| 372 |
},
|
| 373 |
{
|
| 374 |
-
"epoch": 1.
|
| 375 |
-
"learning_rate": 3.
|
| 376 |
-
"loss":
|
| 377 |
"step": 600
|
| 378 |
},
|
| 379 |
{
|
| 380 |
-
"epoch": 1.
|
| 381 |
-
"learning_rate": 3.
|
| 382 |
-
"loss": 1.
|
| 383 |
"step": 610
|
| 384 |
},
|
| 385 |
{
|
| 386 |
-
"epoch": 1.
|
| 387 |
-
"learning_rate":
|
| 388 |
-
"loss": 1.
|
| 389 |
"step": 620
|
| 390 |
},
|
| 391 |
{
|
| 392 |
-
"epoch": 1.
|
| 393 |
-
"learning_rate":
|
| 394 |
-
"loss":
|
| 395 |
"step": 630
|
| 396 |
},
|
| 397 |
{
|
| 398 |
-
"epoch": 1.
|
| 399 |
-
"learning_rate":
|
| 400 |
-
"loss": 1.
|
| 401 |
"step": 640
|
| 402 |
},
|
| 403 |
{
|
| 404 |
-
"epoch": 1.
|
| 405 |
-
"learning_rate":
|
| 406 |
-
"loss":
|
| 407 |
"step": 650
|
| 408 |
},
|
| 409 |
{
|
| 410 |
-
"epoch": 1.
|
| 411 |
-
"learning_rate":
|
| 412 |
-
"loss":
|
| 413 |
"step": 660
|
| 414 |
},
|
| 415 |
{
|
| 416 |
-
"epoch": 1.
|
| 417 |
-
"learning_rate":
|
| 418 |
-
"loss":
|
| 419 |
-
"step": 670
|
| 420 |
-
},
|
| 421 |
-
{
|
| 422 |
-
"epoch": 1.25,
|
| 423 |
-
"eval_accuracy": 0.5,
|
| 424 |
-
"eval_loss": 1.0956870317459106,
|
| 425 |
-
"eval_runtime": 21.8938,
|
| 426 |
-
"eval_samples_per_second": 3.38,
|
| 427 |
-
"eval_steps_per_second": 3.38,
|
| 428 |
"step": 670
|
| 429 |
},
|
| 430 |
{
|
| 431 |
-
"epoch":
|
| 432 |
-
"learning_rate": 2.
|
| 433 |
-
"loss": 0.
|
| 434 |
"step": 680
|
| 435 |
},
|
| 436 |
{
|
| 437 |
-
"epoch":
|
| 438 |
-
"learning_rate": 2.
|
| 439 |
-
"loss": 0.
|
| 440 |
"step": 690
|
| 441 |
},
|
| 442 |
{
|
| 443 |
-
"epoch":
|
| 444 |
-
"learning_rate": 2.
|
| 445 |
-
"loss": 0.
|
| 446 |
"step": 700
|
| 447 |
},
|
| 448 |
{
|
| 449 |
-
"epoch":
|
| 450 |
-
"learning_rate": 2.
|
| 451 |
-
"loss":
|
| 452 |
"step": 710
|
| 453 |
},
|
| 454 |
{
|
| 455 |
-
"epoch":
|
| 456 |
-
"learning_rate": 2.
|
| 457 |
-
"loss":
|
| 458 |
"step": 720
|
| 459 |
},
|
| 460 |
{
|
| 461 |
-
"epoch":
|
| 462 |
-
"learning_rate": 2.
|
| 463 |
-
"loss":
|
| 464 |
"step": 730
|
| 465 |
},
|
| 466 |
{
|
| 467 |
-
"epoch":
|
| 468 |
-
"
|
| 469 |
-
"
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
| 470 |
"step": 740
|
| 471 |
},
|
| 472 |
{
|
| 473 |
-
"epoch": 2.
|
| 474 |
-
"learning_rate": 2.
|
| 475 |
-
"loss":
|
| 476 |
"step": 750
|
| 477 |
},
|
| 478 |
{
|
| 479 |
-
"epoch": 2.
|
| 480 |
-
"learning_rate": 2.
|
| 481 |
-
"loss": 1.
|
| 482 |
"step": 760
|
| 483 |
},
|
| 484 |
{
|
| 485 |
-
"epoch": 2.
|
| 486 |
-
"learning_rate": 2.
|
| 487 |
-
"loss": 0.
|
| 488 |
"step": 770
|
| 489 |
},
|
| 490 |
{
|
| 491 |
-
"epoch": 2.
|
| 492 |
-
"learning_rate": 2.
|
| 493 |
-
"loss":
|
| 494 |
"step": 780
|
| 495 |
},
|
| 496 |
{
|
| 497 |
-
"epoch": 2.
|
| 498 |
-
"learning_rate": 2.
|
| 499 |
-
"loss":
|
| 500 |
"step": 790
|
| 501 |
},
|
| 502 |
{
|
| 503 |
-
"epoch": 2.
|
| 504 |
-
"learning_rate": 2.
|
| 505 |
-
"loss":
|
| 506 |
"step": 800
|
| 507 |
},
|
| 508 |
{
|
| 509 |
-
"epoch": 2.
|
| 510 |
-
"learning_rate": 2.
|
| 511 |
-
"loss":
|
| 512 |
"step": 810
|
| 513 |
},
|
| 514 |
{
|
| 515 |
-
"epoch": 2.
|
| 516 |
-
"learning_rate": 2.
|
| 517 |
-
"loss": 0.
|
| 518 |
"step": 820
|
| 519 |
},
|
| 520 |
{
|
| 521 |
-
"epoch": 2.
|
| 522 |
-
"learning_rate": 2.
|
| 523 |
-
"loss": 1.
|
| 524 |
"step": 830
|
| 525 |
},
|
| 526 |
{
|
| 527 |
-
"epoch": 2.
|
| 528 |
-
"learning_rate": 2.
|
| 529 |
-
"loss":
|
| 530 |
"step": 840
|
| 531 |
},
|
| 532 |
{
|
| 533 |
-
"epoch": 2.
|
| 534 |
-
"learning_rate": 2.
|
| 535 |
-
"loss": 0.
|
| 536 |
"step": 850
|
| 537 |
},
|
| 538 |
{
|
| 539 |
-
"epoch": 2.
|
| 540 |
-
"learning_rate":
|
| 541 |
-
"loss":
|
| 542 |
"step": 860
|
| 543 |
},
|
| 544 |
{
|
| 545 |
-
"epoch": 2.
|
| 546 |
-
"learning_rate":
|
| 547 |
-
"loss": 0.
|
| 548 |
"step": 870
|
| 549 |
},
|
| 550 |
{
|
| 551 |
-
"epoch": 2.
|
| 552 |
-
"learning_rate":
|
| 553 |
-
"loss":
|
| 554 |
"step": 880
|
| 555 |
},
|
| 556 |
{
|
| 557 |
-
"epoch": 2.
|
| 558 |
-
"learning_rate":
|
| 559 |
-
"loss": 0.
|
| 560 |
"step": 890
|
| 561 |
},
|
| 562 |
{
|
| 563 |
-
"epoch": 2.
|
| 564 |
-
"learning_rate":
|
| 565 |
-
"loss":
|
| 566 |
"step": 900
|
| 567 |
},
|
| 568 |
{
|
| 569 |
-
"epoch": 2.
|
| 570 |
-
"learning_rate":
|
| 571 |
-
"loss": 0.
|
| 572 |
"step": 910
|
| 573 |
},
|
| 574 |
{
|
| 575 |
-
"epoch": 2.
|
| 576 |
-
"learning_rate":
|
| 577 |
-
"loss": 0.
|
| 578 |
"step": 920
|
| 579 |
},
|
| 580 |
{
|
| 581 |
-
"epoch": 2.
