| { | |
| "best_metric": 1.2383606433868408, | |
| "best_model_checkpoint": "./vit-base-renovation2/checkpoint-25", | |
| "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": 2.575977325439453, | |
| "learning_rate": 0.00019596774193548388, | |
| "loss": 0.3595, | |
| "step": 10 | |
| }, | |
| { | |
| "epoch": 0.16, | |
| "grad_norm": 0.9806227684020996, | |
| "learning_rate": 0.00019193548387096775, | |
| "loss": 0.273, | |
| "step": 20 | |
| }, | |
| { | |
| "epoch": 0.2, | |
| "eval_accuracy": 0.6027397260273972, | |
| "eval_loss": 1.2383606433868408, | |
| "eval_runtime": 7.9057, | |
| "eval_samples_per_second": 27.702, | |
| "eval_steps_per_second": 3.542, | |
| "step": 25 | |
| }, | |
| { | |
| "epoch": 0.24, | |
| "grad_norm": 2.519897699356079, | |
| "learning_rate": 0.00018790322580645164, | |
| "loss": 0.3335, | |
| "step": 30 | |
| }, | |
| { | |
| "epoch": 0.32, | |
| "grad_norm": 4.2508063316345215, | |
| "learning_rate": 0.00018387096774193548, | |
| "loss": 0.3046, | |
| "step": 40 | |
| }, | |
| { | |
| "epoch": 0.4, | |
| "grad_norm": 1.2349441051483154, | |
| "learning_rate": 0.00017983870967741935, | |
| "loss": 0.5153, | |
| "step": 50 | |
| }, | |
| { | |
| "epoch": 0.4, | |
| "eval_accuracy": 0.5844748858447488, | |
| "eval_loss": 1.40598464012146, | |
| "eval_runtime": 8.1033, | |
| "eval_samples_per_second": 27.026, | |
| "eval_steps_per_second": 3.455, | |
| "step": 50 | |
| }, | |
| { | |
| "epoch": 0.48, | |
| "grad_norm": 2.6730687618255615, | |
| "learning_rate": 0.00017580645161290325, | |
| "loss": 0.3137, | |
| "step": 60 | |
| }, | |
| { | |
| "epoch": 0.56, | |
| "grad_norm": 4.323659420013428, | |
| "learning_rate": 0.00017177419354838711, | |
| "loss": 0.2792, | |
| "step": 70 | |
| }, | |
| { | |
| "epoch": 0.6, | |
| "eval_accuracy": 0.593607305936073, | |
| "eval_loss": 1.30255126953125, | |
| "eval_runtime": 7.4143, | |
| "eval_samples_per_second": 29.538, | |
| "eval_steps_per_second": 3.776, | |
| "step": 75 | |
| }, | |
| { | |
| "epoch": 0.65, | |
| "grad_norm": 1.2822186946868896, | |
| "learning_rate": 0.00016774193548387098, | |
| "loss": 0.3751, | |
| "step": 80 | |
| }, | |
| { | |
| "epoch": 0.73, | |
| "grad_norm": 3.15390944480896, | |
| "learning_rate": 0.00016370967741935485, | |
| "loss": 0.5094, | |
| "step": 90 | |
| }, | |
| { | |
| "epoch": 0.81, | |
| "grad_norm": 3.2424371242523193, | |
| "learning_rate": 0.00016008064516129033, | |
| "loss": 0.5516, | |
| "step": 100 | |
| }, | |
| { | |
| "epoch": 0.81, | |
| "eval_accuracy": 0.6027397260273972, | |
| "eval_loss": 1.3998719453811646, | |
| "eval_runtime": 7.9973, | |
| "eval_samples_per_second": 27.384, | |
| "eval_steps_per_second": 3.501, | |
| "step": 100 | |
| }, | |
| { | |
| "epoch": 0.89, | |
| "grad_norm": 1.8185967206954956, | |
| "learning_rate": 0.0001560483870967742, | |
| "loss": 0.2684, | |
| "step": 110 | |
| }, | |
| { | |
| "epoch": 0.97, | |
| "grad_norm": 3.6117029190063477, | |
| "learning_rate": 0.0001520161290322581, | |
| "loss": 0.4247, | |
| "step": 120 | |
| }, | |
| { | |
| "epoch": 1.01, | |
| "eval_accuracy": 0.5981735159817352, | |
| "eval_loss": 1.2620913982391357, | |
| "eval_runtime": 8.1833, | |
| "eval_samples_per_second": 26.762, | |
| "eval_steps_per_second": 3.422, | |
