Prot10/vit-base-patch16-224-for-pre_evaluation
023
1{2 "best_metric": 0.39285714285714285,3 "best_model_checkpoint": "vit-base-patch16-224-for-pre_evaluation/checkpoint-308",4 "epoch": 29.53846153846154,5 "eval_steps": 500,6 "global_step": 480,7 "is_hyper_param_search": false,8 "is_local_process_zero": true,9 "is_world_process_zero": true,10 "log_history": [11 {12 "epoch": 0.62,13 "learning_rate": 1.0416666666666668e-05,14 "loss": 1.5774,15 "step": 1016 },17 {18 "epoch": 0.98,19 "eval_accuracy": 0.3021978021978022,20 "eval_loss": 1.510853886604309,21 "eval_runtime": 6.2821,22 "eval_samples_per_second": 57.943,23 "eval_steps_per_second": 1.91,24 "step": 1625 },26 {27 "epoch": 1.23,28 "learning_rate": 2.0833333333333336e-05,29 "loss": 1.5237,30 "step": 2031 },32 {33 "epoch": 1.85,34 "learning_rate": 3.125e-05,35 "loss": 1.4794,36 "step": 3037 },38 {39 "epoch": 1.97,40 "eval_accuracy": 0.3241758241758242,41 "eval_loss": 1.494186282157898,42 "eval_runtime": 6.3421,43 "eval_samples_per_second": 57.395,44 "eval_steps_per_second": 1.892,45 "step": 3246 },47 {48 "epoch": 2.46,49 "learning_rate": 4.166666666666667e-05,50 "loss": 1.4536,51 "step": 4052 },53 {54 "epoch": 2.95,55 "eval_accuracy": 0.31868131868131866,56 "eval_loss": 1.4943327903747559,57 "eval_runtime": 5.7214,58 "eval_samples_per_second": 63.621,59 "eval_steps_per_second": 2.097,60 "step": 4861 },62 {63 "epoch": 3.08,64 "learning_rate": 4.976851851851852e-05,65 "loss": 1.4643,66 "step": 5067 },68 {69 "epoch": 3.69,70 "learning_rate": 4.8611111111111115e-05,71 "loss": 1.421,72 "step": 6073 },74 {75 "epoch": 4.0,76 "eval_accuracy": 0.34065934065934067,77 "eval_loss": 1.4246565103530884,78 "eval_runtime": 5.9355,79 "eval_samples_per_second": 61.326,80 "eval_steps_per_second": 2.022,81 "step": 6582 },83 {84 "epoch": 4.31,85 "learning_rate": 4.745370370370371e-05,86 "loss": 1.4268,87 "step": 7088 },89 {90 "epoch": 4.92,91 "learning_rate": 4.62962962962963e-05,92 "loss": 1.3882,93 "step": 8094 },95 {96 "epoch": 4.98,97 "eval_accuracy": 0.34615384615384615,98 "eval_loss": 1.4944226741790771,99 "eval_runtime": 6.3968,100 "eval_samples_per_second": 56.903,101 "eval_steps_per_second": 1.876,102 "step": 81103 },104 {105 "epoch": 5.54,106 "learning_rate": 4.5138888888888894e-05,107 "loss": 1.3579,108 "step": 90109 },110 {111 "epoch": 5.97,112 "eval_accuracy": 0.35714285714285715,113 "eval_loss": 1.4180346727371216,114 "eval_runtime": 6.5996,115 "eval_samples_per_second": 55.154,116 "eval_steps_per_second": 1.818,117 "step": 97118 },119 {120 "epoch": 6.15,121 "learning_rate": 4.3981481481481486e-05,122 "loss": 1.3075,123 "step": 100124 },125 {126 "epoch": 6.77,127 "learning_rate": 4.282407407407408e-05,128 "loss": 1.2838,129 "step": 110130 },131 {132 "epoch": 6.95,133 "eval_accuracy": 0.36813186813186816,134 "eval_loss": 1.4692732095718384,135 "eval_runtime": 5.8126,136 "eval_samples_per_second": 62.623,137 "eval_steps_per_second": 2.064,138 "step": 113139 },140 {141 "epoch": 7.38,142 "learning_rate": 4.166666666666667e-05,143 "loss": 1.2877,144 "step": 120145 },146 {147 "epoch": 8.0,148 "learning_rate": 