|
| 582 |
-
"learning_rate":
|
| 583 |
-
"loss": 0.
|
| 584 |
"step": 930
|
| 585 |
},
|
| 586 |
{
|
| 587 |
-
"epoch": 2.
|
| 588 |
-
"learning_rate": 1.
|
| 589 |
-
"loss": 0.
|
| 590 |
"step": 940
|
| 591 |
},
|
| 592 |
{
|
| 593 |
-
"epoch": 2.
|
| 594 |
-
"learning_rate": 1.
|
| 595 |
-
"loss": 0.
|
| 596 |
"step": 950
|
| 597 |
},
|
| 598 |
{
|
| 599 |
-
"epoch": 2.
|
| 600 |
-
"learning_rate": 1.
|
| 601 |
-
"loss":
|
| 602 |
"step": 960
|
| 603 |
},
|
| 604 |
{
|
| 605 |
-
"epoch": 2.
|
| 606 |
-
"learning_rate": 1.
|
| 607 |
-
"loss": 0.
|
| 608 |
"step": 970
|
| 609 |
},
|
| 610 |
{
|
| 611 |
-
"epoch": 2.
|
| 612 |
-
"learning_rate": 1.
|
| 613 |
-
"loss": 0.
|
| 614 |
"step": 980
|
| 615 |
},
|
| 616 |
{
|
| 617 |
-
"epoch": 2.
|
| 618 |
-
"learning_rate": 1.
|
| 619 |
-
"loss": 0.
|
| 620 |
"step": 990
|
| 621 |
},
|
| 622 |
{
|
| 623 |
-
"epoch": 2.
|
| 624 |
-
"learning_rate": 1.
|
| 625 |
-
"loss": 0.
|
| 626 |
"step": 1000
|
| 627 |
},
|
| 628 |
{
|
| 629 |
-
"epoch": 2.
|
| 630 |
-
"
|
| 631 |
-
"
|
| 632 |
-
"eval_runtime": 23.049,
|
| 633 |
-
"eval_samples_per_second": 3.211,
|
| 634 |
-
"eval_steps_per_second": 3.211,
|
| 635 |
-
"step": 1005
|
| 636 |
-
},
|
| 637 |
-
{
|
| 638 |
-
"epoch": 3.0,
|
| 639 |
-
"learning_rate": 1.3681592039800995e-05,
|
| 640 |
-
"loss": 0.1508,
|
| 641 |
"step": 1010
|
| 642 |
},
|
| 643 |
{
|
| 644 |
-
"epoch":
|
| 645 |
-
"learning_rate": 1.
|
| 646 |
-
"loss": 0.
|
| 647 |
"step": 1020
|
| 648 |
},
|
| 649 |
{
|
| 650 |
-
"epoch":
|
| 651 |
-
"learning_rate": 1.
|
| 652 |
-
"loss": 0.
|
| 653 |
"step": 1030
|
| 654 |
},
|
| 655 |
{
|
| 656 |
-
"epoch":
|
| 657 |
-
"learning_rate": 1.
|
| 658 |
-
"loss": 0.
|
| 659 |
"step": 1040
|
| 660 |
},
|
| 661 |
{
|
| 662 |
-
"epoch":
|
| 663 |
-
"learning_rate": 1.
|
| 664 |
-
"loss": 0.
|
| 665 |
"step": 1050
|
| 666 |
},
|
| 667 |
{
|
| 668 |
-
"epoch":
|
| 669 |
-
"learning_rate": 1.
|
| 670 |
-
"loss":
|
| 671 |
"step": 1060
|
| 672 |
},
|
| 673 |
{
|
| 674 |
-
"epoch":
|
| 675 |
-
"learning_rate": 1.
|
| 676 |
-
"loss": 0.
|
| 677 |
"step": 1070
|
| 678 |
},
|
| 679 |
{
|
| 680 |
-
"epoch":
|
| 681 |
-
"learning_rate": 1.
|
| 682 |
-
"loss":
|
| 683 |
"step": 1080
|
| 684 |
},
|
| 685 |
{
|
| 686 |
-
"epoch":
|
| 687 |
-
"learning_rate": 1.
|
| 688 |
-
"loss": 0.
|
| 689 |
"step": 1090
|
| 690 |
},
|
| 691 |
{
|
| 692 |
-
"epoch":
|
| 693 |
-
"
|
| 694 |
-
"
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
| 695 |
"step": 1100
|
| 696 |
},
|
| 697 |
{
|
| 698 |
-
"epoch": 3.
|
| 699 |
-
"learning_rate":
|
| 700 |
-
"loss": 0.
|
| 701 |
"step": 1110
|
| 702 |
},
|
| 703 |
{
|
| 704 |
-
"epoch": 3.
|
| 705 |
-
"learning_rate":
|
| 706 |
-
"loss": 0.
|
| 707 |
"step": 1120
|
| 708 |
},
|
| 709 |
{
|
| 710 |
-
"epoch": 3.
|
| 711 |
-
"learning_rate":
|
| 712 |
-
"loss": 0.
|
| 713 |
"step": 1130
|
| 714 |
},
|
| 715 |
{
|
| 716 |
-
"epoch": 3.
|
| 717 |
-
"learning_rate":
|
| 718 |
-
"loss": 0.
|
| 719 |
"step": 1140
|
| 720 |
},
|
| 721 |
{
|
| 722 |
-
"epoch": 3.
|
| 723 |
-
"learning_rate":
|
| 724 |
-
"loss": 0.
|
| 725 |
"step": 1150
|
| 726 |
},
|
| 727 |
{
|
| 728 |
-
"epoch": 3.
|
| 729 |
-
"learning_rate":
|
| 730 |
-
"loss": 0.
|
| 731 |
"step": 1160
|
| 732 |
},
|
| 733 |
{
|
| 734 |
-
"epoch": 3.
|
| 735 |
-
"learning_rate":
|
| 736 |
-
"loss": 0.
|
| 737 |
"step": 1170
|
| 738 |
},
|
| 739 |
{
|
| 740 |
-
"epoch": 3.
|
| 741 |
-
"learning_rate":
|
| 742 |
-
"loss": 0.
|
| 743 |
"step": 1180
|
| 744 |
},
|
| 745 |
{
|
| 746 |
-
"epoch": 3.
|
| 747 |
-
"learning_rate":
|
| 748 |
-
"loss": 0.
|
| 749 |
"step": 1190
|
| 750 |
},
|
| 751 |
{
|
| 752 |
-
"epoch": 3.
|
| 753 |
-
"learning_rate":
|
| 754 |
-
"loss": 0.
|
| 755 |
"step": 1200
|
| 756 |
},
|
| 757 |
{
|
| 758 |
-
"epoch": 3.
|
| 759 |
-
"learning_rate":
|
| 760 |
-
"loss": 0.
|
| 761 |
"step": 1210
|
| 762 |
},
|
| 763 |
{
|
| 764 |
-
"epoch": 3.
|
| 765 |
-
"learning_rate":
|
| 766 |
-
"loss": 0.
|
| 767 |
"step": 1220
|
| 768 |
},
|
| 769 |
{
|
| 770 |
-
"epoch": 3.
|
| 771 |
-
"learning_rate":
|
| 772 |
-
"loss": 0.
|
| 773 |
"step": 1230
|
| 774 |
},
|
| 775 |
{
|
| 776 |
-
"epoch": 3.
|
| 777 |
-
"learning_rate":
|
| 778 |
-
"loss":
|
| 779 |
"step": 1240
|
| 780 |
},
|
| 781 |
{
|
| 782 |
-
"epoch": 3.
|
| 783 |
-
"learning_rate":
|
| 784 |
-
"loss": 0.
|
| 785 |
"step": 1250
|
| 786 |
},
|
| 787 |
{
|
| 788 |
-
"epoch": 3.
