| { | |
| "best_metric": 1.0025324821472168, | |
| "best_model_checkpoint": "./vit-base-renovation/checkpoint-75", | |
| "epoch": 4.0, | |
| "eval_steps": 25, | |
| "global_step": 496, | |
| "is_hyper_param_search": false, | |
| "is_local_process_zero": true, | |
| "is_world_process_zero": true, | |
| "log_history": [ | |
| { | |
| "epoch": 0.08, | |
| "grad_norm": 4.653707504272461, | |
| "learning_rate": 0.00019596774193548388, | |
| "loss": 0.4901, | |
| "step": 10 | |
| }, | |
| { | |
| "epoch": 0.16, | |
| "grad_norm": 1.4146682024002075, | |
| "learning_rate": 0.00019193548387096775, | |
| "loss": 0.382, | |
| "step": 20 | |
| }, | |
| { | |
| "epoch": 0.2, | |
| "eval_accuracy": 0.6073059360730594, | |
| "eval_loss": 1.1102608442306519, | |
| "eval_runtime": 7.3313, | |
| "eval_samples_per_second": 29.872, | |
| "eval_steps_per_second": 3.819, | |
| "step": 25 | |
| }, | |
| { | |
| "epoch": 0.24, | |
| "grad_norm": 4.3104939460754395, | |
| "learning_rate": 0.00018790322580645164, | |
| "loss": 0.3537, | |
| "step": 30 | |
| }, | |
| { | |
| "epoch": 0.32, | |
| "grad_norm": 6.281633377075195, | |
| "learning_rate": 0.00018387096774193548, | |
| "loss": 0.5919, | |
| "step": 40 | |
| }, | |
| { | |
| "epoch": 0.4, | |
| "grad_norm": 1.1501330137252808, | |
| "learning_rate": 0.00017983870967741935, | |
| "loss": 0.5741, | |
| "step": 50 | |
| }, | |
| { | |
| "epoch": 0.4, | |
| "eval_accuracy": 0.6210045662100456, | |
| "eval_loss": 1.0627731084823608, | |
| "eval_runtime": 7.7376, | |
| "eval_samples_per_second": 28.303, | |
| "eval_steps_per_second": 3.619, | |
| "step": 50 | |
| }, | |
| { | |
| "epoch": 0.48, | |
| "grad_norm": 4.414377689361572, | |
| "learning_rate": 0.00017580645161290325, | |
| "loss": 0.4974, | |
| "step": 60 | |
| }, | |
| { | |
| "epoch": 0.56, | |
| "grad_norm": 2.614995241165161, | |
| "learning_rate": 0.00017177419354838711, | |
| "loss": 0.5589, | |
| "step": 70 | |
| }, | |
| { | |
| "epoch": 0.6, | |
| "eval_accuracy": 0.6666666666666666, | |
| "eval_loss": 1.0025324821472168, | |
| "eval_runtime": 8.1199, | |
| "eval_samples_per_second": 26.971, | |
| "eval_steps_per_second": 3.448, | |
| "step": 75 | |
| }, | |
| { | |
| "epoch": 0.65, | |
| "grad_norm": 2.475475549697876, | |
| "learning_rate": 0.00016774193548387098, | |
| "loss": 0.3703, | |
| "step": 80 | |
| }, | |
| { | |
| "epoch": 0.73, | |
| "grad_norm": 5.414321422576904, | |
| "learning_rate": 0.00016370967741935485, | |
| "loss": 0.4059, | |
| "step": 90 | |
| }, | |
| { | |
| "epoch": 0.81, | |
| "grad_norm": 2.9260778427124023, | |
| "learning_rate": 0.00015967741935483872, | |
| "loss": 0.4074, | |
| "step": 100 | |
| }, | |
| { | |
| "epoch": 0.81, | |
| "eval_accuracy": 0.6073059360730594, | |
| "eval_loss": 1.1324039697647095, | |
| "eval_runtime": 7.9656, | |
| "eval_samples_per_second": 27.493, | |
| "eval_steps_per_second": 3.515, | |
| "step": 100 | |
| }, | |
| { | |
| "epoch": 0.89, | |
| "grad_norm": 1.745337724685669, | |
| "learning_rate": 0.0001556451612903226, | |
| "loss": 0.2648, | |
| "step": 110 | |
| }, | |
| { | |
| "epoch": 0.97, | |
| "grad_norm": 0.7676700353622437, | |
| "learning_rate": 0.00015161290322580646, | |
| "loss": 0.3581, | |
| "step": 120 | |
| }, | |
| { | |
| "epoch": 1.01, | |
| "eval_accuracy": 0.6438356164383562, | |
| "eval_loss": 1.1934621334075928, | |