| "step": 125 | |
| }, | |
| { | |
| "epoch": 1.05, | |
| "grad_norm": 0.7001488208770752, | |
| "learning_rate": 0.00014798387096774193, | |
| "loss": 0.1693, | |
| "step": 130 | |
| }, | |
| { | |
| "epoch": 1.13, | |
| "grad_norm": 3.214643716812134, | |
| "learning_rate": 0.0001439516129032258, | |
| "loss": 0.2085, | |
| "step": 140 | |
| }, | |
| { | |
| "epoch": 1.21, | |
| "grad_norm": 2.9545109272003174, | |
| "learning_rate": 0.00013991935483870967, | |
| "loss": 0.1556, | |
| "step": 150 | |
| }, | |
| { | |
| "epoch": 1.21, | |
| "eval_accuracy": 0.5570776255707762, | |
| "eval_loss": 1.5660632848739624, | |
| "eval_runtime": 8.1924, | |
| "eval_samples_per_second": 26.732, | |
| "eval_steps_per_second": 3.418, | |
| "step": 150 | |
| }, | |
| { | |
| "epoch": 1.29, | |
| "grad_norm": 3.702380418777466, | |
| "learning_rate": 0.00013588709677419357, | |
| "loss": 0.1262, | |
| "step": 160 | |
| }, | |
| { | |
| "epoch": 1.37, | |
| "grad_norm": 2.634195327758789, | |
| "learning_rate": 0.00013185483870967743, | |
| "loss": 0.1458, | |
| "step": 170 | |
| }, | |
| { | |
| "epoch": 1.41, | |
| "eval_accuracy": 0.634703196347032, | |
| "eval_loss": 1.3458657264709473, | |
| "eval_runtime": 7.2989, | |
| "eval_samples_per_second": 30.005, | |
| "eval_steps_per_second": 3.836, | |
| "step": 175 | |
| }, | |
| { | |
| "epoch": 1.45, | |
| "grad_norm": 1.4022108316421509, | |
| "learning_rate": 0.0001278225806451613, | |
| "loss": 0.0859, | |
| "step": 180 | |
| }, | |
| { | |
| "epoch": 1.53, | |
| "grad_norm": 6.311899185180664, | |
| "learning_rate": 0.00012379032258064514, | |
| "loss": 0.0868, | |
| "step": 190 | |
| }, | |
| { | |
| "epoch": 1.61, | |
| "grad_norm": 1.674466609954834, | |
| "learning_rate": 0.00011975806451612903, | |
| "loss": 0.1595, | |
| "step": 200 | |
| }, | |
| { | |
| "epoch": 1.61, | |
| "eval_accuracy": 0.5981735159817352, | |
| "eval_loss": 1.5277855396270752, | |
| "eval_runtime": 8.087, | |
| "eval_samples_per_second": 27.081, | |
| "eval_steps_per_second": 3.462, | |
| "step": 200 | |
| }, | |
| { | |
| "epoch": 1.69, | |
| "grad_norm": 1.2903810739517212, | |
| "learning_rate": 0.00011572580645161291, | |
| "loss": 0.1657, | |
| "step": 210 | |
| }, | |
| { | |
| "epoch": 1.77, | |
| "grad_norm": 1.2086151838302612, | |
| "learning_rate": 0.00011169354838709678, | |
| "loss": 0.1195, | |
| "step": 220 | |
| }, | |
| { | |
| "epoch": 1.81, | |
| "eval_accuracy": 0.6255707762557078, | |
| "eval_loss": 1.5303133726119995, | |
| "eval_runtime": 7.5329, | |
| "eval_samples_per_second": 29.072, | |
| "eval_steps_per_second": 3.717, | |
| "step": 225 | |
| }, | |
| { | |
| "epoch": 1.85, | |
| "grad_norm": 7.646343231201172, | |
| "learning_rate": 0.00010766129032258066, | |
| "loss": 0.1381, | |
| "step": 230 | |
| }, | |
| { | |
| "epoch": 1.94, | |
| "grad_norm": 5.495790481567383, | |
| "learning_rate": 0.00010362903225806453, | |
| "loss": 0.1761, | |
| "step": 240 | |
| }, | |
| { | |
| "epoch": 2.02, | |
| "grad_norm": 0.2532494366168976, | |
| "learning_rate": 9.95967741935484e-05, | |
| "loss": 0.1507, | |
| "step": 250 | |
| }, | |
| { | |
| "epoch": 2.02, | |
| "eval_accuracy": 0.5844748858447488, | |
| "eval_loss": 1.7701371908187866, | |
| "eval_runtime": 8.229, | |
