Training in progress, epoch 0
Browse files- all_results.json +6 -6
- pytorch_model.bin +1 -1
- runs/Nov06_05-04-24_76172e0bfbf9/events.out.tfevents.1699247079.76172e0bfbf9.578.2 +3 -0
- train_results.json +6 -6
- trainer_state.json +505 -181
- training_args.bin +1 -1
all_results.json
CHANGED
|
@@ -1,13 +1,13 @@
|
|
| 1 |
{
|
| 2 |
-
"epoch":
|
| 3 |
"eval_accuracy": 0.30788520479242126,
|
| 4 |
"eval_loss": 1.6947710514068604,
|
| 5 |
"eval_runtime": 17.149,
|
| 6 |
"eval_samples_per_second": 209.283,
|
| 7 |
"eval_steps_per_second": 6.589,
|
| 8 |
-
"total_flos":
|
| 9 |
-
"train_loss": 1.
|
| 10 |
-
"train_runtime":
|
| 11 |
-
"train_samples_per_second":
|
| 12 |
-
"train_steps_per_second": 0.
|
| 13 |
}
|
|
|
|
| 1 |
{
|
| 2 |
+
"epoch": 4.99,
|
| 3 |
"eval_accuracy": 0.30788520479242126,
|
| 4 |
"eval_loss": 1.6947710514068604,
|
| 5 |
"eval_runtime": 17.149,
|
| 6 |
"eval_samples_per_second": 209.283,
|
| 7 |
"eval_steps_per_second": 6.589,
|
| 8 |
+
"total_flos": 3.424468651849728e+18,
|
| 9 |
+
"train_loss": 1.653014714377267,
|
| 10 |
+
"train_runtime": 2002.0643,
|
| 11 |
+
"train_samples_per_second": 80.662,
|
| 12 |
+
"train_steps_per_second": 0.629
|
| 13 |
}
|
pytorch_model.bin
CHANGED
|
@@ -1,3 +1,3 @@
|
|
| 1 |
version https://git-lfs.github.com/spec/v1
|
| 2 |
-
oid sha256:
|
| 3 |
size 94416394
|
|
|
|
| 1 |
version https://git-lfs.github.com/spec/v1
|
| 2 |
+
oid sha256:eeca89952dba0bd69eb866dcb80f92e7db5657894e4b35a6b2d1f931a10da0b0
|
| 3 |
size 94416394
|
runs/Nov06_05-04-24_76172e0bfbf9/events.out.tfevents.1699247079.76172e0bfbf9.578.2
ADDED
|
@@ -0,0 +1,3 @@
|
|
|
|
|
|
|
|
|
|
|
|
|
| 1 |
+
version https://git-lfs.github.com/spec/v1
|
| 2 |
+
oid sha256:d7ddefc09b51c51f0767e8dd4acc0907204119e7574dde24dfd9a4059141b684
|
| 3 |
+
size 8542
|
train_results.json
CHANGED
|
@@ -1,8 +1,8 @@
|
|
| 1 |
{
|
| 2 |
-
"epoch":
|
| 3 |
-
"total_flos":
|
| 4 |
-
"train_loss": 1.
|
| 5 |
-
"train_runtime":
|
| 6 |
-
"train_samples_per_second":
|
| 7 |
-
"train_steps_per_second": 0.
|
| 8 |
}
|
|
|
|
| 1 |
{
|
| 2 |
+
"epoch": 4.99,
|
| 3 |
+
"total_flos": 3.424468651849728e+18,
|
| 4 |
+
"train_loss": 1.653014714377267,
|
| 5 |
+
"train_runtime": 2002.0643,
|
| 6 |
+
"train_samples_per_second": 80.662,
|
| 7 |
+
"train_steps_per_second": 0.629
|
| 8 |
}
|
trainer_state.json
CHANGED
|
@@ -1,502 +1,826 @@
|
|
| 1 |
{
|
| 2 |
-
"best_metric": 0.
|
| 3 |
-
"best_model_checkpoint": "microsoft/resnet-50/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": 0.04,
|
| 12 |
-
"learning_rate":
|
| 13 |
-
"loss": 1.
|
| 14 |
"step": 10
|
| 15 |
},
|
| 16 |
{
|
| 17 |
"epoch": 0.08,
|
| 18 |
-
"learning_rate":
|
| 19 |
-
"loss": 1.
|
| 20 |
"step": 20
|
| 21 |
},
|
| 22 |
{
|
| 23 |
"epoch": 0.12,
|
| 24 |
-
"learning_rate": 1.
|
| 25 |
-
"loss": 1.
|
| 26 |
"step": 30
|
| 27 |
},
|
| 28 |
{
|
| 29 |
"epoch": 0.16,
|
| 30 |
-
"learning_rate":
|
| 31 |
-
"loss": 1.
|
| 32 |
"step": 40
|
| 33 |
},
|
| 34 |
{
|
| 35 |
"epoch": 0.2,
|
| 36 |
-
"learning_rate":
|
| 37 |
-
"loss": 1.
|
| 38 |
"step": 50
|
| 39 |
},
|
| 40 |
{
|
| 41 |
"epoch": 0.24,
|
| 42 |
-
"learning_rate":
|
| 43 |
-
"loss": 1.
|
| 44 |
"step": 60
|
| 45 |
},
|
| 46 |
{
|
| 47 |
"epoch": 0.28,
|
| 48 |
-
"learning_rate":
|
| 49 |
-
"loss": 1.
|
| 50 |
"step": 70
|
| 51 |
},
|
| 52 |
{
|
| 53 |
"epoch": 0.32,
|
| 54 |
-
"learning_rate":
|
| 55 |
-
"loss": 1.
|
| 56 |
"step": 80
|
| 57 |
},
|
| 58 |
{
|
| 59 |
"epoch": 0.36,
|
| 60 |
-
"learning_rate":
|
| 61 |
-
"loss": 1.
|
| 62 |
"step": 90
|
| 63 |
},
|
| 64 |
{
|
| 65 |
"epoch": 0.4,
|
| 66 |
-
"learning_rate":
|
| 67 |
-
"loss": 1.
|
| 68 |
"step": 100
|
| 69 |
},
|
| 70 |
{
|
| 71 |
"epoch": 0.44,
|
| 72 |
-
"learning_rate": 4.
|
| 73 |
-
"loss": 1.
|
| 74 |
"step": 110
|
| 75 |
},
|
| 76 |
{
|
| 77 |
"epoch": 0.48,
|
| 78 |
-
"learning_rate": 4.
|
| 79 |
-
"loss": 1.
|
| 80 |
"step": 120
|
| 81 |
},
|
| 82 |
{
|
| 83 |
"epoch": 0.51,
|
| 84 |
-
"learning_rate": 4.
|
| 85 |
-
"loss": 1.
|
| 86 |
"step": 130
|
| 87 |
},
|
| 88 |
{
|
| 89 |
"epoch": 0.55,
|
| 90 |
-
"learning_rate": 4.
|
| 91 |
-
"loss": 1.
|
| 92 |
"step": 140
|
| 93 |
},
|
| 94 |
{
|
| 95 |
"epoch": 0.59,
|
| 96 |
-
"learning_rate": 4.
|
| 97 |
-
"loss": 1.
|
| 98 |
"step": 150
|
| 99 |
},
|
| 100 |
{
|
| 101 |
"epoch": 0.63,
|
| 102 |
-
"learning_rate": 4.
|
| 103 |
-
"loss": 1.
|
| 104 |
"step": 160
|
| 105 |
},
|
| 106 |
{
|
| 107 |
"epoch": 0.67,
|
| 108 |
-
"learning_rate": 4.