|
| 789 |
-
"learning_rate":
|
| 790 |
-
"loss": 0.
|
| 791 |
"step": 1260
|
| 792 |
},
|
| 793 |
{
|
| 794 |
-
"epoch": 3.
|
| 795 |
-
"learning_rate":
|
| 796 |
-
"loss":
|
| 797 |
"step": 1270
|
| 798 |
},
|
| 799 |
{
|
| 800 |
-
"epoch": 3.
|
| 801 |
-
"learning_rate":
|
| 802 |
-
"loss": 0.
|
| 803 |
"step": 1280
|
| 804 |
},
|
| 805 |
{
|
| 806 |
-
"epoch": 3.
|
| 807 |
-
"learning_rate":
|
| 808 |
-
"loss": 0.
|
| 809 |
"step": 1290
|
| 810 |
},
|
| 811 |
{
|
| 812 |
-
"epoch": 3.
|
| 813 |
-
"learning_rate":
|
| 814 |
-
"loss": 0.
|
| 815 |
"step": 1300
|
| 816 |
},
|
| 817 |
{
|
| 818 |
-
"epoch": 3.
|
| 819 |
-
"learning_rate":
|
| 820 |
-
"loss": 0.
|
| 821 |
"step": 1310
|
| 822 |
},
|
| 823 |
{
|
| 824 |
-
"epoch": 3.
|
| 825 |
-
"learning_rate":
|
| 826 |
-
"loss": 0.
|
| 827 |
"step": 1320
|
| 828 |
},
|
| 829 |
{
|
| 830 |
-
"epoch": 3.
|
| 831 |
-
"learning_rate": 4.
|
| 832 |
-
"loss": 0.
|
| 833 |
"step": 1330
|
| 834 |
},
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
| 835 |
{
|
| 836 |
"epoch": 3.25,
|
| 837 |
"learning_rate": 0.0,
|
| 838 |
-
"loss": 0.
|
| 839 |
-
"step":
|
| 840 |
},
|
| 841 |
{
|
| 842 |
"epoch": 3.25,
|
| 843 |
-
"eval_accuracy": 0.
|
| 844 |
-
"eval_loss":
|
| 845 |
-
"eval_runtime":
|
| 846 |
-
"eval_samples_per_second": 3.
|
| 847 |
-
"eval_steps_per_second": 3.
|
| 848 |
-
"step":
|
| 849 |
},
|
| 850 |
{
|
| 851 |
"epoch": 3.25,
|
| 852 |
-
"step":
|
| 853 |
-
"total_flos": 1.
|
| 854 |
-
"train_loss": 1.
|
| 855 |
-
"train_runtime":
|
| 856 |
-
"train_samples_per_second": 1.
|
| 857 |
-
"train_steps_per_second": 1.
|
| 858 |
},
|
| 859 |
{
|
| 860 |
"epoch": 3.25,
|
| 861 |
-
"eval_accuracy": 0.
|
| 862 |
-
"eval_loss": 0.
|
| 863 |
-
"eval_runtime":
|
| 864 |
-
"eval_samples_per_second": 3.
|
| 865 |
-
"eval_steps_per_second": 3.
|
| 866 |
-
"step":
|
| 867 |
},
|
| 868 |
{
|
| 869 |
"epoch": 3.25,
|
| 870 |
-
"eval_accuracy": 0.
|
| 871 |
-
"eval_loss": 0.
|
| 872 |
-
"eval_runtime":
|
| 873 |
-
"eval_samples_per_second": 3.
|
| 874 |
-
"eval_steps_per_second": 3.
|
| 875 |
-
"step":
|
| 876 |
}
|
| 877 |
],
|
| 878 |
-
"max_steps":
|
| 879 |
"num_train_epochs": 9223372036854775807,
|
| 880 |
-
"total_flos": 1.
|
| 881 |
"trial_name": null,
|
| 882 |
"trial_params": null
|
| 883 |
}
|
|
|
|
| 1 |
{
|
| 2 |
+
"best_metric": 0.7422680412371134,
|
| 3 |
+
"best_model_checkpoint": "videomae-base-finetuned-ucf101-subset/checkpoint-1460",
|
| 4 |
"epoch": 3.25,
|
| 5 |
+
"global_step": 1460,
|
| 6 |
"is_hyper_param_search": false,
|
| 7 |
"is_local_process_zero": true,
|
| 8 |
"is_world_process_zero": true,
|
| 9 |
"log_history": [
|
| 10 |
{
|
| 11 |
"epoch": 0.01,
|
| 12 |
+
"learning_rate": 3.4246575342465754e-06,
|
| 13 |
+
"loss": 2.5993,
|
| 14 |
"step": 10
|
| 15 |
},
|
| 16 |
{
|
| 17 |
"epoch": 0.01,
|
| 18 |
+
"learning_rate": 6.849315068493151e-06,
|
| 19 |
+
"loss": 2.5853,
|
| 20 |
"step": 20
|
| 21 |
},
|
| 22 |
{
|
| 23 |
"epoch": 0.02,
|
| 24 |
+
"learning_rate": 1.0273972602739726e-05,
|
| 25 |
+
"loss": 2.6647,
|
| 26 |
"step": 30
|
| 27 |
},
|
| 28 |
{
|
| 29 |
"epoch": 0.03,
|
| 30 |
+
"learning_rate": 1.3698630136986302e-05,
|
| 31 |
+
"loss": 2.5315,
|
| 32 |
"step": 40
|
| 33 |
},
|
| 34 |
{
|
| 35 |
+
"epoch": 0.03,
|
| 36 |
+
"learning_rate": 1.7123287671232875e-05,
|
| 37 |
+
"loss": 2.4656,
|
| 38 |
"step": 50
|
| 39 |
},
|
| 40 |
{
|
| 41 |
"epoch": 0.04,
|
| 42 |
+
"learning_rate": 2.0547945205479453e-05,
|
| 43 |
+
"loss": 2.6351,
|
| 44 |
"step": 60
|
| 45 |
},
|
| 46 |
{
|
| 47 |
"epoch": 0.05,
|
| 48 |
+
"learning_rate": 2.3972602739726026e-05,
|
| 49 |
+
"loss": 2.662,
|
| 50 |
"step": 70
|
| 51 |
},
|
| 52 |
{
|
| 53 |
+
"epoch": 0.05,
|
| 54 |
+
"learning_rate": 2.7397260273972603e-05,
|
| 55 |
+
"loss": 2.3767,
|
| 56 |
"step": 80
|
| 57 |
},
|
| 58 |
{
|
| 59 |
+
"epoch": 0.06,
|
| 60 |
+
"learning_rate": 3.082191780821918e-05,
|
| 61 |
+
"loss": 2.3325,
|
| 62 |
"step": 90
|
| 63 |
},
|
| 64 |
{
|
| 65 |
"epoch": 0.07,
|
| 66 |
+
"learning_rate": 3.424657534246575e-05,
|
| 67 |
+
"loss": 2.7198,
|
| 68 |
"step": 100
|
| 69 |
},
|
| 70 |
{
|
| 71 |
"epoch": 0.08,
|
| 72 |
+
"learning_rate": 3.767123287671233e-05,
|
| 73 |
+
"loss": 2.4336,
|
| 74 |
"step": 110
|
| 75 |
},
|
| 76 |
{
|
| 77 |
+
"epoch": 0.08,
|
| 78 |
+
"learning_rate": 4.1095890410958905e-05,
|
| 79 |
+
"loss": 2.4116,
|
| 80 |
"step": 120
|
| 81 |
},
|
| 82 |
{
|
| 83 |
+
"epoch": 0.09,
|
| 84 |
+
"learning_rate": 4.452054794520548e-05,
|
| 85 |
+
"loss": 2.6112,
|
| 86 |
"step": 130
|
| 87 |
},
|
| 88 |
{
|
| 89 |
"epoch": 0.1,
|
| 90 |
+