| "eval_samples_per_second": 26.613, | |
| "eval_steps_per_second": 3.403, | |
| "step": 250 | |
| }, | |
| { | |
| "epoch": 2.1, | |
| "grad_norm": 2.001723527908325, | |
| "learning_rate": 9.556451612903226e-05, | |
| "loss": 0.141, | |
| "step": 260 | |
| }, | |
| { | |
| "epoch": 2.18, | |
| "grad_norm": 0.07203235477209091, | |
| "learning_rate": 9.153225806451613e-05, | |
| "loss": 0.023, | |
| "step": 270 | |
| }, | |
| { | |
| "epoch": 2.22, | |
| "eval_accuracy": 0.6301369863013698, | |
| "eval_loss": 1.535449743270874, | |
| "eval_runtime": 7.2739, | |
| "eval_samples_per_second": 30.108, | |
| "eval_steps_per_second": 3.849, | |
| "step": 275 | |
| }, | |
| { | |
| "epoch": 2.26, | |
| "grad_norm": 0.10443231463432312, | |
| "learning_rate": 8.75e-05, | |
| "loss": 0.0434, | |
| "step": 280 | |
| }, | |
| { | |
| "epoch": 2.34, | |
| "grad_norm": 1.9513520002365112, | |
| "learning_rate": 8.346774193548388e-05, | |
| "loss": 0.0221, | |
| "step": 290 | |
| }, | |
| { | |
| "epoch": 2.42, | |
| "grad_norm": 4.193704605102539, | |
| "learning_rate": 7.943548387096774e-05, | |
| "loss": 0.028, | |
| "step": 300 | |
| }, | |
| { | |
| "epoch": 2.42, | |
| "eval_accuracy": 0.6301369863013698, | |
| "eval_loss": 1.6535084247589111, | |
| "eval_runtime": 8.3156, | |
| "eval_samples_per_second": 26.336, | |
| "eval_steps_per_second": 3.367, | |
| "step": 300 | |
| }, | |
| { | |
| "epoch": 2.5, | |
| "grad_norm": 0.11177629232406616, | |
| "learning_rate": 7.540322580645162e-05, | |
| "loss": 0.0573, | |
| "step": 310 | |
| }, | |
| { | |
| "epoch": 2.58, | |
| "grad_norm": 0.0510709173977375, | |
| "learning_rate": 7.137096774193549e-05, | |
| "loss": 0.0698, | |
| "step": 320 | |
| }, | |
| { | |
| "epoch": 2.62, | |
| "eval_accuracy": 0.6438356164383562, | |
| "eval_loss": 1.6771588325500488, | |
| "eval_runtime": 7.5025, | |
| "eval_samples_per_second": 29.19, | |
| "eval_steps_per_second": 3.732, | |
| "step": 325 | |
| }, | |
| { | |
| "epoch": 2.66, | |
| "grad_norm": 4.947625160217285, | |
| "learning_rate": 6.733870967741935e-05, | |
| "loss": 0.0964, | |
| "step": 330 | |
| }, | |
| { | |
| "epoch": 2.74, | |
| "grad_norm": 0.07066596299409866, | |
| "learning_rate": 6.330645161290322e-05, | |
| "loss": 0.0792, | |
| "step": 340 | |
| }, | |
| { | |
| "epoch": 2.82, | |
| "grad_norm": 6.108673572540283, | |
| "learning_rate": 5.9274193548387104e-05, | |
| "loss": 0.0516, | |
| "step": 350 | |
| }, | |
| { | |
| "epoch": 2.82, | |
| "eval_accuracy": 0.680365296803653, | |
| "eval_loss": 1.4380197525024414, | |
| "eval_runtime": 8.9956, | |
| "eval_samples_per_second": 24.345, | |
| "eval_steps_per_second": 3.113, | |
| "step": 350 | |
| }, | |
| { | |
| "epoch": 2.9, | |
| "grad_norm": 0.04943337291479111, | |
| "learning_rate": 5.5241935483870966e-05, | |
| "loss": 0.024, | |
| "step": 360 | |
| }, | |
| { | |
| "epoch": 2.98, | |
| "grad_norm": 0.2530857026576996, | |
| "learning_rate": 5.120967741935484e-05, | |
| "loss": 0.0136, | |
| "step": 370 | |
| }, | |
| { | |
| "epoch": 3.02, | |
| "eval_accuracy": 0.6484018264840182, | |
| "eval_loss": 1.6561062335968018, | |
| "eval_runtime": 7.9464, | |
| "eval_samples_per_second": 27.56, | |
| "eval_steps_per_second": 3.524, | |
| "step": 375 | |
| }, | |
| { | |