| "eval_runtime": 7.3834, | |
| "eval_samples_per_second": 29.661, | |
| "eval_steps_per_second": 3.792, | |
| "step": 125 | |
| }, | |
| { | |
| "epoch": 1.05, | |
| "grad_norm": 0.40987706184387207, | |
| "learning_rate": 0.00014758064516129032, | |
| "loss": 0.1496, | |
| "step": 130 | |
| }, | |
| { | |
| "epoch": 1.13, | |
| "grad_norm": 0.4426233768463135, | |
| "learning_rate": 0.00014354838709677422, | |
| "loss": 0.1734, | |
| "step": 140 | |
| }, | |
| { | |
| "epoch": 1.21, | |
| "grad_norm": 4.183903694152832, | |
| "learning_rate": 0.0001395161290322581, | |
| "loss": 0.2618, | |
| "step": 150 | |
| }, | |
| { | |
| "epoch": 1.21, | |
| "eval_accuracy": 0.502283105022831, | |
| "eval_loss": 1.8300055265426636, | |
| "eval_runtime": 8.5064, | |
| "eval_samples_per_second": 25.745, | |
| "eval_steps_per_second": 3.292, | |
| "step": 150 | |
| }, | |
| { | |
| "epoch": 1.29, | |
| "grad_norm": 5.4273681640625, | |
| "learning_rate": 0.00013548387096774193, | |
| "loss": 0.4102, | |
| "step": 160 | |
| }, | |
| { | |
| "epoch": 1.37, | |
| "grad_norm": 0.8633623123168945, | |
| "learning_rate": 0.0001314516129032258, | |
| "loss": 0.1299, | |
| "step": 170 | |
| }, | |
| { | |
| "epoch": 1.41, | |
| "eval_accuracy": 0.6301369863013698, | |
| "eval_loss": 1.2576818466186523, | |
| "eval_runtime": 8.1086, | |
| "eval_samples_per_second": 27.008, | |
| "eval_steps_per_second": 3.453, | |
| "step": 175 | |
| }, | |
| { | |
| "epoch": 1.45, | |
| "grad_norm": 4.694212436676025, | |
| "learning_rate": 0.0001274193548387097, | |
| "loss": 0.2189, | |
| "step": 180 | |
| }, | |
| { | |
| "epoch": 1.53, | |
| "grad_norm": 3.287609338760376, | |
| "learning_rate": 0.00012338709677419356, | |
| "loss": 0.1625, | |
| "step": 190 | |
| }, | |
| { | |
| "epoch": 1.61, | |
| "grad_norm": 3.42903470993042, | |
| "learning_rate": 0.00011935483870967743, | |
| "loss": 0.2562, | |
| "step": 200 | |
| }, | |
| { | |
| "epoch": 1.61, | |
| "eval_accuracy": 0.6894977168949772, | |
| "eval_loss": 1.092441439628601, | |
| "eval_runtime": 8.405, | |
| "eval_samples_per_second": 26.056, | |
| "eval_steps_per_second": 3.331, | |
| "step": 200 | |
| }, | |
| { | |
| "epoch": 1.69, | |
| "grad_norm": 3.9116501808166504, | |
| "learning_rate": 0.00011532258064516131, | |
| "loss": 0.3386, | |
| "step": 210 | |
| }, | |
| { | |
| "epoch": 1.77, | |
| "grad_norm": 1.250923752784729, | |
| "learning_rate": 0.00011129032258064515, | |
| "loss": 0.2573, | |
| "step": 220 | |
| }, | |
| { | |
| "epoch": 1.81, | |
| "eval_accuracy": 0.684931506849315, | |
| "eval_loss": 1.1285330057144165, | |
| "eval_runtime": 8.1457, | |
| "eval_samples_per_second": 26.885, | |
| "eval_steps_per_second": 3.437, | |
| "step": 225 | |
| }, | |
| { | |
| "epoch": 1.85, | |
| "grad_norm": 0.8015751242637634, | |
| "learning_rate": 0.00010725806451612903, | |
| "loss": 0.1359, | |
| "step": 230 | |
| }, | |
| { | |
| "epoch": 1.94, | |
| "grad_norm": 3.1585347652435303, | |
| "learning_rate": 0.0001032258064516129, | |
| "loss": 0.2325, | |
| "step": 240 | |
| }, | |
| { | |
| "epoch": 2.02, | |
| "grad_norm": 1.220850944519043, | |
| "learning_rate": 9.919354838709678e-05, | |
| "loss": 0.2471, | |
| "step": 250 | |
| }, | |
| { | |
| "epoch": 2.02, | |
| "eval_accuracy": 0.6255707762557078, | |
| "eval_loss": 1.3386634588241577, | |
| "eval_runtime": 8.4514, | |
