| { | |
| "best_metric": 0.8272958397865295, | |
| "best_model_checkpoint": "./vit-base-renovation2/checkpoint-250", | |
| "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": 1.815894365310669, | |
| "learning_rate": 0.00019596774193548388, | |
| "loss": 1.5846, | |
| "step": 10 | |
| }, | |
| { | |
| "epoch": 0.16, | |
| "grad_norm": 2.0016415119171143, | |
| "learning_rate": 0.00019193548387096775, | |
| "loss": 1.359, | |
| "step": 20 | |
| }, | |
| { | |
| "epoch": 0.2, | |
| "eval_accuracy": 0.4657534246575342, | |
| "eval_loss": 1.2074170112609863, | |
| "eval_runtime": 7.6692, | |
| "eval_samples_per_second": 28.556, | |
| "eval_steps_per_second": 3.651, | |
| "step": 25 | |
| }, | |
| { | |
| "epoch": 0.24, | |
| "grad_norm": 1.6700384616851807, | |
| "learning_rate": 0.00018790322580645164, | |
| "loss": 1.182, | |
| "step": 30 | |
| }, | |
| { | |
| "epoch": 0.32, | |
| "grad_norm": 2.4473767280578613, | |
| "learning_rate": 0.00018387096774193548, | |
| "loss": 1.1559, | |
| "step": 40 | |
| }, | |
| { | |
| "epoch": 0.4, | |
| "grad_norm": 2.757627010345459, | |
| "learning_rate": 0.00017983870967741935, | |
| "loss": 1.1384, | |
| "step": 50 | |
| }, | |
| { | |
| "epoch": 0.4, | |
| "eval_accuracy": 0.5205479452054794, | |
| "eval_loss": 1.1213266849517822, | |
| "eval_runtime": 8.2359, | |
| "eval_samples_per_second": 26.591, | |
| "eval_steps_per_second": 3.4, | |
| "step": 50 | |
| }, | |
| { | |
| "epoch": 0.48, | |
| "grad_norm": 1.8923784494400024, | |
| "learning_rate": 0.00017580645161290325, | |
| "loss": 1.147, | |
| "step": 60 | |
| }, | |
| { | |
| "epoch": 0.56, | |
| "grad_norm": 2.0101828575134277, | |
| "learning_rate": 0.00017177419354838711, | |
| "loss": 1.0866, | |
| "step": 70 | |
| }, | |
| { | |
| "epoch": 0.6, | |
| "eval_accuracy": 0.6301369863013698, | |
| "eval_loss": 0.9746416807174683, | |
| "eval_runtime": 8.0785, | |
| "eval_samples_per_second": 27.109, | |
| "eval_steps_per_second": 3.466, | |
| "step": 75 | |
| }, | |
| { | |
| "epoch": 0.65, | |
| "grad_norm": 3.479001522064209, | |
| "learning_rate": 0.00016774193548387098, | |
| "loss": 0.875, | |
| "step": 80 | |
| }, | |
| { | |
| "epoch": 0.73, | |
| "grad_norm": 3.4011082649230957, | |
| "learning_rate": 0.00016370967741935485, | |
| "loss": 1.1928, | |
| "step": 90 | |
| }, | |
| { | |
| "epoch": 0.81, | |
| "grad_norm": 2.3756794929504395, | |
| "learning_rate": 0.00015967741935483872, | |
| "loss": 1.1787, | |
| "step": 100 | |
| }, | |
| { | |
| "epoch": 0.81, | |
| "eval_accuracy": 0.5662100456621004, | |
| "eval_loss": 1.0523391962051392, | |
| "eval_runtime": 9.1559, | |
| "eval_samples_per_second": 23.919, | |
| "eval_steps_per_second": 3.058, | |
| "step": 100 | |
| }, | |
| { | |
| "epoch": 0.89, | |
| "grad_norm": 1.5179095268249512, | |
| "learning_rate": 0.0001556451612903226, | |
| "loss": 0.9827, | |
| "step": 110 | |
| }, | |
| { | |
| "epoch": 0.97, | |
| "grad_norm": 2.3550424575805664, | |