| "epoch": 3.06, | |
| "grad_norm": 0.07669265568256378, | |
| "learning_rate": 4.7177419354838716e-05, | |
| "loss": 0.0114, | |
| "step": 380 | |
| }, | |
| { | |
| "epoch": 3.15, | |
| "grad_norm": 0.04155677929520607, | |
| "learning_rate": 4.3145161290322584e-05, | |
| "loss": 0.0109, | |
| "step": 390 | |
| }, | |
| { | |
| "epoch": 3.23, | |
| "grad_norm": 0.8200215101242065, | |
| "learning_rate": 3.911290322580645e-05, | |
| "loss": 0.0325, | |
| "step": 400 | |
| }, | |
| { | |
| "epoch": 3.23, | |
| "eval_accuracy": 0.6621004566210046, | |
| "eval_loss": 1.6027804613113403, | |
| "eval_runtime": 8.0216, | |
| "eval_samples_per_second": 27.301, | |
| "eval_steps_per_second": 3.491, | |
| "step": 400 | |
| }, | |
| { | |
| "epoch": 3.31, | |
| "grad_norm": 0.028287330642342567, | |
| "learning_rate": 3.508064516129033e-05, | |
| "loss": 0.0085, | |
| "step": 410 | |
| }, | |
| { | |
| "epoch": 3.39, | |
| "grad_norm": 0.05001891031861305, | |
| "learning_rate": 3.1048387096774195e-05, | |
| "loss": 0.0149, | |
| "step": 420 | |
| }, | |
| { | |
| "epoch": 3.43, | |
| "eval_accuracy": 0.6621004566210046, | |
| "eval_loss": 1.6261166334152222, | |
| "eval_runtime": 7.3858, | |
| "eval_samples_per_second": 29.652, | |
| "eval_steps_per_second": 3.791, | |
| "step": 425 | |
| }, | |
| { | |
| "epoch": 3.47, | |
| "grad_norm": 0.07340113818645477, | |
| "learning_rate": 2.7016129032258064e-05, | |
| "loss": 0.0104, | |
| "step": 430 | |
| }, | |
| { | |
| "epoch": 3.55, | |
| "grad_norm": 0.04165619611740112, | |
| "learning_rate": 2.2983870967741935e-05, | |
| "loss": 0.0119, | |
| "step": 440 | |
| }, | |
| { | |
| "epoch": 3.63, | |
| "grad_norm": 0.032720282673835754, | |
| "learning_rate": 1.8951612903225807e-05, | |
| "loss": 0.0082, | |
| "step": 450 | |
| }, | |
| { | |
| "epoch": 3.63, | |
| "eval_accuracy": 0.6621004566210046, | |
| "eval_loss": 1.6614654064178467, | |
| "eval_runtime": 7.1739, | |
| "eval_samples_per_second": 30.527, | |
| "eval_steps_per_second": 3.903, | |
| "step": 450 | |
| }, | |
| { | |
| "epoch": 3.71, | |
| "grad_norm": 0.07252111285924911, | |
| "learning_rate": 1.4919354838709679e-05, | |
| "loss": 0.0079, | |
| "step": 460 | |
| }, | |
| { | |
| "epoch": 3.79, | |
| "grad_norm": 0.029246076941490173, | |
| "learning_rate": 1.0887096774193549e-05, | |
| "loss": 0.0093, | |
| "step": 470 | |
| }, | |
| { | |
| "epoch": 3.83, | |
| "eval_accuracy": 0.6529680365296804, | |
| "eval_loss": 1.6878423690795898, | |
| "eval_runtime": 7.9237, | |
| "eval_samples_per_second": 27.639, | |
| "eval_steps_per_second": 3.534, | |
| "step": 475 | |
| }, | |
| { | |
| "epoch": 3.87, | |
| "grad_norm": 0.029909249395132065, | |
| "learning_rate": 6.854838709677419e-06, | |
| "loss": 0.0087, | |
| "step": 480 | |
| }, | |
| { | |
| "epoch": 3.95, | |
| "grad_norm": 0.042575616389513016, | |
| "learning_rate": 2.82258064516129e-06, | |
| "loss": 0.0293, | |
| "step": 490 | |
| }, | |
| { | |
| "epoch": 4.0, | |
| "step": 496, | |
| "total_flos": 6.10974224738132e+17, | |
| "train_loss": 0.14545002643529686, | |
| "train_runtime": 906.411, | |
| "train_samples_per_second": 8.698, | |
| "train_steps_per_second": 0.547 | |
| } | |
| ], | |
| "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 | |
| } | |