4.0509259259259265e-05,149 "loss": 1.2695,150 "step": 130151 },152 {153 "epoch": 8.0,154 "eval_accuracy": 0.3434065934065934,155 "eval_loss": 1.4359365701675415,156 "eval_runtime": 5.7907,157 "eval_samples_per_second": 62.859,158 "eval_steps_per_second": 2.072,159 "step": 130160 },161 {162 "epoch": 8.62,163 "learning_rate": 3.935185185185186e-05,164 "loss": 1.2016,165 "step": 140166 },167 {168 "epoch": 8.98,169 "eval_accuracy": 0.3598901098901099,170 "eval_loss": 1.4656463861465454,171 "eval_runtime": 6.0359,172 "eval_samples_per_second": 60.306,173 "eval_steps_per_second": 1.988,174 "step": 146175 },176 {177 "epoch": 9.23,178 "learning_rate": 3.8194444444444444e-05,179 "loss": 1.2361,180 "step": 150181 },182 {183 "epoch": 9.85,184 "learning_rate": 3.7037037037037037e-05,185 "loss": 1.2087,186 "step": 160187 },188 {189 "epoch": 9.97,190 "eval_accuracy": 0.33791208791208793,191 "eval_loss": 1.4549881219863892,192 "eval_runtime": 6.5216,193 "eval_samples_per_second": 55.814,194 "eval_steps_per_second": 1.84,195 "step": 162196 },197 {198 "epoch": 10.46,199 "learning_rate": 3.587962962962963e-05,200 "loss": 1.206,201 "step": 170202 },203 {204 "epoch": 10.95,205 "eval_accuracy": 0.3516483516483517,206 "eval_loss": 1.5055769681930542,207 "eval_runtime": 5.8519,208 "eval_samples_per_second": 62.202,209 "eval_steps_per_second": 2.051,210 "step": 178211 },212 {213 "epoch": 11.08,214 "learning_rate": 3.472222222222222e-05,215 "loss": 1.1296,216 "step": 180217 },218 {219 "epoch": 11.69,220 "learning_rate": 3.3564814814814815e-05,221 "loss": 1.1236,222 "step": 190223 },224 {225 "epoch": 12.0,226 "eval_accuracy": 0.3434065934065934,227 "eval_loss": 1.5003132820129395,228 "eval_runtime": 6.3872,229 "eval_samples_per_second": 56.989,230 "eval_steps_per_second": 1.879,231 "step": 195232 },233 {234 "epoch": 12.31,235 "learning_rate": 3.240740740740741e-05,236 "loss": 1.0955,237 "step": 200238 },239 {240 "epoch": 12.92,241 "learning_rate": 3.125e-05,242 "loss": 1.0534,243 "step": 210244 },245 {246 "epoch": 12.98,247 "eval_accuracy": 0.3269230769230769,248 "eval_loss": 1.5192676782608032,249 "eval_runtime": 6.8957,250 "eval_samples_per_second": 52.786,251 "eval_steps_per_second": 1.74,252 "step": 211253 },254 {255 "epoch": 13.54,256 "learning_rate": 3.0092592592592593e-05,257 "loss": 1.0024,258 "step": 220259 },260 {261 "epoch": 13.97,262 "eval_accuracy": 0.36813186813186816,263 "eval_loss": 1.4890482425689697,264 "eval_runtime": 5.7451,265 "eval_samples_per_second": 63.358,266 "eval_steps_per_second": 2.089,267 "step": 227268 },269 {270 "epoch": 14.15,271 "learning_rate": 2.8935185185185186e-05,272 "loss": 0.9924,273 "step": 230274 },275 {276 "epoch": 14.77,277 "learning_rate": 2.777777777777778e-05,278 "loss": 0.9767,279 "step": 240280 },281 {282 "epoch": 14.95,283 "eval_accuracy": 0.3434065934065934,284 "eval_loss": 1.5628184080123901,285 "eval_runtime": 6.5373,286 "eval_samples_per_second": 55.68,287 "eval_steps_per_second": 1.836,288 "step": 243289 },290 {291 "epoch": 15.38,292 "learning_rate": 2.6620370370370372e-05,293 "loss": 0.9337,294 "step": 