|
| 109 |
-
"loss": 1.
|
| 110 |
"step": 170
|
| 111 |
},
|
| 112 |
{
|
| 113 |
"epoch": 0.71,
|
| 114 |
-
"learning_rate": 4.
|
| 115 |
-
"loss": 1.
|
| 116 |
"step": 180
|
| 117 |
},
|
| 118 |
{
|
| 119 |
"epoch": 0.75,
|
| 120 |
-
"learning_rate": 4.
|
| 121 |
-
"loss": 1.
|
| 122 |
"step": 190
|
| 123 |
},
|
| 124 |
{
|
| 125 |
"epoch": 0.79,
|
| 126 |
-
"learning_rate": 4.
|
| 127 |
-
"loss": 1.
|
| 128 |
"step": 200
|
| 129 |
},
|
| 130 |
{
|
| 131 |
"epoch": 0.83,
|
| 132 |
-
"learning_rate": 4.
|
| 133 |
-
"loss": 1.
|
| 134 |
"step": 210
|
| 135 |
},
|
| 136 |
{
|
| 137 |
"epoch": 0.87,
|
| 138 |
-
"learning_rate":
|
| 139 |
-
"loss": 1.
|
| 140 |
"step": 220
|
| 141 |
},
|
| 142 |
{
|
| 143 |
"epoch": 0.91,
|
| 144 |
-
"learning_rate":
|
| 145 |
-
"loss": 1.
|
| 146 |
"step": 230
|
| 147 |
},
|
| 148 |
{
|
| 149 |
"epoch": 0.95,
|
| 150 |
-
"learning_rate":
|
| 151 |
-
"loss": 1.
|
| 152 |
"step": 240
|
| 153 |
},
|
| 154 |
{
|
| 155 |
"epoch": 0.99,
|
| 156 |
-
"learning_rate":
|
| 157 |
-
"loss": 1.
|
| 158 |
"step": 250
|
| 159 |
},
|
| 160 |
{
|
| 161 |
"epoch": 1.0,
|
| 162 |
-
"eval_accuracy": 0.
|
| 163 |
-
"eval_loss": 1.
|
| 164 |
-
"eval_runtime":
|
| 165 |
-
"eval_samples_per_second":
|
| 166 |
-
"eval_steps_per_second": 6.
|
| 167 |
"step": 252
|
| 168 |
},
|
| 169 |
{
|
| 170 |
"epoch": 1.03,
|
| 171 |
-
"learning_rate":
|
| 172 |
-
"loss": 1.
|
| 173 |
"step": 260
|
| 174 |
},
|
| 175 |
{
|
| 176 |
"epoch": 1.07,
|
| 177 |
-
"learning_rate":
|
| 178 |
-
"loss": 1.
|
| 179 |
"step": 270
|
| 180 |
},
|
| 181 |
{
|
| 182 |
"epoch": 1.11,
|
| 183 |
-
"learning_rate":
|
| 184 |
-
"loss": 1.
|
| 185 |
"step": 280
|
| 186 |
},
|
| 187 |
{
|
| 188 |
"epoch": 1.15,
|
| 189 |
-
"learning_rate":
|
| 190 |
-
"loss": 1.
|
| 191 |
"step": 290
|
| 192 |
},
|
| 193 |
{
|
| 194 |
"epoch": 1.19,
|
| 195 |
-
"learning_rate":
|
| 196 |
-
"loss": 1.
|
| 197 |
"step": 300
|
| 198 |
},
|
| 199 |
{
|
| 200 |
"epoch": 1.23,
|
| 201 |
-
"learning_rate":
|
| 202 |
-
"loss": 1.
|
| 203 |
"step": 310
|
| 204 |
},
|
| 205 |
{
|
| 206 |
"epoch": 1.27,
|
| 207 |
-
"learning_rate":
|
| 208 |
-
"loss": 1.
|
| 209 |
"step": 320
|
| 210 |
},
|
| 211 |
{
|
| 212 |
"epoch": 1.31,
|
| 213 |
-
"learning_rate":
|
| 214 |
-
"loss": 1.
|
| 215 |
"step": 330
|
| 216 |
},
|
| 217 |
{
|
| 218 |
"epoch": 1.35,
|
| 219 |
-
"learning_rate":
|
| 220 |
-
"loss": 1.
|
| 221 |
"step": 340
|
| 222 |
},
|
| 223 |
{
|
| 224 |
"epoch": 1.39,
|
| 225 |
-
"learning_rate":
|
| 226 |
-
"loss": 1.
|
| 227 |
"step": 350
|
| 228 |
},
|
| 229 |
{
|
| 230 |
"epoch": 1.43,
|
| 231 |
-
"learning_rate":
|
| 232 |
-
"loss": 1.
|
| 233 |
"step": 360
|
| 234 |
},
|
| 235 |
{
|
| 236 |
"epoch": 1.47,
|
| 237 |
-
"learning_rate":
|
| 238 |
-
"loss": 1.
|
| 239 |
"step": 370
|
| 240 |
},
|
| 241 |
{
|
| 242 |
"epoch": 1.5,
|
| 243 |
-
"learning_rate":
|
| 244 |
-
"loss": 1.
|
| 245 |
"step": 380
|
| 246 |
},
|
| 247 |
{
|
| 248 |
"epoch": 1.54,
|
| 249 |
-
"learning_rate":
|
| 250 |
-
"loss": 1.
|
| 251 |
"step": 390
|
| 252 |
},
|
| 253 |
{
|
| 254 |
"epoch": 1.58,
|
| 255 |
-
"learning_rate":
|
| 256 |
-
"loss": 1.
|
| 257 |
"step": 400
|
| 258 |
},
|
| 259 |
{
|
| 260 |
"epoch": 1.62,
|
| 261 |
-
"learning_rate":
|
| 262 |
-
"loss": 1.
|
| 263 |
"step": 410
|
| 264 |
},
|
| 265 |
{
|
| 266 |
"epoch": 1.66,
|
| 267 |
-
"learning_rate":
|
| 268 |
-
"loss": 1.
|
| 269 |
"step": 420
|
| 270 |
},
|
| 271 |
{
|
| 272 |
"epoch": 1.7,
|
| 273 |
-
"learning_rate":
|
| 274 |
-
"loss": 1.
|
| 275 |
"step": 430
|
| 276 |
},
|
| 277 |
{
|
| 278 |
"epoch": 1.74,
|
| 279 |
-
"learning_rate":
|
| 280 |
-
"loss": 1.
|
| 281 |
"step": 440
|
| 282 |
},
|
| 283 |
{
|
| 284 |
"epoch": 1.78,
|
| 285 |
-
"learning_rate":
|
| 286 |
-
"loss": 1.
|
| 287 |
"step": 450
|
| 288 |
},
|
| 289 |
{
|
| 290 |
"epoch": 1.82,
|
| 291 |
-
"learning_rate":
|
| 292 |
-
"loss": 1.
|
| 293 |
"step": 460
|
| 294 |
},
|
| 295 |
{
|
| 296 |
"epoch": 1.86,
|
| 297 |
-
"learning_rate":
|
| 298 |
-
"loss": 1.
|
| 299 |
"step": 470
|
| 300 |
},
|
| 301 |
{
|
| 302 |
"epoch": 1.9,
|
| 303 |
-
"learning_rate":
|
| 304 |
-
"loss": 1.
|
| 305 |
"step": 480
|
| 306 |
},
|
| 307 |
{
|
| 308 |
"epoch": 1.94,
|
| 309 |
-
"learning_rate":
|
| 310 |
-
"loss": 1.