| "eval_samples_per_second": 25.913, | |
| "eval_steps_per_second": 3.313, | |
| "step": 250 | |
| }, | |
| { | |
| "epoch": 2.1, | |
| "grad_norm": 0.14146128296852112, | |
| "learning_rate": 9.516129032258065e-05, | |
| "loss": 0.0773, | |
| "step": 260 | |
| }, | |
| { | |
| "epoch": 2.18, | |
| "grad_norm": 0.13016600906848907, | |
| "learning_rate": 9.112903225806452e-05, | |
| "loss": 0.0618, | |
| "step": 270 | |
| }, | |
| { | |
| "epoch": 2.22, | |
| "eval_accuracy": 0.6666666666666666, | |
| "eval_loss": 1.2246029376983643, | |
| "eval_runtime": 7.483, | |
| "eval_samples_per_second": 29.267, | |
| "eval_steps_per_second": 3.742, | |
| "step": 275 | |
| }, | |
| { | |
| "epoch": 2.26, | |
| "grad_norm": 0.1758795827627182, | |
| "learning_rate": 8.709677419354839e-05, | |
| "loss": 0.108, | |
| "step": 280 | |
| }, | |
| { | |
| "epoch": 2.34, | |
| "grad_norm": 1.5025098323822021, | |
| "learning_rate": 8.306451612903227e-05, | |
| "loss": 0.0662, | |
| "step": 290 | |
| }, | |
| { | |
| "epoch": 2.42, | |
| "grad_norm": 8.820311546325684, | |
| "learning_rate": 7.903225806451613e-05, | |
| "loss": 0.0658, | |
| "step": 300 | |
| }, | |
| { | |
| "epoch": 2.42, | |
| "eval_accuracy": 0.634703196347032, | |
| "eval_loss": 1.4131733179092407, | |
| "eval_runtime": 8.0515, | |
| "eval_samples_per_second": 27.2, | |
| "eval_steps_per_second": 3.478, | |
| "step": 300 | |
| }, | |
| { | |
| "epoch": 2.5, | |
| "grad_norm": 0.2889515161514282, | |
| "learning_rate": 7.500000000000001e-05, | |
| "loss": 0.1059, | |
| "step": 310 | |
| }, | |
| { | |
| "epoch": 2.58, | |
| "grad_norm": 0.25573813915252686, | |
| "learning_rate": 7.096774193548388e-05, | |
| "loss": 0.0592, | |
| "step": 320 | |
| }, | |
| { | |
| "epoch": 2.62, | |
| "eval_accuracy": 0.6529680365296804, | |
| "eval_loss": 1.4326461553573608, | |
| "eval_runtime": 8.3973, | |
| "eval_samples_per_second": 26.08, | |
| "eval_steps_per_second": 3.334, | |
| "step": 325 | |
| }, | |
| { | |
| "epoch": 2.66, | |
| "grad_norm": 4.445276260375977, | |
| "learning_rate": 6.693548387096774e-05, | |
| "loss": 0.0717, | |
| "step": 330 | |
| }, | |
| { | |
| "epoch": 2.74, | |
| "grad_norm": 0.5493516325950623, | |
| "learning_rate": 6.290322580645161e-05, | |
| "loss": 0.0981, | |
| "step": 340 | |
| }, | |
| { | |
| "epoch": 2.82, | |
| "grad_norm": 3.7721645832061768, | |
| "learning_rate": 5.887096774193549e-05, | |
| "loss": 0.0464, | |
| "step": 350 | |
| }, | |
| { | |
| "epoch": 2.82, | |
| "eval_accuracy": 0.684931506849315, | |
| "eval_loss": 1.2483667135238647, | |
| "eval_runtime": 7.3404, | |
| "eval_samples_per_second": 29.835, | |
| "eval_steps_per_second": 3.815, | |
| "step": 350 | |
| }, | |
| { | |
| "epoch": 2.9, | |
| "grad_norm": 0.07467525452375412, | |
| "learning_rate": 5.4838709677419355e-05, | |
| "loss": 0.0249, | |
| "step": 360 | |
| }, | |
| { | |
| "epoch": 2.98, | |
| "grad_norm": 1.5583500862121582, | |
| "learning_rate": 5.080645161290323e-05, | |
| "loss": 0.0567, | |
| "step": 370 | |
| }, | |
| { | |
| "epoch": 3.02, | |
| "eval_accuracy": 0.634703196347032, | |
| "eval_loss": 1.5350439548492432, | |
| "eval_runtime": 7.9868, | |
| "eval_samples_per_second": 27.42, | |
| "eval_steps_per_second": 3.506, | |
| "step": 375 | |
| }, | |
| { | |
| "epoch": 3.06, | |
| "grad_norm": 0.06753262132406235, | |
| "learning_rate": 4.67741935483871e-05, | |