| "learning_rate": 0.00015161290322580646, | |
| "loss": 0.9242, | |
| "step": 120 | |
| }, | |
| { | |
| "epoch": 1.01, | |
| "eval_accuracy": 0.6255707762557078, | |
| "eval_loss": 0.9543189406394958, | |
| "eval_runtime": 7.3462, | |
| "eval_samples_per_second": 29.812, | |
| "eval_steps_per_second": 3.812, | |
| "step": 125 | |
| }, | |
| { | |
| "epoch": 1.05, | |
| "grad_norm": 1.8633378744125366, | |
| "learning_rate": 0.00014758064516129032, | |
| "loss": 0.8227, | |
| "step": 130 | |
| }, | |
| { | |
| "epoch": 1.13, | |
| "grad_norm": 2.0424883365631104, | |
| "learning_rate": 0.00014354838709677422, | |
| "loss": 0.747, | |
| "step": 140 | |
| }, | |
| { | |
| "epoch": 1.21, | |
| "grad_norm": 2.2865993976593018, | |
| "learning_rate": 0.0001395161290322581, | |
| "loss": 0.7945, | |
| "step": 150 | |
| }, | |
| { | |
| "epoch": 1.21, | |
| "eval_accuracy": 0.6118721461187214, | |
| "eval_loss": 0.9199973940849304, | |
| "eval_runtime": 7.7884, | |
| "eval_samples_per_second": 28.119, | |
| "eval_steps_per_second": 3.595, | |
| "step": 150 | |
| }, | |
| { | |
| "epoch": 1.29, | |
| "grad_norm": 1.884401798248291, | |
| "learning_rate": 0.00013548387096774193, | |
| "loss": 0.9408, | |
| "step": 160 | |
| }, | |
| { | |
| "epoch": 1.37, | |
| "grad_norm": 2.085380792617798, | |
| "learning_rate": 0.0001314516129032258, | |
| "loss": 0.8379, | |
| "step": 170 | |
| }, | |
| { | |
| "epoch": 1.41, | |
| "eval_accuracy": 0.6712328767123288, | |
| "eval_loss": 0.8447285294532776, | |
| "eval_runtime": 7.8794, | |
| "eval_samples_per_second": 27.794, | |
| "eval_steps_per_second": 3.554, | |
| "step": 175 | |
| }, | |
| { | |
| "epoch": 1.45, | |
| "grad_norm": 2.1437325477600098, | |
| "learning_rate": 0.0001274193548387097, | |
| "loss": 0.8348, | |
| "step": 180 | |
| }, | |
| { | |
| "epoch": 1.53, | |
| "grad_norm": 1.6616462469100952, | |
| "learning_rate": 0.00012338709677419356, | |
| "loss": 0.6521, | |
| "step": 190 | |
| }, | |
| { | |
| "epoch": 1.61, | |
| "grad_norm": 2.3329973220825195, | |
| "learning_rate": 0.00011935483870967743, | |
| "loss": 0.7253, | |
| "step": 200 | |
| }, | |
| { | |
| "epoch": 1.61, | |
| "eval_accuracy": 0.6575342465753424, | |
| "eval_loss": 0.8642416596412659, | |
| "eval_runtime": 7.2392, | |
| "eval_samples_per_second": 30.252, | |
| "eval_steps_per_second": 3.868, | |
| "step": 200 | |
| }, | |
| { | |
| "epoch": 1.69, | |
| "grad_norm": 3.4008474349975586, | |
| "learning_rate": 0.00011532258064516131, | |
| "loss": 0.6683, | |
| "step": 210 | |
| }, | |
| { | |
| "epoch": 1.77, | |
| "grad_norm": 2.7154157161712646, | |
| "learning_rate": 0.00011129032258064515, | |
| "loss": 0.6344, | |
| "step": 220 | |
| }, | |
| { | |
| "epoch": 1.81, | |
| "eval_accuracy": 0.6438356164383562, | |
| "eval_loss": 0.8443467020988464, | |
| "eval_runtime": 7.7267, | |
| "eval_samples_per_second": 28.343, | |
| "eval_steps_per_second": 3.624, | |
| "step": 225 | |
| }, | |
| { | |
| "epoch": 1.85, | |