|
| 311 |
"step": 490
|
| 312 |
},
|
| 313 |
{
|
| 314 |
"epoch": 1.98,
|
| 315 |
-
"learning_rate":
|
| 316 |
-
"loss": 1.
|
| 317 |
"step": 500
|
| 318 |
},
|
| 319 |
{
|
| 320 |
"epoch": 2.0,
|
| 321 |
-
"eval_accuracy": 0.
|
| 322 |
-
"eval_loss": 1.
|
| 323 |
-
"eval_runtime":
|
| 324 |
-
"eval_samples_per_second":
|
| 325 |
-
"eval_steps_per_second": 6.
|
| 326 |
"step": 505
|
| 327 |
},
|
| 328 |
{
|
| 329 |
"epoch": 2.02,
|
| 330 |
-
"learning_rate":
|
| 331 |
-
"loss": 1.
|
| 332 |
"step": 510
|
| 333 |
},
|
| 334 |
{
|
| 335 |
"epoch": 2.06,
|
| 336 |
-
"learning_rate":
|
| 337 |
-
"loss": 1.
|
| 338 |
"step": 520
|
| 339 |
},
|
| 340 |
{
|
| 341 |
"epoch": 2.1,
|
| 342 |
-
"learning_rate":
|
| 343 |
-
"loss": 1.
|
| 344 |
"step": 530
|
| 345 |
},
|
| 346 |
{
|
| 347 |
"epoch": 2.14,
|
| 348 |
-
"learning_rate":
|
| 349 |
-
"loss": 1.
|
| 350 |
"step": 540
|
| 351 |
},
|
| 352 |
{
|
| 353 |
"epoch": 2.18,
|
| 354 |
-
"learning_rate":
|
| 355 |
-
"loss": 1.
|
| 356 |
"step": 550
|
| 357 |
},
|
| 358 |
{
|
| 359 |
"epoch": 2.22,
|
| 360 |
-
"learning_rate":
|
| 361 |
-
"loss": 1.
|
| 362 |
"step": 560
|
| 363 |
},
|
| 364 |
{
|
| 365 |
"epoch": 2.26,
|
| 366 |
-
"learning_rate":
|
| 367 |
-
"loss": 1.
|
| 368 |
"step": 570
|
| 369 |
},
|
| 370 |
{
|
| 371 |
"epoch": 2.3,
|
| 372 |
-
"learning_rate":
|
| 373 |
-
"loss": 1.
|
| 374 |
"step": 580
|
| 375 |
},
|
| 376 |
{
|
| 377 |
"epoch": 2.34,
|
| 378 |
-
"learning_rate":
|
| 379 |
-
"loss": 1.
|
| 380 |
"step": 590
|
| 381 |
},
|
| 382 |
{
|
| 383 |
"epoch": 2.38,
|
| 384 |
-
"learning_rate":
|
| 385 |
-
"loss": 1.
|
| 386 |
"step": 600
|
| 387 |
},
|
| 388 |
{
|
| 389 |
"epoch": 2.42,
|
| 390 |
-
"learning_rate":
|
| 391 |
-
"loss": 1.
|
| 392 |
"step": 610
|
| 393 |
},
|
| 394 |
{
|
| 395 |
"epoch": 2.46,
|
| 396 |
-
"learning_rate":
|
| 397 |
-
"loss": 1.
|
| 398 |
"step": 620
|
| 399 |
},
|
| 400 |
{
|
| 401 |
"epoch": 2.5,
|
| 402 |
-
"learning_rate":
|
| 403 |
-
"loss": 1.
|
| 404 |
"step": 630
|
| 405 |
},
|
| 406 |
{
|
| 407 |
"epoch": 2.53,
|
| 408 |
-
"learning_rate":
|
| 409 |
-
"loss": 1.
|
| 410 |
"step": 640
|
| 411 |
},
|
| 412 |
{
|
| 413 |
"epoch": 2.57,
|
| 414 |
-
"learning_rate":
|
| 415 |
-
"loss": 1.
|
| 416 |
"step": 650
|
| 417 |
},
|
| 418 |
{
|
| 419 |
"epoch": 2.61,
|
| 420 |
-
"learning_rate":
|
| 421 |
-
"loss": 1.
|
| 422 |
"step": 660
|
| 423 |
},
|
| 424 |
{
|
| 425 |
"epoch": 2.65,
|
| 426 |
-
"learning_rate":
|
| 427 |
-
"loss": 1.
|
| 428 |
"step": 670
|
| 429 |
},
|
| 430 |
{
|
| 431 |
"epoch": 2.69,
|
| 432 |
-
"learning_rate":
|
| 433 |
-
"loss": 1.
|
| 434 |
"step": 680
|
| 435 |
},
|
| 436 |
{
|
| 437 |
"epoch": 2.73,
|
| 438 |
-
"learning_rate":
|
| 439 |
-
"loss": 1.
|
| 440 |
"step": 690
|
| 441 |
},
|
| 442 |
{
|
| 443 |
"epoch": 2.77,
|
| 444 |
-
"learning_rate":
|
| 445 |
-
"loss": 1.
|
| 446 |
"step": 700
|
| 447 |
},
|
| 448 |
{
|
| 449 |
"epoch": 2.81,
|
| 450 |
-
"learning_rate":
|
| 451 |
-
"loss": 1.
|
| 452 |
"step": 710
|
| 453 |
},
|
| 454 |
{
|
| 455 |
"epoch": 2.85,
|
| 456 |
-
"learning_rate": 2.
|
| 457 |
-
"loss": 1.
|
| 458 |
"step": 720
|
| 459 |
},
|
| 460 |
{
|
| 461 |
"epoch": 2.89,
|
| 462 |
-
"learning_rate":
|
| 463 |
-
"loss": 1.
|
| 464 |
"step": 730
|
| 465 |
},
|
| 466 |
{
|
| 467 |
"epoch": 2.93,
|
| 468 |
-
"learning_rate":
|
| 469 |
-
"loss": 1.
|
| 470 |
"step": 740
|
| 471 |
},
|
| 472 |
{
|
| 473 |
"epoch": 2.97,
|
| 474 |
-
"learning_rate":
|
| 475 |
-
"loss": 1.
|
| 476 |
"step": 750
|
| 477 |
},
|
| 478 |
{
|
| 479 |
-
"epoch":
|
| 480 |
-
"eval_accuracy": 0.
|
| 481 |
-
"eval_loss": 1.
|
| 482 |
-
"eval_runtime":
|
| 483 |
-
"eval_samples_per_second":
|
| 484 |
-
"eval_steps_per_second": 6.
|
| 485 |
-
"step":
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
| 486 |
},
|
| 487 |
{
|
| 488 |
-
"epoch":
|
| 489 |
-
"step":
|
| 490 |
-
"total_flos":
|
| 491 |
-
"train_loss": 1.
|
| 492 |
-
"train_runtime":
|
| 493 |
-
"train_samples_per_second":
|
| 494 |
-
"train_steps_per_second": 0.