| "loss": 0.0185, | |
| "step": 380 | |
| }, | |
| { | |
| "epoch": 3.15, | |
| "grad_norm": 0.0861305296421051, | |
| "learning_rate": 4.2741935483870973e-05, | |
| "loss": 0.0179, | |
| "step": 390 | |
| }, | |
| { | |
| "epoch": 3.23, | |
| "grad_norm": 0.07843153923749924, | |
| "learning_rate": 3.870967741935484e-05, | |
| "loss": 0.0269, | |
| "step": 400 | |
| }, | |
| { | |
| "epoch": 3.23, | |
| "eval_accuracy": 0.6666666666666666, | |
| "eval_loss": 1.479716181755066, | |
| "eval_runtime": 10.862, | |
| "eval_samples_per_second": 20.162, | |
| "eval_steps_per_second": 2.578, | |
| "step": 400 | |
| }, | |
| { | |
| "epoch": 3.31, | |
| "grad_norm": 0.5874373912811279, | |
| "learning_rate": 3.467741935483872e-05, | |
| "loss": 0.0181, | |
| "step": 410 | |
| }, | |
| { | |
| "epoch": 3.39, | |
| "grad_norm": 0.15166939795017242, | |
| "learning_rate": 3.0645161290322585e-05, | |
| "loss": 0.0239, | |
| "step": 420 | |
| }, | |
| { | |
| "epoch": 3.43, | |
| "eval_accuracy": 0.6529680365296804, | |
| "eval_loss": 1.444357991218567, | |
| "eval_runtime": 8.4171, | |
| "eval_samples_per_second": 26.019, | |
| "eval_steps_per_second": 3.327, | |
| "step": 425 | |
| }, | |
| { | |
| "epoch": 3.47, | |
| "grad_norm": 0.2967914938926697, | |
| "learning_rate": 2.661290322580645e-05, | |
| "loss": 0.0217, | |
| "step": 430 | |
| }, | |
| { | |
| "epoch": 3.55, | |
| "grad_norm": 0.084845632314682, | |
| "learning_rate": 2.258064516129032e-05, | |
| "loss": 0.021, | |
| "step": 440 | |
| }, | |
| { | |
| "epoch": 3.63, | |
| "grad_norm": 0.06501659750938416, | |
| "learning_rate": 1.8548387096774193e-05, | |
| "loss": 0.0184, | |
| "step": 450 | |
| }, | |
| { | |
| "epoch": 3.63, | |
| "eval_accuracy": 0.6575342465753424, | |
| "eval_loss": 1.4473841190338135, | |
| "eval_runtime": 8.0514, | |
| "eval_samples_per_second": 27.2, | |
| "eval_steps_per_second": 3.478, | |
| "step": 450 | |
| }, | |
| { | |
| "epoch": 3.71, | |
| "grad_norm": 0.09078585356473923, | |
| "learning_rate": 1.4516129032258066e-05, | |
| "loss": 0.0162, | |
| "step": 460 | |
| }, | |
| { | |
| "epoch": 3.79, | |
| "grad_norm": 0.06605294346809387, | |
| "learning_rate": 1.0483870967741936e-05, | |
| "loss": 0.0286, | |
| "step": 470 | |
| }, | |
| { | |
| "epoch": 3.83, | |
| "eval_accuracy": 0.6666666666666666, | |
| "eval_loss": 1.4620742797851562, | |
| "eval_runtime": 8.3261, | |
| "eval_samples_per_second": 26.303, | |
| "eval_steps_per_second": 3.363, | |
| "step": 475 | |
| }, | |
| { | |
| "epoch": 3.87, | |
| "grad_norm": 0.05681619793176651, | |
| "learning_rate": 6.451612903225806e-06, | |
| "loss": 0.0162, | |
| "step": 480 | |
| }, | |
| { | |
| "epoch": 3.95, | |
| "grad_norm": 0.07769191265106201, | |
| "learning_rate": 2.4193548387096776e-06, | |
| "loss": 0.0261, | |
| "step": 490 | |
| }, | |
| { | |
| "epoch": 4.0, | |
| "step": 496, | |
| "total_flos": 6.10974224738132e+17, | |
| "train_loss": 0.18820726821920084, | |
| "train_runtime": 922.3598, | |
| "train_samples_per_second": 8.548, | |
| "train_steps_per_second": 0.538 | |
| } | |
| ], | |
| "logging_steps": 10, | |
| "max_steps": 496, | |
| "num_input_tokens_seen": 0, | |
| "num_train_epochs": 4, | |
| "save_steps": 25, | |
| "total_flos": 6.10974224738132e+17, | |
| "train_batch_size": 16, | |
| "trial_name": null, | |
| "trial_params": null | |
| } | |