| "grad_norm": 2.2653236389160156, | |
| "learning_rate": 0.00010725806451612903, | |
| "loss": 0.768, | |
| "step": 230 | |
| }, | |
| { | |
| "epoch": 1.94, | |
| "grad_norm": 2.3153140544891357, | |
| "learning_rate": 0.0001032258064516129, | |
| "loss": 0.8194, | |
| "step": 240 | |
| }, | |
| { | |
| "epoch": 2.02, | |
| "grad_norm": 1.3728018999099731, | |
| "learning_rate": 9.919354838709678e-05, | |
| "loss": 0.6521, | |
| "step": 250 | |
| }, | |
| { | |
| "epoch": 2.02, | |
| "eval_accuracy": 0.6666666666666666, | |
| "eval_loss": 0.8272958397865295, | |
| "eval_runtime": 7.9935, | |
| "eval_samples_per_second": 27.397, | |
| "eval_steps_per_second": 3.503, | |
| "step": 250 | |
| }, | |
| { | |
| "epoch": 2.1, | |
| "grad_norm": 1.0636628866195679, | |
| "learning_rate": 9.516129032258065e-05, | |
| "loss": 0.3979, | |
| "step": 260 | |
| }, | |
| { | |
| "epoch": 2.18, | |
| "grad_norm": 1.3106963634490967, | |
| "learning_rate": 9.112903225806452e-05, | |
| "loss": 0.3627, | |
| "step": 270 | |
| }, | |
| { | |
| "epoch": 2.22, | |
| "eval_accuracy": 0.6712328767123288, | |
| "eval_loss": 0.865295946598053, | |
| "eval_runtime": 7.9627, | |
| "eval_samples_per_second": 27.503, | |
| "eval_steps_per_second": 3.516, | |
| "step": 275 | |
| }, | |
| { | |
| "epoch": 2.26, | |
| "grad_norm": 1.7618376016616821, | |
| "learning_rate": 8.709677419354839e-05, | |
| "loss": 0.3624, | |
| "step": 280 | |
| }, | |
| { | |
| "epoch": 2.34, | |
| "grad_norm": 1.3475382328033447, | |
| "learning_rate": 8.306451612903227e-05, | |
| "loss": 0.4034, | |
| "step": 290 | |
| }, | |
| { | |
| "epoch": 2.42, | |
| "grad_norm": 3.0443389415740967, | |
| "learning_rate": 7.903225806451613e-05, | |
| "loss": 0.2523, | |
| "step": 300 | |
| }, | |
| { | |
| "epoch": 2.42, | |
| "eval_accuracy": 0.6894977168949772, | |
| "eval_loss": 0.874794065952301, | |
| "eval_runtime": 7.2976, | |
| "eval_samples_per_second": 30.01, | |
| "eval_steps_per_second": 3.837, | |
| "step": 300 | |
| }, | |
| { | |
| "epoch": 2.5, | |
| "grad_norm": 2.961167812347412, | |
| "learning_rate": 7.500000000000001e-05, | |
| "loss": 0.3129, | |
| "step": 310 | |
| }, | |
| { | |
| "epoch": 2.58, | |
| "grad_norm": 1.5659066438674927, | |
| "learning_rate": 7.096774193548388e-05, | |
| "loss": 0.363, | |
| "step": 320 | |
| }, | |
| { | |
| "epoch": 2.62, | |
| "eval_accuracy": 0.684931506849315, | |
| "eval_loss": 0.8407436609268188, | |
| "eval_runtime": 7.6113, | |
| "eval_samples_per_second": 28.773, | |
| "eval_steps_per_second": 3.679, | |
| "step": 325 | |
| }, | |
| { | |
| "epoch": 2.66, | |
| "grad_norm": 1.1653140783309937, | |
| "learning_rate": 6.733870967741935e-05, | |
| "loss": 0.3924, | |
| "step": 330 | |
| }, | |
| { | |
| "epoch": 2.74, | |
| "grad_norm": 3.533656358718872, | |
| "learning_rate": 6.330645161290322e-05, | |
| "loss": 0.3786, | |
| "step": 340 | |
| }, | |
| { | |
| "epoch": 2.82, | |
| "grad_norm": 3.4866058826446533, | |