|
| 495 |
}
|
| 496 |
],
|
| 497 |
-
"max_steps":
|
| 498 |
-
"num_train_epochs":
|
| 499 |
-
"total_flos":
|
| 500 |
"trial_name": null,
|
| 501 |
"trial_params": null
|
| 502 |
}
|
|
|
|
| 1 |
{
|
| 2 |
+
"best_metric": 0.4446921147952076,
|
| 3 |
+
"best_model_checkpoint": "microsoft/resnet-50/checkpoint-1260",
|
| 4 |
+
"epoch": 4.99009900990099,
|
| 5 |
+
"global_step": 1260,
|
| 6 |
"is_hyper_param_search": false,
|
| 7 |
"is_local_process_zero": true,
|
| 8 |
"is_world_process_zero": true,
|
| 9 |
"log_history": [
|
| 10 |
{
|
| 11 |
"epoch": 0.04,
|
| 12 |
+
"learning_rate": 3.968253968253968e-06,
|
| 13 |
+
"loss": 1.7784,
|
| 14 |
"step": 10
|
| 15 |
},
|
| 16 |
{
|
| 17 |
"epoch": 0.08,
|
| 18 |
+
"learning_rate": 7.936507936507936e-06,
|
| 19 |
+
"loss": 1.786,
|
| 20 |
"step": 20
|
| 21 |
},
|
| 22 |
{
|
| 23 |
"epoch": 0.12,
|
| 24 |
+
"learning_rate": 1.1904761904761905e-05,
|
| 25 |
+
"loss": 1.8132,
|
| 26 |
"step": 30
|
| 27 |
},
|
| 28 |
{
|
| 29 |
"epoch": 0.16,
|
| 30 |
+
"learning_rate": 1.5873015873015872e-05,
|
| 31 |
+
"loss": 1.7916,
|
| 32 |
"step": 40
|
| 33 |
},
|
| 34 |
{
|
| 35 |
"epoch": 0.2,
|
| 36 |
+
"learning_rate": 1.984126984126984e-05,
|
| 37 |
+
"loss": 1.7859,
|
| 38 |
"step": 50
|
| 39 |
},
|
| 40 |
{
|
| 41 |
"epoch": 0.24,
|
| 42 |
+
"learning_rate": 2.380952380952381e-05,
|
| 43 |
+
"loss": 1.7751,
|
| 44 |
"step": 60
|
| 45 |
},
|
| 46 |
{
|
| 47 |
"epoch": 0.28,
|
| 48 |
+
"learning_rate": 2.777777777777778e-05,
|
| 49 |
+
"loss": 1.8154,
|
| 50 |
"step": 70
|
| 51 |
},
|
| 52 |
{
|
| 53 |
"epoch": 0.32,
|
| 54 |
+
"learning_rate": 3.1746031746031745e-05,
|
| 55 |
+
"loss": 1.7868,
|
| 56 |
"step": 80
|
| 57 |
},
|
| 58 |
{
|
| 59 |
"epoch": 0.36,
|
| 60 |
+
"learning_rate": 3.571428571428572e-05,
|
| 61 |
+
"loss": 1.7624,
|
| 62 |
"step": 90
|
| 63 |
},
|
| 64 |
{
|
| 65 |
"epoch": 0.4,
|
| 66 |
+
"learning_rate": 3.968253968253968e-05,
|
| 67 |
+
"loss": 1.7823,
|
| 68 |
"step": 100
|
| 69 |
},
|
| 70 |
{
|
| 71 |
"epoch": 0.44,
|
| 72 |
+
"learning_rate": 4.3650793650793655e-05,
|
| 73 |
+
"loss": 1.7823,
|
| 74 |
"step": 110
|
| 75 |
},
|
| 76 |
{
|
| 77 |
"epoch": 0.48,
|
| 78 |
+
"learning_rate": 4.761904761904762e-05,
|
| 79 |
+
"loss": 1.7932,
|
| 80 |
"step": 120
|
| 81 |
},
|
| 82 |
{
|
| 83 |
"epoch": 0.51,
|
| 84 |
+
"learning_rate": 4.982363315696649e-05,
|
| 85 |
+
"loss": 1.7863,
|
| 86 |
"step": 130
|
| 87 |
},
|
| 88 |
{
|
| 89 |
"epoch": 0.55,
|
| 90 |
+
"learning_rate": 4.938271604938271e-05,
|
| 91 |
+
"loss": 1.7806,
|
| 92 |
"step": 140
|
| 93 |
},
|
| 94 |
{
|
| 95 |
"epoch": 0.59,
|
| 96 |
+
"learning_rate": 4.894179894179895e-05,
|
| 97 |
+
"loss": 1.7706,
|
| 98 |
"step": 150
|
| 99 |
},
|
| 100 |
{
|
| 101 |
"epoch": 0.63,
|
| 102 |
+
"learning_rate": 4.850088183421517e-05,
|
| 103 |
+
"loss": 1.7629,
|
| 104 |
"step": 160
|
| 105 |
},
|
| 106 |
{
|
| 107 |
"epoch": 0.67,
|
| 108 |
+
"learning_rate": 4.8059964726631394e-05,
|
| 109 |
+
"loss": 1.7731,
|
| 110 |
"step": 170
|
| 111 |
},
|
| 112 |
{
|
| 113 |
"epoch": 0.71,
|
| 114 |
+
"learning_rate": 4.761904761904762e-05,
|
| 115 |
+
"loss": 1.7435,
|
| 116 |
"step": 180
|
| 117 |
},
|
| 118 |
{
|
| 119 |
"epoch": 0.75,
|
| 120 |
+
"learning_rate": 4.717813051146385e-05,
|
| 121 |
+
"loss": 1.7371,
|
| 122 |
"step": 190
|
| 123 |
},
|
| 124 |
{
|
| 125 |
"epoch": 0.79,
|
| 126 |
+
"learning_rate": 4.673721340388007e-05,
|
| 127 |
+
"loss": 1.729,
|
| 128 |
"step": 200
|
| 129 |
},
|
| 130 |
{
|
| 131 |
"epoch": 0.83,
|
| 132 |
+
"learning_rate": 4.62962962962963e-05,
|
| 133 |
+
"loss": 1.7516,
|
| 134 |
"step": 210
|
| 135 |
},
|
| 136 |
{
|
| 137 |
"epoch": 0.87,
|
| 138 |
+
"learning_rate": 4.585537918871252e-05,
|
| 139 |
+
"loss": 1.7337,
|
| 140 |
"step": 220
|
| 141 |
},
|
| 142 |
{
|
| 143 |
"epoch": 0.91,
|
| 144 |
+
"learning_rate": 4.541446208112875e-05,
|
| 145 |
+
"loss": 1.7118,
|
| 146 |
"step": 230
|
| 147 |
},
|
| 148 |
{
|
| 149 |
"epoch": 0.95,
|
| 150 |
+
"learning_rate": 4.4973544973544974e-05,
|
| 151 |
+
"loss": 1.7141,
|
| 152 |
"step": 240
|
| 153 |
},
|
| 154 |
{
|
| 155 |
"epoch": 0.99,
|
| 156 |
+
"learning_rate": 4.45326278659612e-05,
|
| 157 |
+
"loss": 1.7252,
|
| 158 |
"step": 250
|
| 159 |
},
|
| 160 |
{
|
| 161 |
"epoch": 1.0,
|
| 162 |
+
"eval_accuracy": 0.27528559487322374,
|
| 163 |
+
"eval_loss": 1.7069034576416016,
|
| 164 |
+
"eval_runtime": 16.7761,
|
| 165 |
+
"eval_samples_per_second": 213.935,
|
| 166 |
+