250295 },296 {297 "epoch": 16.0,298 "learning_rate": 2.5462962962962965e-05,299 "loss": 0.9201,300 "step": 260301 },302 {303 "epoch": 16.0,304 "eval_accuracy": 0.3516483516483517,305 "eval_loss": 1.6305893659591675,306 "eval_runtime": 6.6234,307 "eval_samples_per_second": 54.957,308 "eval_steps_per_second": 1.812,309 "step": 260310 },311 {312 "epoch": 16.62,313 "learning_rate": 2.4305555555555558e-05,314 "loss": 0.9136,315 "step": 270316 },317 {318 "epoch": 16.98,319 "eval_accuracy": 0.3626373626373626,320 "eval_loss": 1.5715110301971436,321 "eval_runtime": 5.8274,322 "eval_samples_per_second": 62.463,323 "eval_steps_per_second": 2.059,324 "step": 276325 },326 {327 "epoch": 17.23,328 "learning_rate": 2.314814814814815e-05,329 "loss": 0.8228,330 "step": 280331 },332 {333 "epoch": 17.85,334 "learning_rate": 2.1990740740740743e-05,335 "loss": 0.8566,336 "step": 290337 },338 {339 "epoch": 17.97,340 "eval_accuracy": 0.36538461538461536,341 "eval_loss": 1.5965826511383057,342 "eval_runtime": 5.8771,343 "eval_samples_per_second": 61.935,344 "eval_steps_per_second": 2.042,345 "step": 292346 },347 {348 "epoch": 18.46,349 "learning_rate": 2.0833333333333336e-05,350 "loss": 0.8273,351 "step": 300352 },353 {354 "epoch": 18.95,355 "eval_accuracy": 0.39285714285714285,356 "eval_loss": 1.604812502861023,357 "eval_runtime": 6.2505,358 "eval_samples_per_second": 58.235,359 "eval_steps_per_second": 1.92,360 "step": 308361 },362 {363 "epoch": 19.08,364 "learning_rate": 1.967592592592593e-05,365 "loss": 0.8217,366 "step": 310367 },368 {369 "epoch": 19.69,370 "learning_rate": 1.8518518518518518e-05,371 "loss": 0.7825,372 "step": 320373 },374 {375 "epoch": 20.0,376 "eval_accuracy": 0.38461538461538464,377 "eval_loss": 1.6174668073654175,378 "eval_runtime": 5.8325,379 "eval_samples_per_second": 62.409,380 "eval_steps_per_second": 2.057,381 "step": 325382 },383 {384 "epoch": 20.31,385 "learning_rate": 1.736111111111111e-05,386 "loss": 0.8128,387 "step": 330388 },389 {390 "epoch": 20.92,391 "learning_rate": 1.6203703703703704e-05,392 "loss": 0.736,393 "step": 340394 },395 {396 "epoch": 20.98,397 "eval_accuracy": 0.39285714285714285,398 "eval_loss": 1.652581810951233,399 "eval_runtime": 6.2966,400 "eval_samples_per_second": 57.809,401 "eval_steps_per_second": 1.906,402 "step": 341403 },404 {405 "epoch": 21.54,406 "learning_rate": 1.5046296296296297e-05,407 "loss": 0.7008,408 "step": 350409 },410 {411 "epoch": 21.97,412 "eval_accuracy": 0.37362637362637363,413 "eval_loss": 1.65627920627594,414 "eval_runtime": 6.0612,415 "eval_samples_per_second": 60.054,416 "eval_steps_per_second": 1.98,417 "step": 357418 },419 {420 "epoch": 22.15,421 "learning_rate": 1.388888888888889e-05,422 "loss": 0.7074,423 "step": 360424 },425 {426 "epoch": 22.77,427 "learning_rate": 1.2731481481481482e-05,428 "loss": 0.6714,429 "step": 370430 },431 {432 "epoch": 22.95,433 "eval_accuracy": 0.3901098901098901,434 "eval_loss": 1.7319426536560059,435 "eval_runtime": 6.3113,436 "eval_samples_per_second": 57.674,437 "eval_steps_per_second": 1.901,438 "step": 373439 },440 {441 "epoch": 23.38,442 "learning_rate": 1.1574074074074075e-05,443 "loss": 0.6723,444 "step": 380445 },446 {447 "epoch": 24.0,448 "learning_rate": 1.0416666666666668e-05,449 "loss": 0.7039,450 "step": 390451 },452 {453 "epoch": 24.0,454 "eval_accuracy": 0.39285714285714285,455 "eval_loss": 1.686637282371521,456 "eval_runtime": 5.7311,457 "eval_samples_per_second": 63.514,458 "eval_steps_per_second": 2.094,459 "step": 390460 },461 {462 "epoch": 24.62,463 "learning_rate": 9.259259259259259e-06,464 "loss": 0.628,465 "step": 400466 },467 {468 "epoch": 24.98,469 "eval_accuracy": 0.3791208791208791,470 "eval_loss": 1.7022507190704346,471 "eval_runtime": 5.7107,472 "eval_samples_per_second": 63.74,473 "eval_steps_per_second": 2.101,474 "step": 406475 },476 {477 "epoch": 25.23,478 "learning_rate": 8.101851851851852e-06,479 "loss": 0.6386,480 "step": 410481 },482 {483 "epoch": 25.85,484 "learning_rate": 6.944444444444445e-06,485 "loss": 0.6182,486 "step": 420487 },488 {489 "epoch": 25.97,490 "eval_accuracy": 0.3901098901098901,491 "eval_loss": 1.7301020622253418,492 "eval_runtime": 5.9062,493 "eval_samples_per_second": 61.63,494 "eval_steps_per_second": 2.032,495 "step": 422496 },497 {498 "epoch": 26.46,499 "learning_rate": 5.787037037037038e-06,500 "loss": 0.5957,501 "step": 430502 },503 {504 "epoch": 26.95,505 "eval_accuracy": 0.38461538461538464,506 "eval_loss": 1.7156624794006348,507 "eval_runtime": 6.0021,508 "eval_samples_per_second": 60.646,509 "eval_steps_per_second": 1.999,510 "step": 438511 },512 {513 "epoch": 27.08,514 "learning_rate": 4.6296296296296296e-06,515 "loss": 0.595,516 "step": 440517 },518 {519 "epoch": 27.69,520 "learning_rate": 3.4722222222222224e-06,521 "loss": 0.5973,522 "step": 450523 },524 {525 "epoch": 28.0,526 "eval_accuracy": 0.3708791208791209,527 "eval_loss": 1.7478266954421997,528 "eval_runtime": 6.4181,529 "eval_samples_per_second": 56.715,530 "eval_steps_per_second": 1.87,531 "step": 455532 },533 {534 "epoch": 28.31,535 "learning_rate": 2.3148148148148148e-06,536 "loss": 0.5767,537 "step": 460538 },539 {540 "epoch": 28.92,541 "learning_rate": 1.1574074074074074e-06,542 "loss": 0.5655,543 "step": 470544 },545 {546 "epoch": 28.98,547 "eval_accuracy": 0.37362637362637363,548 "eval_loss": 1.7376786470413208,549 "eval_runtime": 5.7918,550 "eval_samples_per_second": 62.847,551 "eval_steps_per_second": 2.072,552 "step": 471553 },554 {555 "epoch": 29.54,556 "learning_rate": 0.0,557 "loss": 0.5631,558 "step": 480559 },560 {561 "epoch": 29.54,562 "eval_accuracy": 0.37362637362637363,563 "eval_loss": 1.737407922744751,564 "eval_runtime": 6.3261,565 "eval_samples_per_second": 57.54,566 "eval_steps_per_second": 1.897,567 "step": 480568 },569 {570 "epoch": 29.54,571 "step": 480,572 "total_flos": 4.691568687003814e+18,573 "train_loss": 0.9943243801593781,574 "train_runtime": 2664.7299,575 "train_samples_per_second": 23.068,576 "train_steps_per_second": 0.18577 }578 ],579 "logging_steps": 10,580 "max_steps": 480,581 "num_train_epochs": 30,582 "save_steps": 500,583 "total_flos": 4.691568687003814e+18,584 "trial_name": null,585 "trial_params": null586}587 