"learning_rate": 4.794520547945205e-05,
|
| 91 |
+
"loss": 2.3558,
|
| 92 |
"step": 140
|
| 93 |
},
|
| 94 |
{
|
| 95 |
+
"epoch": 0.1,
|
| 96 |
+
"learning_rate": 4.984779299847793e-05,
|
| 97 |
+
"loss": 2.1907,
|
| 98 |
"step": 150
|
| 99 |
},
|
| 100 |
{
|
| 101 |
+
"epoch": 0.11,
|
| 102 |
+
"learning_rate": 4.9467275494672755e-05,
|
| 103 |
+
"loss": 2.1687,
|
| 104 |
"step": 160
|
| 105 |
},
|
| 106 |
{
|
| 107 |
+
"epoch": 0.12,
|
| 108 |
+
"learning_rate": 4.908675799086758e-05,
|
| 109 |
+
"loss": 2.2809,
|
| 110 |
"step": 170
|
| 111 |
},
|
| 112 |
{
|
| 113 |
+
"epoch": 0.12,
|
| 114 |
+
"learning_rate": 4.8706240487062404e-05,
|
| 115 |
+
"loss": 2.4247,
|
| 116 |
"step": 180
|
| 117 |
},
|
| 118 |
{
|
| 119 |
+
"epoch": 0.13,
|
| 120 |
+
"learning_rate": 4.8325722983257235e-05,
|
| 121 |
+
"loss": 2.2747,
|
| 122 |
"step": 190
|
| 123 |
},
|
| 124 |
{
|
| 125 |
+
"epoch": 0.14,
|
| 126 |
+
"learning_rate": 4.794520547945205e-05,
|
| 127 |
+
"loss": 2.2174,
|
| 128 |
"step": 200
|
| 129 |
},
|
| 130 |
{
|
| 131 |
+
"epoch": 0.14,
|
| 132 |
+
"learning_rate": 4.7564687975646884e-05,
|
| 133 |
+
"loss": 2.4167,
|
| 134 |
"step": 210
|
| 135 |
},
|
| 136 |
{
|
| 137 |
+
"epoch": 0.15,
|
| 138 |
+
"learning_rate": 4.71841704718417e-05,
|
| 139 |
+
"loss": 2.2652,
|
| 140 |
"step": 220
|
| 141 |
},
|
| 142 |
{
|
| 143 |
+
"epoch": 0.16,
|
| 144 |
+
"learning_rate": 4.680365296803653e-05,
|
| 145 |
+
"loss": 2.2426,
|
| 146 |
"step": 230
|
| 147 |
},
|
| 148 |
{
|
| 149 |
+
"epoch": 0.16,
|
| 150 |
+
"learning_rate": 4.642313546423136e-05,
|
| 151 |
+
"loss": 2.3514,
|
| 152 |
"step": 240
|
| 153 |
},
|
| 154 |
{
|
| 155 |
+
"epoch": 0.17,
|
| 156 |
+
"learning_rate": 4.604261796042618e-05,
|
| 157 |
+
"loss": 1.8138,
|
| 158 |
"step": 250
|
| 159 |
},
|
| 160 |
{
|
| 161 |
+
"epoch": 0.18,
|
| 162 |
+
"learning_rate": 4.5662100456621006e-05,
|
| 163 |
+
"loss": 2.6617,
|
| 164 |
"step": 260
|
| 165 |
},
|
| 166 |
{
|
| 167 |
+
"epoch": 0.18,
|
| 168 |
+
"learning_rate": 4.528158295281583e-05,
|
| 169 |
+
"loss": 1.6823,
|
| 170 |
"step": 270
|
| 171 |
},
|
| 172 |
{
|
| 173 |
+
"epoch": 0.19,
|
| 174 |
+
"learning_rate": 4.4901065449010655e-05,
|
| 175 |
+
"loss": 2.382,
|
| 176 |
"step": 280
|
| 177 |
},
|
| 178 |
{
|
| 179 |
+
"epoch": 0.2,
|
| 180 |
+
"learning_rate": 4.452054794520548e-05,
|
| 181 |
+
"loss": 2.1086,
|
| 182 |
"step": 290
|
| 183 |
},
|
| 184 |
{
|
| 185 |
+
"epoch": 0.21,
|
| 186 |
+
"learning_rate": 4.41400304414003e-05,
|
| 187 |
+
"loss": 1.8853,
|
| 188 |
"step": 300
|
| 189 |
},
|
| 190 |
{
|
| 191 |
+
"epoch": 0.21,
|
| 192 |
+
"learning_rate": 4.3759512937595135e-05,
|
| 193 |
+
"loss": 1.8514,
|
| 194 |
"step": 310
|
| 195 |
},
|
| 196 |
{
|
| 197 |
+
"epoch": 0.22,
|
| 198 |
+
"learning_rate": 4.337899543378995e-05,
|
| 199 |
+
"loss": 2.6215,
|
| 200 |
"step": 320
|
| 201 |
},
|
| 202 |
{
|
| 203 |
+
"epoch": 0.23,
|
| 204 |
+
"learning_rate": 4.299847792998478e-05,
|
| 205 |
+
"loss": 2.2046,
|
| 206 |
"step": 330
|
| 207 |
},
|
| 208 |
{
|
| 209 |
+
"epoch": 0.23,
|
| 210 |
+
"learning_rate": 4.26179604261796e-05,
|
| 211 |
+
"loss": 2.602,
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
| 212 |
"step": 340
|
| 213 |
},
|
| 214 |
{
|
| 215 |
+
"epoch": 0.24,
|
| 216 |
+
"learning_rate": 4.223744292237443e-05,
|
| 217 |
+
"loss": 2.3634,
|
| 218 |
"step": 350
|
| 219 |
},
|
| 220 |
{
|
| 221 |
+
"epoch": 0.25,
|
| 222 |
+
"learning_rate": 4.1856925418569256e-05,
|
| 223 |
+
"loss": 2.0135,
|
| 224 |
"step": 360
|
| 225 |
},
|
| 226 |
{
|
| 227 |
+
"epoch": 0.25,
|
| 228 |
+
"eval_accuracy": 0.14432989690721648,
|
| 229 |
+
"eval_loss": 2.1087806224823,
|
| 230 |
+
"eval_runtime": 27.8992,
|
| 231 |
+
"eval_samples_per_second": 3.477,
|
| 232 |
+
"eval_steps_per_second": 3.477,
|
| 233 |
+
"step": 365
|
| 234 |
+
},
|
| 235 |
+
{
|
| 236 |
+
"epoch": 1.0,
|
| 237 |
+
"learning_rate": 4.147640791476408e-05,
|
| 238 |
+
"loss": 2.381,
|
| 239 |
"step": 370
|
| 240 |
},
|
| 241 |
{
|
| 242 |
+
"epoch": 1.01,
|
| 243 |
+
"learning_rate": 4.1095890410958905e-05,
|
| 244 |
+
"loss": 1.991,
|
| 245 |
"step": 380
|
| 246 |
},
|
| 247 |
{
|
| 248 |
+
"epoch": 1.02,
|
| 249 |
+
"learning_rate": 4.071537290715373e-05,
|
| 250 |
+
"loss": 1.4985,
|
| 251 |
"step": 390
|
| 252 |
},
|
| 253 |
{
|
| 254 |
+
"epoch": 1.02,
|
| 255 |
+
"learning_rate": 4.0334855403348554e-05,
|
| 256 |
+
"loss": 1.761,
|
| 257 |
"step": 400
|
| 258 |
},
|
| 259 |
{
|
| 260 |
+
"epoch": 1.03,
|
| 261 |
+
"learning_rate": 3.995433789954338e-05,
|
| 262 |
+
"loss": 2.366,
|
| 263 |
"step": 410
|
| 264 |
},
|
| 265 |