"eval_steps_per_second": 6.736,
|
| 167 |
"step": 252
|
| 168 |
},
|
| 169 |
{
|
| 170 |
"epoch": 1.03,
|
| 171 |
+
"learning_rate": 4.409171075837743e-05,
|
| 172 |
+
"loss": 1.7246,
|
| 173 |
"step": 260
|
| 174 |
},
|
| 175 |
{
|
| 176 |
"epoch": 1.07,
|
| 177 |
+
"learning_rate": 4.3650793650793655e-05,
|
| 178 |
+
"loss": 1.7388,
|
| 179 |
"step": 270
|
| 180 |
},
|
| 181 |
{
|
| 182 |
"epoch": 1.11,
|
| 183 |
+
"learning_rate": 4.3209876543209875e-05,
|
| 184 |
+
"loss": 1.7169,
|
| 185 |
"step": 280
|
| 186 |
},
|
| 187 |
{
|
| 188 |
"epoch": 1.15,
|
| 189 |
+
"learning_rate": 4.27689594356261e-05,
|
| 190 |
+
"loss": 1.6987,
|
| 191 |
"step": 290
|
| 192 |
},
|
| 193 |
{
|
| 194 |
"epoch": 1.19,
|
| 195 |
+
"learning_rate": 4.232804232804233e-05,
|
| 196 |
+
"loss": 1.7117,
|
| 197 |
"step": 300
|
| 198 |
},
|
| 199 |
{
|
| 200 |
"epoch": 1.23,
|
| 201 |
+
"learning_rate": 4.1887125220458555e-05,
|
| 202 |
+
"loss": 1.7117,
|
| 203 |
"step": 310
|
| 204 |
},
|
| 205 |
{
|
| 206 |
"epoch": 1.27,
|
| 207 |
+
"learning_rate": 4.144620811287478e-05,
|
| 208 |
+
"loss": 1.7248,
|
| 209 |
"step": 320
|
| 210 |
},
|
| 211 |
{
|
| 212 |
"epoch": 1.31,
|
| 213 |
+
"learning_rate": 4.100529100529101e-05,
|
| 214 |
+
"loss": 1.7123,
|
| 215 |
"step": 330
|
| 216 |
},
|
| 217 |
{
|
| 218 |
"epoch": 1.35,
|
| 219 |
+
"learning_rate": 4.056437389770723e-05,
|
| 220 |
+
"loss": 1.7096,
|
| 221 |
"step": 340
|
| 222 |
},
|
| 223 |
{
|
| 224 |
"epoch": 1.39,
|
| 225 |
+
"learning_rate": 4.012345679012346e-05,
|
| 226 |
+
"loss": 1.6905,
|
| 227 |
"step": 350
|
| 228 |
},
|
| 229 |
{
|
| 230 |
"epoch": 1.43,
|
| 231 |
+
"learning_rate": 3.968253968253968e-05,
|
| 232 |
+
"loss": 1.7309,
|
| 233 |
"step": 360
|
| 234 |
},
|
| 235 |
{
|
| 236 |
"epoch": 1.47,
|
| 237 |
+
"learning_rate": 3.924162257495591e-05,
|
| 238 |
+
"loss": 1.7151,
|
| 239 |
"step": 370
|
| 240 |
},
|
| 241 |
{
|
| 242 |
"epoch": 1.5,
|
| 243 |
+
"learning_rate": 3.8800705467372136e-05,
|
| 244 |
+
"loss": 1.7067,
|
| 245 |
"step": 380
|
| 246 |
},
|
| 247 |
{
|
| 248 |
"epoch": 1.54,
|
| 249 |
+
"learning_rate": 3.835978835978836e-05,
|
| 250 |
+
"loss": 1.6982,
|
| 251 |
"step": 390
|
| 252 |
},
|
| 253 |
{
|
| 254 |
"epoch": 1.58,
|
| 255 |
+
"learning_rate": 3.791887125220458e-05,
|
| 256 |
+
"loss": 1.6959,
|
| 257 |
"step": 400
|
| 258 |
},
|
| 259 |
{
|
| 260 |
"epoch": 1.62,
|
| 261 |
+
"learning_rate": 3.7477954144620817e-05,
|
| 262 |
+
"loss": 1.6794,
|
| 263 |
"step": 410
|
| 264 |
},
|
| 265 |
{
|
| 266 |
"epoch": 1.66,
|
| 267 |
+
"learning_rate": 3.7037037037037037e-05,
|
| 268 |
+
"loss": 1.6696,
|
| 269 |
"step": 420
|
| 270 |
},
|
| 271 |
{
|
| 272 |
"epoch": 1.7,
|
| 273 |
+
"learning_rate": 3.659611992945326e-05,
|
| 274 |
+
"loss": 1.6761,
|
| 275 |
"step": 430
|
| 276 |
},
|
| 277 |
{
|
| 278 |
"epoch": 1.74,
|
| 279 |
+
"learning_rate": 3.615520282186949e-05,
|
| 280 |
+
"loss": 1.6844,
|
| 281 |
"step": 440
|
| 282 |
},
|
| 283 |
{
|
| 284 |
"epoch": 1.78,
|
| 285 |
+
"learning_rate": 3.571428571428572e-05,
|
| 286 |
+
"loss": 1.6666,
|
| 287 |
"step": 450
|
| 288 |
},
|
| 289 |
{
|
| 290 |
"epoch": 1.82,
|
| 291 |
+
"learning_rate": 3.527336860670194e-05,
|
| 292 |
+
"loss": 1.6767,
|
| 293 |
"step": 460
|
| 294 |
},
|
| 295 |
{
|
| 296 |
"epoch": 1.86,
|
| 297 |
+
"learning_rate": 3.483245149911817e-05,
|
| 298 |
+
"loss": 1.6578,
|
| 299 |
"step": 470
|
| 300 |
},
|
| 301 |
{
|
| 302 |
"epoch": 1.9,
|
| 303 |
+
"learning_rate": 3.439153439153439e-05,
|
| 304 |
+
"loss": 1.6493,
|
| 305 |
"step": 480
|
| 306 |
},
|
| 307 |
{
|
| 308 |
"epoch": 1.94,
|
| 309 |
+
"learning_rate": 3.395061728395062e-05,
|
| 310 |
+
"loss": 1.6547,
|
| 311 |
"step": 490
|
| 312 |
},
|
| 313 |
{
|
| 314 |
"epoch": 1.98,
|
| 315 |
+
"learning_rate": 3.3509700176366844e-05,
|
| 316 |
+
"loss": 1.6386,
|
| 317 |
"step": 500
|
| 318 |
},
|
| 319 |
{
|
| 320 |
"epoch": 2.0,
|
| 321 |
+
"eval_accuracy": 0.3911953190303706,
|
| 322 |
+
"eval_loss": 1.5798698663711548,
|
| 323 |
+
"eval_runtime": 16.5653,
|
| 324 |
+
"eval_samples_per_second": 216.658,
|
| 325 |
+
"eval_steps_per_second": 6.821,
|
| 326 |
"step": 505
|
| 327 |
},
|
| 328 |
{
|
| 329 |
"epoch": 2.02,
|
| 330 |
+
"learning_rate": 3.306878306878307e-05,
|
| 331 |
+
"loss": 1.675,
|
| 332 |
"step": 510
|
| 333 |
},
|
| 334 |
{
|
| 335 |
"epoch": 2.06,
|
| 336 |
+
"learning_rate": 3.262786596119929e-05,
|
| 337 |
+
"loss": 1.6512,
|
| 338 |
"step": 520
|
| 339 |
},
|
| 340 |
{
|
| 341 |
"epoch": 2.1,
|
| 342 |
+
"learning_rate": 3.2186948853615525e-05,
|
| 343 |
+
"loss": 1.6484,
|
| 344 |
"step": 530
|
| 345 |
},
|
| 346 |
{
|
| 347 |
"epoch": 2.14,
|
| 348 |
+
"learning_rate": 3.1746031746031745e-05,
|
| 349 |
+
"loss": 1.6573,
|
| 350 |
"step": 540
|
| 351 |
},
|
| 352 |
{
|
| 353 |
"epoch": 2.18,
|
| 354 |
+
"learning_rate": 3.130511463844797e-05,
|
| 355 |
+
"loss": 1.6452,
|
| 356 |
"step": 550
|
| 357 |
},
|
| 358 |
{
|
| 359 |
"epoch": 2.22,
|
| 360 |
+
"learning_rate": 3.08641975308642e-05,
|
| 361 |
+
"loss": 1.6462,
|
| 362 |
"step": 560
|
| 363 |
},
|
| 364 |
{
|
| 365 |
"epoch": 2.26,
|
| 366 |
+
"learning_rate": 3.0423280423280425e-05,
|
| 367 |
+
"loss": 1.6164,
|
| 368 |
"step": 570
|
| 369 |
},
|
| 370 |
{
|
| 371 |
"epoch": 2.3,
|
| 372 |
+
"learning_rate": 2.998236331569665e-05,
|
| 373 |
+
"loss": 1.6062,
|
| 374 |
"step": 580
|
| 375 |
},
|
| 376 |
{
|
| 377 |
"epoch": 2.34,
|
| 378 |
+
"learning_rate": 2.954144620811288e-05,
|
| 379 |
+
"loss": 1.6517,
|
| 380 |
"step": 590
|
| 381 |
},
|
| 382 |
{
|
| 383 |
"epoch": 2.38,
|
| 384 |
+
"learning_rate": 2.91005291005291e-05,
|
| 385 |
+
"loss": 1.6156,
|
| 386 |
"step": 600
|
| 387 |
},
|
| 388 |
{
|
| 389 |
"epoch": 2.42,
|
| 390 |
+
"learning_rate": 2.865961199294533e-05,
|
| 391 |
+
"loss": 1.6497,
|
| 392 |
"step": 610
|
| 393 |
},
|
| 394 |
{
|
| 395 |
"epoch": 2.46,
|
| 396 |
+
"learning_rate": 2.8218694885361552e-05,
|
| 397 |
+
"loss": 1.6526,
|
| 398 |
"step": 620
|
| 399 |
},
|
| 400 |
{
|
| 401 |
"epoch": 2.5,
|
| 402 |
+
"learning_rate": 2.777777777777778e-05,
|
| 403 |
+
"loss": 1.6249,
|
| 404 |
"step": 630
|
| 405 |
},
|
| 406 |
{
|
| 407 |
"epoch": 2.53,
|
| 408 |
+
"learning_rate": 2.7336860670194003e-05,
|
| 409 |
+
"loss": 1.6367,
|
| 410 |
"step": 640
|
| 411 |
},
|
| 412 |
{
|
| 413 |
"epoch": 2.57,
|
| 414 |
+
"learning_rate": 2.6895943562610233e-05,
|
| 415 |
+
"loss": 1.6321,
|
| 416 |
"step": 650
|
| 417 |
},
|
| 418 |
{
|
| 419 |
"epoch": 2.61,
|
| 420 |
+
"learning_rate": 2.6455026455026456e-05,
|
| 421 |
+
"loss": 1.6255,
|
| 422 |
"step": 660
|
| 423 |
},
|
| 424 |
{
|
| 425 |
"epoch": 2.65,
|
| 426 |
+
"learning_rate": 2.6014109347442683e-05,
|
| 427 |
+
"loss": 1.6215,
|
| 428 |
"step": 670
|
| 429 |
},
|
| 430 |
{
|
| 431 |
"epoch": 2.69,
|
| 432 |
+
"learning_rate": 2.5573192239858906e-05,
|
| 433 |
+
"loss": 1.6425,
|
| 434 |
"step": 680
|
| 435 |
},
|
| 436 |
{
|
| 437 |
"epoch": 2.73,
|
| 438 |
+
"learning_rate": 2.5132275132275137e-05,
|
| 439 |
+
"loss": 1.6064,
|
| 440 |
"step": 690
|
| 441 |
},
|
| 442 |
{
|
| 443 |
"epoch": 2.77,
|
| 444 |
+
"learning_rate": 2.4691358024691357e-05,
|
| 445 |
+
"loss": 1.6093,
|
| 446 |
"step": 700
|
| 447 |
},
|
| 448 |
{
|
| 449 |
"epoch": 2.81,
|
| 450 |
+
"learning_rate": 2.4250440917107583e-05,
|
| 451 |
+
"loss": 1.6022,
|
| 452 |
"step": 710
|
| 453 |
},
|
| 454 |
{
|
| 455 |
"epoch": 2.85,
|
| 456 |
+
"learning_rate": 2.380952380952381e-05,
|
| 457 |
+
"loss": 1.6309,
|
| 458 |
"step": 720
|
| 459 |
},
|
| 460 |
{
|
| 461 |
"epoch": 2.89,
|
| 462 |
+
"learning_rate": 2.3368606701940034e-05,
|
| 463 |
+
"loss": 1.6112,
|
| 464 |
"step": 730
|
| 465 |
},
|
| 466 |
{
|
| 467 |
"epoch": 2.93,
|
| 468 |
+
"learning_rate": 2.292768959435626e-05,
|
| 469 |
+
"loss": 1.6253,
|
| 470 |
"step": 740
|
| 471 |
},
|
| 472 |
{
|
| 473 |
"epoch": 2.97,
|
| 474 |
+
"learning_rate": 2.2486772486772487e-05,
|
| 475 |
+
"loss": 1.617,
|
| 476 |
"step": 750
|
| 477 |
},
|
| 478 |
{
|
| 479 |
+
"epoch": 3.0,
|
| 480 |
+
"eval_accuracy": 0.42741710782947895,
|
| 481 |
+
"eval_loss": 1.5154473781585693,
|
| 482 |
+
"eval_runtime": 17.1814,
|
| 483 |
+
"eval_samples_per_second": 208.889,
|
| 484 |
+
"eval_steps_per_second": 6.577,
|
| 485 |
+
"step": 757
|
| 486 |
+
},
|
| 487 |
+
{
|
| 488 |
+
"epoch": 3.01,
|
| 489 |
+
"learning_rate": 2.2045855379188714e-05,
|
| 490 |
+
"loss": 1.6162,
|
| 491 |
+
"step": 760
|
| 492 |
+
},
|
| 493 |
+
{
|
| 494 |
+
"epoch": 3.05,
|
| 495 |
+
"learning_rate": 2.1604938271604937e-05,
|
| 496 |
+
"loss": 1.6166,
|
| 497 |
+
"step": 770
|
| 498 |
+
},
|
| 499 |
+
{
|
| 500 |
+
"epoch": 3.09,
|
| 501 |
+
"learning_rate": 2.1164021164021164e-05,
|
| 502 |
+
"loss": 1.5919,
|
| 503 |
+
"step": 780
|
| 504 |
+
},
|
| 505 |
+
{
|
| 506 |
+
"epoch": 3.13,
|
| 507 |
+
"learning_rate": 2.072310405643739e-05,
|
| 508 |
+
"loss": 1.6002,
|
| 509 |
+
"step": 790
|
| 510 |
+
},
|
| 511 |
+
{
|
| 512 |