{
|
| 266 |
+
"epoch": 1.04,
|
| 267 |
+
"learning_rate": 3.95738203957382e-05,
|
| 268 |
+
"loss": 2.3994,
|
| 269 |
"step": 420
|
| 270 |
},
|
| 271 |
{
|
| 272 |
+
"epoch": 1.04,
|
| 273 |
+
"learning_rate": 3.9193302891933034e-05,
|
| 274 |
+
"loss": 2.032,
|
| 275 |
"step": 430
|
| 276 |
},
|
| 277 |
{
|
| 278 |
+
"epoch": 1.05,
|
| 279 |
+
"learning_rate": 3.881278538812785e-05,
|
| 280 |
+
"loss": 1.9747,
|
| 281 |
"step": 440
|
| 282 |
},
|
| 283 |
{
|
| 284 |
+
"epoch": 1.06,
|
| 285 |
+
"learning_rate": 3.843226788432268e-05,
|
| 286 |
+
"loss": 1.925,
|
| 287 |
"step": 450
|
| 288 |
},
|
| 289 |
{
|
| 290 |
+
"epoch": 1.07,
|
| 291 |
+
"learning_rate": 3.80517503805175e-05,
|
| 292 |
+
"loss": 1.7874,
|
| 293 |
"step": 460
|
| 294 |
},
|
| 295 |
{
|
| 296 |
+
"epoch": 1.07,
|
| 297 |
+
"learning_rate": 3.767123287671233e-05,
|
| 298 |
+
"loss": 1.9562,
|
| 299 |
"step": 470
|
| 300 |
},
|
| 301 |
{
|
| 302 |
+
"epoch": 1.08,
|
| 303 |
+
"learning_rate": 3.7290715372907156e-05,
|
| 304 |
+
"loss": 2.988,
|
| 305 |
"step": 480
|
| 306 |
},
|
| 307 |
{
|
| 308 |
+
"epoch": 1.09,
|
| 309 |
+
"learning_rate": 3.691019786910198e-05,
|
| 310 |
+
"loss": 1.6423,
|
| 311 |
"step": 490
|
| 312 |
},
|
| 313 |
{
|
| 314 |
+
"epoch": 1.09,
|
| 315 |
+
"learning_rate": 3.6529680365296805e-05,
|
| 316 |
+
"loss": 2.0119,
|
| 317 |
"step": 500
|
| 318 |
},
|
| 319 |
{
|
| 320 |
+
"epoch": 1.1,
|
| 321 |
+
"learning_rate": 3.614916286149163e-05,
|
| 322 |
+
"loss": 1.6883,
|
| 323 |
"step": 510
|
| 324 |
},
|
| 325 |
{
|
| 326 |
+
"epoch": 1.11,
|
| 327 |
+
"learning_rate": 3.5768645357686453e-05,
|
| 328 |
+
"loss": 1.7275,
|
| 329 |
"step": 520
|
| 330 |
},
|
| 331 |
{
|
| 332 |
+
"epoch": 1.11,
|
| 333 |
+
"learning_rate": 3.538812785388128e-05,
|
| 334 |
+
"loss": 1.0309,
|
| 335 |
"step": 530
|
| 336 |
},
|
| 337 |
{
|
| 338 |
+
"epoch": 1.12,
|
| 339 |
+
"learning_rate": 3.50076103500761e-05,
|
| 340 |
+
"loss": 2.0449,
|
| 341 |
"step": 540
|
| 342 |
},
|
| 343 |
{
|
| 344 |
+
"epoch": 1.13,
|
| 345 |
+
"learning_rate": 3.462709284627093e-05,
|
| 346 |
+
"loss": 1.1414,
|
| 347 |
"step": 550
|
| 348 |
},
|
| 349 |
{
|
| 350 |
+
"epoch": 1.13,
|
| 351 |
+
"learning_rate": 3.424657534246575e-05,
|
| 352 |
+
"loss": 1.0559,
|
| 353 |
"step": 560
|
| 354 |
},
|
| 355 |
{
|
| 356 |
+
"epoch": 1.14,
|
| 357 |
+
"learning_rate": 3.386605783866058e-05,
|
| 358 |
+
"loss": 1.4151,
|
| 359 |
"step": 570
|
| 360 |
},
|
| 361 |
{
|
| 362 |
+
"epoch": 1.15,
|
| 363 |
+
"learning_rate": 3.34855403348554e-05,
|
| 364 |
+
"loss": 1.5515,
|
| 365 |
"step": 580
|
| 366 |
},
|
| 367 |
{
|
| 368 |
+
"epoch": 1.15,
|
| 369 |
+
"learning_rate": 3.310502283105023e-05,
|
| 370 |
+
"loss": 1.2185,
|
| 371 |
"step": 590
|
| 372 |
},
|
| 373 |
{
|
| 374 |
+
"epoch": 1.16,
|
| 375 |
+
"learning_rate": 3.2724505327245055e-05,
|
| 376 |
+
"loss": 1.3974,
|
| 377 |
"step": 600
|
| 378 |
},
|
| 379 |
{
|
| 380 |
+
"epoch": 1.17,
|
| 381 |
+
"learning_rate": 3.234398782343988e-05,
|
| 382 |
+
"loss": 1.3453,
|
| 383 |
"step": 610
|
| 384 |
},
|
| 385 |
{
|
| 386 |
+
"epoch": 1.17,
|
| 387 |
+
"learning_rate": 3.1963470319634704e-05,
|
| 388 |
+
"loss": 1.8057,
|
| 389 |
"step": 620
|
| 390 |
},
|
| 391 |
{
|
| 392 |
+
"epoch": 1.18,
|
| 393 |
+
"learning_rate": 3.158295281582953e-05,
|
| 394 |
+
"loss": 0.9315,
|
| 395 |
"step": 630
|
| 396 |
},
|
| 397 |
{
|
| 398 |
+
"epoch": 1.19,
|
| 399 |
+
"learning_rate": 3.120243531202435e-05,
|
| 400 |
+
"loss": 1.2106,
|
| 401 |
"step": 640
|
| 402 |
},
|
| 403 |
{
|
| 404 |
+
"epoch": 1.2,
|
| 405 |
+
"learning_rate": 3.082191780821918e-05,
|
| 406 |
+
"loss": 0.7568,
|
| 407 |
"step": 650
|
| 408 |
},
|
| 409 |
{
|
| 410 |
+
"epoch": 1.2,
|
| 411 |
+
"learning_rate": 3.0441400304414e-05,
|
| 412 |
+
"loss": 1.0881,
|
| 413 |
"step": 660
|
| 414 |
},
|
| 415 |
{
|
| 416 |
+
"epoch": 1.21,
|
| 417 |
+
"learning_rate": 3.006088280060883e-05,
|
| 418 |
+
"loss": 1.1191,
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
| 419 |
"step": 670
|
| 420 |
},
|
| 421 |
{
|
| 422 |
+
"epoch": 1.22,
|
| 423 |
+
"learning_rate": 2.9680365296803654e-05,
|
| 424 |
+
"loss": 0.3567,
|
| 425 |
"step": 680
|
| 426 |
},
|
| 427 |
{
|
| 428 |
+
"epoch": 1.22,
|
| 429 |
+
"learning_rate": 2.929984779299848e-05,
|
| 430 |
+
"loss": 0.828,
|
| 431 |
"step": 690
|
| 432 |
},
|
| 433 |
{
|
| 434 |
+
"epoch": 1.23,
|
| 435 |
+
"learning_rate": 2.8919330289193303e-05,
|
| 436 |
+
"loss": 0.6539,
|
| 437 |
"step": 700
|
| 438 |
},
|
| 439 |
{
|
| 440 |
+
"epoch": 1.24,
|
| 441 |
+
"learning_rate": 2.853881278538813e-05,