+
"epoch": 3.17,
|
| 513 |
+
"learning_rate": 2.0282186948853614e-05,
|
| 514 |
+
"loss": 1.6324,
|
| 515 |
+
"step": 800
|
| 516 |
+
},
|
| 517 |
+
{
|
| 518 |
+
"epoch": 3.21,
|
| 519 |
+
"learning_rate": 1.984126984126984e-05,
|
| 520 |
+
"loss": 1.5883,
|
| 521 |
+
"step": 810
|
| 522 |
+
},
|
| 523 |
+
{
|
| 524 |
+
"epoch": 3.25,
|
| 525 |
+
"learning_rate": 1.9400352733686068e-05,
|
| 526 |
+
"loss": 1.587,
|
| 527 |
+
"step": 820
|
| 528 |
+
},
|
| 529 |
+
{
|
| 530 |
+
"epoch": 3.29,
|
| 531 |
+
"learning_rate": 1.895943562610229e-05,
|
| 532 |
+
"loss": 1.6264,
|
| 533 |
+
"step": 830
|
| 534 |
+
},
|
| 535 |
+
{
|
| 536 |
+
"epoch": 3.33,
|
| 537 |
+
"learning_rate": 1.8518518518518518e-05,
|
| 538 |
+
"loss": 1.5798,
|
| 539 |
+
"step": 840
|
| 540 |
+
},
|
| 541 |
+
{
|
| 542 |
+
"epoch": 3.37,
|
| 543 |
+
"learning_rate": 1.8077601410934745e-05,
|
| 544 |
+
"loss": 1.6045,
|
| 545 |
+
"step": 850
|
| 546 |
+
},
|
| 547 |
+
{
|
| 548 |
+
"epoch": 3.41,
|
| 549 |
+
"learning_rate": 1.763668430335097e-05,
|
| 550 |
+
"loss": 1.6234,
|
| 551 |
+
"step": 860
|
| 552 |
+
},
|
| 553 |
+
{
|
| 554 |
+
"epoch": 3.45,
|
| 555 |
+
"learning_rate": 1.7195767195767195e-05,
|
| 556 |
+
"loss": 1.5851,
|
| 557 |
+
"step": 870
|
| 558 |
+
},
|
| 559 |
+
{
|
| 560 |
+
"epoch": 3.49,
|
| 561 |
+
"learning_rate": 1.6754850088183422e-05,
|
| 562 |
+
"loss": 1.6007,
|
| 563 |
+
"step": 880
|
| 564 |
+
},
|
| 565 |
+
{
|
| 566 |
+
"epoch": 3.52,
|
| 567 |
+
"learning_rate": 1.6313932980599646e-05,
|
| 568 |
+
"loss": 1.6145,
|
| 569 |
+
"step": 890
|
| 570 |
+
},
|
| 571 |
+
{
|
| 572 |
+
"epoch": 3.56,
|
| 573 |
+
"learning_rate": 1.5873015873015872e-05,
|
| 574 |
+
"loss": 1.5742,
|
| 575 |
+
"step": 900
|
| 576 |
+
},
|
| 577 |
+
{
|
| 578 |
+
"epoch": 3.6,
|
| 579 |
+
"learning_rate": 1.54320987654321e-05,
|
| 580 |
+
"loss": 1.5846,
|
| 581 |
+
"step": 910
|
| 582 |
+
},
|
| 583 |
+
{
|
| 584 |
+
"epoch": 3.64,
|
| 585 |
+
"learning_rate": 1.4991181657848324e-05,
|
| 586 |
+
"loss": 1.5921,
|
| 587 |
+
"step": 920
|
| 588 |
+
},
|
| 589 |
+
{
|
| 590 |
+
"epoch": 3.68,
|
| 591 |
+
"learning_rate": 1.455026455026455e-05,
|
| 592 |
+
"loss": 1.6146,
|
| 593 |
+
"step": 930
|
| 594 |
+
},
|
| 595 |
+
{
|
| 596 |
+
"epoch": 3.72,
|
| 597 |
+
"learning_rate": 1.4109347442680776e-05,
|
| 598 |
+
"loss": 1.5937,
|
| 599 |
+
"step": 940
|
| 600 |
+
},
|
| 601 |
+
{
|
| 602 |
+
"epoch": 3.76,
|
| 603 |
+
"learning_rate": 1.3668430335097001e-05,
|
| 604 |
+
"loss": 1.5761,
|
| 605 |
+
"step": 950
|
| 606 |
+
},
|
| 607 |
+
{
|
| 608 |
+
"epoch": 3.8,
|
| 609 |
+
"learning_rate": 1.3227513227513228e-05,
|
| 610 |
+
"loss": 1.6056,
|
| 611 |
+
"step": 960
|
| 612 |
+
},
|
| 613 |
+
{
|
| 614 |
+
"epoch": 3.84,
|
| 615 |
+
"learning_rate": 1.2786596119929453e-05,
|
| 616 |
+
"loss": 1.5614,
|
| 617 |
+
"step": 970
|
| 618 |
+
},
|
| 619 |
+
{
|
| 620 |
+
"epoch": 3.88,
|
| 621 |
+
"learning_rate": 1.2345679012345678e-05,
|
| 622 |
+
"loss": 1.578,
|
| 623 |
+
"step": 980
|
| 624 |
+
},
|
| 625 |
+
{
|
| 626 |
+
"epoch": 3.92,
|
| 627 |
+
"learning_rate": 1.1904761904761905e-05,
|
| 628 |
+
"loss": 1.5495,
|
| 629 |
+
"step": 990
|
| 630 |
+
},
|
| 631 |
+
{
|
| 632 |
+
"epoch": 3.96,
|
| 633 |
+
"learning_rate": 1.146384479717813e-05,
|
| 634 |
+
"loss": 1.5949,
|
| 635 |
+
"step": 1000
|
| 636 |
+
},
|
| 637 |
+
{
|
| 638 |
+
"epoch": 4.0,
|
| 639 |
+
"learning_rate": 1.1022927689594357e-05,
|
| 640 |
+
"loss": 1.5795,
|
| 641 |
+
"step": 1010
|
| 642 |
+
},
|
| 643 |
+
{
|
| 644 |
+
"epoch": 4.0,
|
| 645 |
+
"eval_accuracy": 0.43772638617999443,
|
| 646 |
+
"eval_loss": 1.4835728406906128,
|
| 647 |
+
"eval_runtime": 16.6725,
|
| 648 |
+
"eval_samples_per_second": 215.265,
|
| 649 |
+
"eval_steps_per_second": 6.778,
|
| 650 |
+
"step": 1010
|
| 651 |
+
},
|
| 652 |
+
{
|
| 653 |
+
"epoch": 4.04,
|
| 654 |
+
"learning_rate": 1.0582010582010582e-05,
|
| 655 |
+
"loss": 1.5641,
|
| 656 |
+
"step": 1020
|
| 657 |
+
},
|
| 658 |
+
{
|
| 659 |
+
"epoch": 4.08,
|
| 660 |
+
"learning_rate": 1.0141093474426807e-05,
|
| 661 |
+
"loss": 1.5833,
|
| 662 |
+
"step": 1030
|
| 663 |
+
},
|
| 664 |
+
{
|
| 665 |
+
"epoch": 4.12,
|
| 666 |
+
"learning_rate": 9.700176366843034e-06,
|
| 667 |
+
"loss": 1.5865,
|
| 668 |
+
"step": 1040
|
| 669 |
+
},
|
| 670 |
+
{
|
| 671 |
+
"epoch": 4.16,
|
| 672 |
+
"learning_rate": 9.259259259259259e-06,
|
| 673 |
+
"loss": 1.5846,
|
| 674 |
+
"step": 1050
|
| 675 |
+
},
|
| 676 |
+
{
|
| 677 |
+
"epoch": 4.2,
|
| 678 |
+
"learning_rate": 8.818342151675484e-06,
|
| 679 |
+
"loss": 1.5946,
|
| 680 |
+
"step": 1060
|
| 681 |
+
},
|
| 682 |
+
{
|
| 683 |
+
"epoch": 4.24,
|
| 684 |
+
"learning_rate": 8.377425044091711e-06,
|
| 685 |
+
"loss": 1.6151,
|
| 686 |
+
"step": 1070
|
| 687 |
+
},
|
| 688 |
+
{
|
| 689 |
+
"epoch": 4.28,
|
| 690 |
+
"learning_rate": 7.936507936507936e-06,
|
| 691 |
+
"loss": 1.5523,
|
| 692 |
+
"step": 1080
|
| 693 |
+
},
|
| 694 |
+
{
|
| 695 |
+
"epoch": 4.32,
|
| 696 |
+
"learning_rate": 7.495590828924162e-06,
|
| 697 |
+
"loss": 1.5852,
|
| 698 |
+
"step": 1090
|
| 699 |
+
},
|
| 700 |
+
{
|
| 701 |
+
"epoch": 4.36,
|
| 702 |
+
"learning_rate": 7.054673721340388e-06,
|
| 703 |
+
"loss": 1.5802,
|
| 704 |
+
"step": 1100
|
| 705 |
+
},
|
| 706 |
+
{
|
| 707 |
+
"epoch": 4.4,
|
| 708 |
+
"learning_rate": 6.613756613756614e-06,
|
| 709 |
+
"loss": 1.6024,
|
| 710 |
+
"step": 1110
|
| 711 |
+
},
|
| 712 |
+
{
|
| 713 |
+
"epoch": 4.44,
|
| 714 |
+
"learning_rate": 6.172839506172839e-06,
|
| 715 |
+
"loss": 1.5884,
|
| 716 |
+
"step": 1120
|
| 717 |
+
},
|
| 718 |
+
{
|
| 719 |
+
"epoch": 4.48,
|
| 720 |
+
"learning_rate": 5.731922398589065e-06,
|
| 721 |
+
"loss": 1.5436,
|
| 722 |
+
"step": 1130
|
| 723 |
+
},
|
| 724 |
+
{
|
| 725 |
+
"epoch": 4.51,
|
| 726 |
+
"learning_rate": 5.291005291005291e-06,
|
| 727 |
+
"loss": 1.5893,
|
| 728 |
+
"step": 1140
|
| 729 |
+
},
|
| 730 |
+
{
|
| 731 |
+
"epoch": 4.55,
|
| 732 |
+
"learning_rate": 4.850088183421517e-06,
|
| 733 |
+
"loss": 1.5899,
|
| 734 |
+
"step": 1150
|
| 735 |
+
},
|
| 736 |
+
{
|
| 737 |
+
"epoch": 4.59,
|
| 738 |
+
"learning_rate": 4.409171075837742e-06,
|
| 739 |
+
"loss": 1.5784,
|
| 740 |
+
"step": 1160
|
| 741 |
+
},
|
| 742 |
+
{
|
| 743 |
+
"epoch": 4.63,
|
| 744 |
+
"learning_rate": 3.968253968253968e-06,
|
| 745 |
+
"loss": 1.5623,
|
| 746 |
+
"step": 1170
|
| 747 |
+
},
|
| 748 |
+
{
|
| 749 |
+
"epoch": 4.67,
|
| 750 |
+
"learning_rate": 3.527336860670194e-06,
|
| 751 |
+
"loss": 1.586,
|
| 752 |
+
"step": 1180
|
| 753 |
+
},
|
| 754 |
+
{
|
| 755 |
+
"epoch": 4.71,
|
| 756 |
+
"learning_rate": 3.0864197530864196e-06,
|
| 757 |
+
"loss": 1.5778,
|
| 758 |
+
"step": 1190
|
| 759 |
+
},
|
| 760 |
+
{
|
| 761 |
+
"epoch": 4.75,
|
| 762 |
+
"learning_rate": 2.6455026455026455e-06,
|
| 763 |
+
"loss": 1.5501,
|
| 764 |
+
"step": 1200
|
| 765 |
+
},
|
| 766 |
+
{
|
| 767 |
+
"epoch": 4.79,
|
| 768 |
+
"learning_rate": 2.204585537918871e-06,
|
| 769 |
+
"loss": 1.6087,
|
| 770 |
+
"step": 1210
|
| 771 |
+
},
|
| 772 |
+
{
|
| 773 |
+
"epoch": 4.83,
|
| 774 |
+
"learning_rate": 1.763668430335097e-06,
|
| 775 |
+
"loss": 1.5764,
|
| 776 |
+
"step": 1220
|
| 777 |
+
},
|
| 778 |
+
{
|
| 779 |
+
"epoch": 4.87,
|
| 780 |
+
"learning_rate": 1.3227513227513228e-06,
|
| 781 |
+
"loss": 1.5991,
|
| 782 |
+
"step": 1230
|
| 783 |
+
},
|
| 784 |
+
{
|
| 785 |
+
"epoch": 4.91,
|
| 786 |
+
"learning_rate": 8.818342151675485e-07,
|
| 787 |
+
"loss": 1.5643,
|
| 788 |
+
"step": 1240
|
| 789 |
+
},
|
| 790 |
+
{
|
| 791 |
+
"epoch": 4.95,
|
| 792 |
+
"learning_rate": 4.4091710758377425e-07,
|
| 793 |
+
"loss": 1.5688,
|
| 794 |
+
"step": 1250
|
| 795 |
+
},
|
| 796 |
+
{
|
| 797 |
+
"epoch": 4.99,
|
| 798 |
+
"learning_rate": 0.0,
|
| 799 |
+
"loss": 1.5645,
|
| 800 |
+
"step": 1260
|
| 801 |
+
},
|
| 802 |
+
{
|
| 803 |
+
"epoch": 4.99,
|
| 804 |
+
"eval_accuracy": 0.4446921147952076,
|
| 805 |
+
"eval_loss": 1.4667332172393799,
|
| 806 |
+
"eval_runtime": 16.8572,
|
| 807 |
+
"eval_samples_per_second": 212.906,
|
| 808 |
+
"eval_steps_per_second": 6.703,
|
| 809 |
+
"step": 1260
|
| 810 |
},
|
| 811 |
{
|
| 812 |
+
"epoch": 4.99,
|
| 813 |
+
"step": 1260,
|
| 814 |
+
"total_flos": 3.424468651849728e+18,
|
| 815 |
+
"train_loss": 1.653014714377267,
|
| 816 |
+
"train_runtime": 2002.0643,
|
| 817 |
+
"train_samples_per_second": 80.662,
|
| 818 |
+
"train_steps_per_second": 0.629
|
| 819 |
}
|
| 820 |
],
|
| 821 |
+
"max_steps": 1260,
|
| 822 |
+
"num_train_epochs": 5,
|
| 823 |
+
"total_flos": 3.424468651849728e+18,
|
| 824 |
"trial_name": null,
|
| 825 |
"trial_params": null
|
| 826 |
}
|
training_args.bin
CHANGED
|
@@ -1,3 +1,3 @@
|
|
| 1 |
version https://git-lfs.github.com/spec/v1
|
| 2 |
-
oid sha256:
|
| 3 |
size 4408
|
|
|
|
| 1 |
version https://git-lfs.github.com/spec/v1
|
| 2 |
+
oid sha256:3229e35e3e55df4b186a6a6699879b95c04344282f339abe3fbb728404a6edca
|
| 3 |
size 4408
|