|
| 442 |
+
"loss": 1.2757,
|
| 443 |
"step": 710
|
| 444 |
},
|
| 445 |
{
|
| 446 |
+
"epoch": 1.24,
|
| 447 |
+
"learning_rate": 2.815829528158295e-05,
|
| 448 |
+
"loss": 1.9999,
|
| 449 |
"step": 720
|
| 450 |
},
|
| 451 |
{
|
| 452 |
+
"epoch": 1.25,
|
| 453 |
+
"learning_rate": 2.777777777777778e-05,
|
| 454 |
+
"loss": 1.6817,
|
| 455 |
"step": 730
|
| 456 |
},
|
| 457 |
{
|
| 458 |
+
"epoch": 1.25,
|
| 459 |
+
"eval_accuracy": 0.3917525773195876,
|
| 460 |
+
"eval_loss": 2.0997157096862793,
|
| 461 |
+
"eval_runtime": 25.2952,
|
| 462 |
+
"eval_samples_per_second": 3.835,
|
| 463 |
+
"eval_steps_per_second": 3.835,
|
| 464 |
+
"step": 730
|
| 465 |
+
},
|
| 466 |
+
{
|
| 467 |
+
"epoch": 2.01,
|
| 468 |
+
"learning_rate": 2.7397260273972603e-05,
|
| 469 |
+
"loss": 0.9235,
|
| 470 |
"step": 740
|
| 471 |
},
|
| 472 |
{
|
| 473 |
+
"epoch": 2.01,
|
| 474 |
+
"learning_rate": 2.701674277016743e-05,
|
| 475 |
+
"loss": 1.7029,
|
| 476 |
"step": 750
|
| 477 |
},
|
| 478 |
{
|
| 479 |
+
"epoch": 2.02,
|
| 480 |
+
"learning_rate": 2.6636225266362252e-05,
|
| 481 |
+
"loss": 1.0662,
|
| 482 |
"step": 760
|
| 483 |
},
|
| 484 |
{
|
| 485 |
+
"epoch": 2.03,
|
| 486 |
+
"learning_rate": 2.625570776255708e-05,
|
| 487 |
+
"loss": 0.9232,
|
| 488 |
"step": 770
|
| 489 |
},
|
| 490 |
{
|
| 491 |
+
"epoch": 2.03,
|
| 492 |
+
"learning_rate": 2.58751902587519e-05,
|
| 493 |
+
"loss": 0.9683,
|
| 494 |
"step": 780
|
| 495 |
},
|
| 496 |
{
|
| 497 |
+
"epoch": 2.04,
|
| 498 |
+
"learning_rate": 2.549467275494673e-05,
|
| 499 |
+
"loss": 1.005,
|
| 500 |
"step": 790
|
| 501 |
},
|
| 502 |
{
|
| 503 |
+
"epoch": 2.05,
|
| 504 |
+
"learning_rate": 2.5114155251141553e-05,
|
| 505 |
+
"loss": 1.0024,
|
| 506 |
"step": 800
|
| 507 |
},
|
| 508 |
{
|
| 509 |
+
"epoch": 2.05,
|
| 510 |
+
"learning_rate": 2.4733637747336378e-05,
|
| 511 |
+
"loss": 1.0185,
|
| 512 |
"step": 810
|
| 513 |
},
|
| 514 |
{
|
| 515 |
+
"epoch": 2.06,
|
| 516 |
+
"learning_rate": 2.4353120243531202e-05,
|
| 517 |
+
"loss": 0.289,
|
| 518 |
"step": 820
|
| 519 |
},
|
| 520 |
{
|
| 521 |
+
"epoch": 2.07,
|
| 522 |
+
"learning_rate": 2.3972602739726026e-05,
|
| 523 |
+
"loss": 1.3332,
|
| 524 |
"step": 830
|
| 525 |
},
|
| 526 |
{
|
| 527 |
+
"epoch": 2.08,
|
| 528 |
+
"learning_rate": 2.359208523592085e-05,
|
| 529 |
+
"loss": 1.2137,
|
| 530 |
"step": 840
|
| 531 |
},
|
| 532 |
{
|
| 533 |
+
"epoch": 2.08,
|
| 534 |
+
"learning_rate": 2.321156773211568e-05,
|
| 535 |
+
"loss": 0.8686,
|
| 536 |
"step": 850
|
| 537 |
},
|
| 538 |
{
|
| 539 |
+
"epoch": 2.09,
|
| 540 |
+
"learning_rate": 2.2831050228310503e-05,
|
| 541 |
+
"loss": 1.8639,
|
| 542 |
"step": 860
|
| 543 |
},
|
| 544 |
{
|
| 545 |
+
"epoch": 2.1,
|
| 546 |
+
"learning_rate": 2.2450532724505327e-05,
|
| 547 |
+
"loss": 0.7075,
|
| 548 |
"step": 870
|
| 549 |
},
|
| 550 |
{
|
| 551 |
+
"epoch": 2.1,
|
| 552 |
+
"learning_rate": 2.207001522070015e-05,
|
| 553 |
+
"loss": 0.4791,
|
| 554 |
"step": 880
|
| 555 |
},
|
| 556 |
{
|
| 557 |
+
"epoch": 2.11,
|
| 558 |
+
"learning_rate": 2.1689497716894976e-05,
|
| 559 |
+
"loss": 0.8184,
|
| 560 |
"step": 890
|
| 561 |
},
|
| 562 |
{
|
| 563 |
+
"epoch": 2.12,
|
| 564 |
+
"learning_rate": 2.13089802130898e-05,
|
| 565 |
+
"loss": 1.0992,
|
| 566 |
"step": 900
|
| 567 |
},
|
| 568 |
{
|
| 569 |
+
"epoch": 2.12,
|
| 570 |
+
"learning_rate": 2.0928462709284628e-05,
|
| 571 |
+
"loss": 0.7498,
|
| 572 |
"step": 910
|
| 573 |
},
|
| 574 |
{
|
| 575 |
+
"epoch": 2.13,
|
| 576 |
+
"learning_rate": 2.0547945205479453e-05,
|
| 577 |
+
"loss": 0.2501,
|
| 578 |
"step": 920
|
| 579 |
},
|
| 580 |
{
|
| 581 |
+
"epoch": 2.14,
|
| 582 |
+
"learning_rate": 2.0167427701674277e-05,
|
| 583 |
+
"loss": 0.2121,
|
| 584 |
"step": 930
|
| 585 |
},
|
| 586 |
{
|
| 587 |
+
"epoch": 2.14,
|
| 588 |
+
"learning_rate": 1.97869101978691e-05,
|
| 589 |
+
"loss": 0.3343,
|
| 590 |
"step": 940
|
| 591 |
},
|
| 592 |
{
|
| 593 |
+
"epoch": 2.15,
|
| 594 |
+
"learning_rate": 1.9406392694063926e-05,
|
| 595 |
+
"loss": 0.7544,
|
| 596 |
"step": 950
|
| 597 |
},
|
| 598 |
{
|
| 599 |
+
"epoch": 2.16,
|
| 600 |
+
"learning_rate": 1.902587519025875e-05,
|
| 601 |
+
"loss": 1.3621,
|
| 602 |
"step": 960
|
| 603 |
},
|
| 604 |
{
|
| 605 |
+
"epoch": 2.16,
|
| 606 |
+
"learning_rate": 1.8645357686453578e-05,
|
| 607 |
+
"loss": 0.5267,
|
| 608 |
"step": 970
|
| 609 |
},
|
| 610 |
{
|
| 611 |
+
"epoch": 2.17,
|
| 612 |
+
"learning_rate": 1.8264840182648402e-05,
|
| 613 |
+
"loss": 0.8902,
|
| 614 |
"step": 980
|
| 615 |
},
|
| 616 |
{
|
| 617 |
+
"epoch": 2.18,
|
| 618 |
+
"learning_rate": 1.7884322678843227e-05,
|
| 619 |
+
"loss": 0.5535,
|
| 620 |
"step": 990
|
| 621 |
},
|
| 622 |
{
|
| 623 |
+
"epoch": 2.18,
|
| 624 |
+
"learning_rate": 1.750380517503805e-05,
|
| 625 |
+
"loss": 0.8907,
|
| 626 |
"step": 1000
|
| 627 |
},
|
| 628 |
{
|
| 629 |
+
"epoch": 2.19,
|
| 630 |
+
"learning_rate": 1.7123287671232875e-05,
|
| 631 |
+
"loss": 0.1521,
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
| 632 |
"step": 1010
|
| 633 |
},
|
| 634 |
{
|
| 635 |
+
"epoch": 2.2,
|
| 636 |
+
"learning_rate": 1.67427701674277e-05,
|
| 637 |
+
"loss": 0.2744,
|
| 638 |
"step": 1020
|
| 639 |
},
|
| 640 |
{
|
| 641 |
+
"epoch": 2.21,
|
| 642 |
+
"learning_rate": 1.6362252663622528e-05,
|
| 643 |
+
"loss": 0.17,
|
| 644 |
"step": 1030
|
| 645 |
},
|
| 646 |
{
|
| 647 |
+
"epoch": 2.21,
|
| 648 |
+
"learning_rate": 1.5981735159817352e-05,
|
| 649 |
+
"loss": 0.1956,
|
| 650 |
"step": 1040
|
| 651 |
},
|
| 652 |
{
|
| 653 |
+
"epoch": 2.22,
|
| 654 |
+
"learning_rate": 1.5601217656012176e-05,
|
| 655 |
+
"loss": 0.9093,
|
| 656 |
"step": 1050
|
| 657 |
},
|
| 658 |
{
|
| 659 |
+
"epoch": 2.23,
|
| 660 |
+
"learning_rate": 1.5220700152207e-05,
|
| 661 |
+
"loss": 1.2878,
|
| 662 |
"step": 1060
|
| 663 |
},
|
| 664 |
{
|
| 665 |
+
"epoch": 2.23,
|
| 666 |
+
"learning_rate": 1.4840182648401827e-05,
|
| 667 |
+
"loss": 0.6323,
|
| 668 |
"step": 1070
|
| 669 |
},
|
| 670 |
{
|
| 671 |
+
"epoch": 2.24,
|
| 672 |
+
"learning_rate": 1.4459665144596651e-05,
|
| 673 |
+
"loss": 0.4712,
|
| 674 |
"step": 1080
|
| 675 |
},
|
| 676 |
{
|
| 677 |
+
"epoch": 2.25,
|
| 678 |
+
"learning_rate": 1.4079147640791476e-05,
|
| 679 |
+
"loss": 0.1918,
|
| 680 |
"step": 1090
|
| 681 |
},
|
| 682 |
{
|
| 683 |
+
"epoch": 2.25,
|
| 684 |
+
"eval_accuracy": 0.6185567010309279,
|
| 685 |
+
"eval_loss": 1.2655441761016846,
|
| 686 |
+
"eval_runtime": 25.5435,
|
| 687 |
+
"eval_samples_per_second": 3.797,
|
| 688 |
+
"eval_steps_per_second": 3.797,
|
| 689 |
+
"step": 1095
|
| 690 |
+
},
|
| 691 |
+
{
|
| 692 |
+
"epoch": 3.0,
|
| 693 |
+
"learning_rate": 1.3698630136986302e-05,
|
| 694 |
+
"loss": 0.629,
|
| 695 |
"step": 1100
|
| 696 |
},
|
| 697 |
{
|
| 698 |
+
"epoch": 3.01,
|
| 699 |
+
"learning_rate": 1.3318112633181126e-05,
|
| 700 |
+
"loss": 0.4214,
|
| 701 |
"step": 1110
|
| 702 |
},
|
| 703 |
{
|
| 704 |
+
"epoch": 3.02,
|
| 705 |
+
"learning_rate": 1.293759512937595e-05,
|
| 706 |
+
"loss": 0.7807,
|
| 707 |
"step": 1120
|
| 708 |
},
|
| 709 |
{
|
| 710 |
+
"epoch": 3.02,
|
| 711 |
+
"learning_rate": 1.2557077625570777e-05,
|
| 712 |
+
"loss": 0.2658,
|
| 713 |
"step": 1130
|
| 714 |
},
|
| 715 |
{
|
| 716 |
+
"epoch": 3.03,
|
| 717 |
+
"learning_rate": 1.2176560121765601e-05,
|
| 718 |
+
"loss": 0.8271,
|
| 719 |
"step": 1140
|
| 720 |
},
|
| 721 |
{
|
| 722 |
+
"epoch": 3.04,
|
| 723 |
+
"learning_rate": 1.1796042617960425e-05,
|
| 724 |
+
"loss": 0.3046,
|
| 725 |
"step": 1150
|
| 726 |
},
|
| 727 |
{
|
| 728 |
+
"epoch": 3.04,
|
| 729 |
+
"learning_rate": 1.1415525114155251e-05,
|
| 730 |
+
"loss": 0.3587,
|
| 731 |
"step": 1160
|
| 732 |
},
|
| 733 |
{
|
| 734 |
+
"epoch": 3.05,
|
| 735 |
+
"learning_rate": 1.1035007610350076e-05,
|
| 736 |
+
"loss": 0.1176,
|
| 737 |
"step": 1170
|
| 738 |
},
|
| 739 |
{
|
| 740 |
+
"epoch": 3.06,
|
| 741 |
+
"learning_rate": 1.06544901065449e-05,
|
| 742 |
+
"loss": 0.1104,
|
| 743 |
"step": 1180
|
| 744 |
},
|
| 745 |
{
|
| 746 |
+
"epoch": 3.07,
|
| 747 |
+
"learning_rate": 1.0273972602739726e-05,
|
| 748 |
+
"loss": 0.8839,
|
| 749 |
"step": 1190
|
| 750 |
},
|
| 751 |
{
|
| 752 |
+
"epoch": 3.07,
|
| 753 |
+
"learning_rate": 9.89345509893455e-06,
|
| 754 |
+
"loss": 0.1449,
|
| 755 |
"step": 1200
|
| 756 |
},
|
| 757 |
{
|
| 758 |
+
"epoch": 3.08,
|
| 759 |
+
"learning_rate": 9.512937595129375e-06,
|
| 760 |
+
"loss": 0.1052,
|
| 761 |
"step": 1210
|
| 762 |
},
|
| 763 |
{
|
| 764 |
+
"epoch": 3.09,
|
| 765 |
+
"learning_rate": 9.132420091324201e-06,
|
| 766 |
+
"loss": 0.0785,
|
| 767 |
"step": 1220
|
| 768 |
},
|
| 769 |
{
|
| 770 |
+
"epoch": 3.09,
|
| 771 |
+
"learning_rate": 8.751902587519026e-06,
|
| 772 |
+
"loss": 0.3151,
|
| 773 |
"step": 1230
|
| 774 |
},
|
| 775 |
{
|
| 776 |
+
"epoch": 3.1,
|
| 777 |
+
"learning_rate": 8.37138508371385e-06,
|
| 778 |
+
"loss": 1.218,
|
| 779 |
"step": 1240
|
| 780 |
},
|
| 781 |
{
|
| 782 |
+
"epoch": 3.11,
|
| 783 |
+
"learning_rate": 7.990867579908676e-06,
|
| 784 |
+
"loss": 0.1809,
|
| 785 |
"step": 1250
|
| 786 |
},
|
| 787 |
{
|
| 788 |
+
"epoch": 3.11,
|
| 789 |
+
"learning_rate": 7.6103500761035e-06,
|
| 790 |
+
"loss": 0.1806,
|
| 791 |
"step": 1260
|
| 792 |
},
|
| 793 |
{
|
| 794 |
+
"epoch": 3.12,
|
| 795 |
+
"learning_rate": 7.229832572298326e-06,
|
| 796 |
+
"loss": 0.1936,
|
| 797 |
"step": 1270
|
| 798 |
},
|
| 799 |
{
|
| 800 |
+
"epoch": 3.13,
|
| 801 |
+
"learning_rate": 6.849315068493151e-06,
|
| 802 |
+
"loss": 0.0504,
|
| 803 |
"step": 1280
|
| 804 |
},
|
| 805 |
{
|
| 806 |
+
"epoch": 3.13,
|
| 807 |
+
"learning_rate": 6.468797564687975e-06,
|
| 808 |
+
"loss": 0.9438,
|
| 809 |
"step": 1290
|
| 810 |
},
|
| 811 |
{
|
| 812 |
+
"epoch": 3.14,
|
| 813 |
+
"learning_rate": 6.0882800608828005e-06,
|
| 814 |
+
"loss": 0.0356,
|
| 815 |
"step": 1300
|
| 816 |
},
|
| 817 |
{
|
| 818 |
+
"epoch": 3.15,
|
| 819 |
+
"learning_rate": 5.707762557077626e-06,
|
| 820 |
+
"loss": 0.9749,
|
| 821 |
"step": 1310
|
| 822 |
},
|
| 823 |
{
|
| 824 |
+
"epoch": 3.15,
|
| 825 |
+
"learning_rate": 5.32724505327245e-06,
|
| 826 |
+
"loss": 0.2376,
|
| 827 |
"step": 1320
|
| 828 |
},
|
| 829 |
{
|
| 830 |
+
"epoch": 3.16,
|
| 831 |
+
"learning_rate": 4.946727549467275e-06,
|
| 832 |
+
"loss": 0.3259,
|
| 833 |
"step": 1330
|
| 834 |
},
|
| 835 |
+
{
|
| 836 |
+
"epoch": 3.17,
|
| 837 |
+
"learning_rate": 4.566210045662101e-06,
|
| 838 |
+
"loss": 0.2959,
|
| 839 |
+
"step": 1340
|
| 840 |
+
},
|
| 841 |
+
{
|
| 842 |
+
"epoch": 3.17,
|
| 843 |
+
"learning_rate": 4.185692541856925e-06,
|
| 844 |
+
"loss": 0.0705,
|
| 845 |
+
"step": 1350
|
| 846 |
+
},
|
| 847 |
+
{
|
| 848 |
+
"epoch": 3.18,
|
| 849 |
+
"learning_rate": 3.80517503805175e-06,
|
| 850 |
+
"loss": 0.5899,
|
| 851 |
+
"step": 1360
|
| 852 |
+
},
|
| 853 |
+
{
|
| 854 |
+
"epoch": 3.19,
|
| 855 |
+
"learning_rate": 3.4246575342465754e-06,
|
| 856 |
+
"loss": 0.2143,
|
| 857 |
+
"step": 1370
|
| 858 |
+
},
|
| 859 |
+
{
|
| 860 |
+
"epoch": 3.2,
|
| 861 |
+
"learning_rate": 3.0441400304414002e-06,
|
| 862 |
+
"loss": 0.265,
|
| 863 |
+
"step": 1380
|
| 864 |
+
},
|
| 865 |
+
{
|
| 866 |
+
"epoch": 3.2,
|
| 867 |
+
"learning_rate": 2.663622526636225e-06,
|
| 868 |
+
"loss": 0.0809,
|
| 869 |
+
"step": 1390
|
| 870 |
+
},
|
| 871 |
+
{
|
| 872 |
+
"epoch": 3.21,
|
| 873 |
+
"learning_rate": 2.2831050228310503e-06,
|
| 874 |
+
"loss": 0.0135,
|
| 875 |
+
"step": 1400
|
| 876 |
+
},
|
| 877 |
+
{
|
| 878 |
+
"epoch": 3.22,
|
| 879 |
+
"learning_rate": 1.902587519025875e-06,
|
| 880 |
+
"loss": 0.0089,
|
| 881 |
+
"step": 1410
|
| 882 |
+
},
|
| 883 |
+
{
|
| 884 |
+
"epoch": 3.22,
|
| 885 |
+
"learning_rate": 1.5220700152207001e-06,
|
| 886 |
+
"loss": 0.0718,
|
| 887 |
+
"step": 1420
|
| 888 |
+
},
|
| 889 |
+
{
|
| 890 |
+
"epoch": 3.23,
|
| 891 |
+
"learning_rate": 1.1415525114155251e-06,
|
| 892 |
+
"loss": 0.5197,
|
| 893 |
+
"step": 1430
|
| 894 |
+
},
|
| 895 |
+
{
|
| 896 |
+
"epoch": 3.24,
|
| 897 |
+
"learning_rate": 7.610350076103501e-07,
|
| 898 |
+
"loss": 0.3288,
|
| 899 |
+
"step": 1440
|
| 900 |
+
},
|
| 901 |
+
{
|
| 902 |
+
"epoch": 3.24,
|
| 903 |
+
"learning_rate": 3.8051750380517503e-07,
|
| 904 |
+
"loss": 0.1716,
|
| 905 |
+
"step": 1450
|
| 906 |
+
},
|
| 907 |
{
|
| 908 |
"epoch": 3.25,
|
| 909 |
"learning_rate": 0.0,
|
| 910 |
+
"loss": 0.4802,
|
| 911 |
+
"step": 1460
|
| 912 |
},
|
| 913 |
{
|
| 914 |
"epoch": 3.25,
|
| 915 |
+
"eval_accuracy": 0.7422680412371134,
|
| 916 |
+
"eval_loss": 1.1125394105911255,
|
| 917 |
+
"eval_runtime": 26.9468,
|
| 918 |
+
"eval_samples_per_second": 3.6,
|
| 919 |
+
"eval_steps_per_second": 3.6,
|
| 920 |
+
"step": 1460
|
| 921 |
},
|
| 922 |
{
|
| 923 |
"epoch": 3.25,
|
| 924 |
+
"step": 1460,
|
| 925 |
+
"total_flos": 1.8194165442831974e+18,
|
| 926 |
+
"train_loss": 1.2557036288133632,
|
| 927 |
+
"train_runtime": 1150.1678,
|
| 928 |
+
"train_samples_per_second": 1.269,
|
| 929 |
+
"train_steps_per_second": 1.269
|
| 930 |
},
|
| 931 |
{
|
| 932 |
"epoch": 3.25,
|
| 933 |
+
"eval_accuracy": 0.8114285714285714,
|
| 934 |
+
"eval_loss": 0.6237284541130066,
|
| 935 |
+
"eval_runtime": 57.0411,
|
| 936 |
+
"eval_samples_per_second": 3.068,
|
| 937 |
+
"eval_steps_per_second": 3.068,
|
| 938 |
+
"step": 1460
|
| 939 |
},
|
| 940 |
{
|
| 941 |
"epoch": 3.25,
|
| 942 |
+
"eval_accuracy": 0.8114285714285714,
|
| 943 |
+
"eval_loss": 0.6237285733222961,
|
| 944 |
+
"eval_runtime": 52.2354,
|
| 945 |
+
"eval_samples_per_second": 3.35,
|
| 946 |
+
"eval_steps_per_second": 3.35,
|
| 947 |
+
"step": 1460
|
| 948 |
}
|
| 949 |
],
|
| 950 |
+
"max_steps": 1460,
|
| 951 |
"num_train_epochs": 9223372036854775807,
|
| 952 |
+
"total_flos": 1.8194165442831974e+18,
|
| 953 |
"trial_name": null,
|
| 954 |
"trial_params": null
|
| 955 |
}
|