| "learning_rate": 5.9274193548387104e-05, | |
| "loss": 0.3433, | |
| "step": 350 | |
| }, | |
| { | |
| "epoch": 2.82, | |
| "eval_accuracy": 0.6484018264840182, | |
| "eval_loss": 0.969610333442688, | |
| "eval_runtime": 8.4636, | |
| "eval_samples_per_second": 25.875, | |
| "eval_steps_per_second": 3.308, | |
| "step": 350 | |
| }, | |
| { | |
| "epoch": 2.9, | |
| "grad_norm": 0.8572694659233093, | |
| "learning_rate": 5.5241935483870966e-05, | |
| "loss": 0.3991, | |
| "step": 360 | |
| }, | |
| { | |
| "epoch": 2.98, | |
| "grad_norm": 2.2343902587890625, | |
| "learning_rate": 5.120967741935484e-05, | |
| "loss": 0.2874, | |
| "step": 370 | |
| }, | |
| { | |
| "epoch": 3.02, | |
| "eval_accuracy": 0.680365296803653, | |
| "eval_loss": 0.9290387034416199, | |
| "eval_runtime": 7.9323, | |
| "eval_samples_per_second": 27.608, | |
| "eval_steps_per_second": 3.53, | |
| "step": 375 | |
| }, | |
| { | |
| "epoch": 3.06, | |
| "grad_norm": 0.40827465057373047, | |
| "learning_rate": 4.7177419354838716e-05, | |
| "loss": 0.1757, | |
| "step": 380 | |
| }, | |
| { | |
| "epoch": 3.15, | |
| "grad_norm": 0.28673678636550903, | |
| "learning_rate": 4.3145161290322584e-05, | |
| "loss": 0.122, | |
| "step": 390 | |
| }, | |
| { | |
| "epoch": 3.23, | |
| "grad_norm": 1.3551135063171387, | |
| "learning_rate": 3.911290322580645e-05, | |
| "loss": 0.1682, | |
| "step": 400 | |
| }, | |
| { | |
| "epoch": 3.23, | |
| "eval_accuracy": 0.6575342465753424, | |
| "eval_loss": 0.9713302254676819, | |
| "eval_runtime": 7.4128, | |
| "eval_samples_per_second": 29.543, | |
| "eval_steps_per_second": 3.777, | |
| "step": 400 | |
| }, | |
| { | |
| "epoch": 3.31, | |
| "grad_norm": 0.3541167676448822, | |
| "learning_rate": 3.508064516129033e-05, | |
| "loss": 0.1371, | |
| "step": 410 | |
| }, | |
| { | |
| "epoch": 3.39, | |
| "grad_norm": 1.2407793998718262, | |
| "learning_rate": 3.1048387096774195e-05, | |
| "loss": 0.1575, | |
| "step": 420 | |
| }, | |
| { | |
| "epoch": 3.43, | |
| "eval_accuracy": 0.680365296803653, | |
| "eval_loss": 0.9963440299034119, | |
| "eval_runtime": 7.3009, | |
| "eval_samples_per_second": 29.996, | |
| "eval_steps_per_second": 3.835, | |
| "step": 425 | |
| }, | |
| { | |
| "epoch": 3.47, | |
| "grad_norm": 2.0023341178894043, | |
| "learning_rate": 2.7016129032258064e-05, | |
| "loss": 0.1432, | |
| "step": 430 | |
| }, | |
| { | |
| "epoch": 3.55, | |
| "grad_norm": 0.3904053568840027, | |
| "learning_rate": 2.2983870967741935e-05, | |
| "loss": 0.1827, | |
| "step": 440 | |
| }, | |
| { | |
| "epoch": 3.63, | |
| "grad_norm": 0.42743560671806335, | |
| "learning_rate": 1.8951612903225807e-05, | |
| "loss": 0.0822, | |
| "step": 450 | |
| }, | |
| { | |
| "epoch": 3.63, | |
| "eval_accuracy": 0.7123287671232876, | |
| "eval_loss": 0.9472664594650269, | |
| "eval_runtime": 7.4585, | |
| "eval_samples_per_second": 29.362, | |
| "eval_steps_per_second": 3.754, | |
| "step": 450 | |
| }, | |
| { | |
| "epoch": 3.71, | |
| "grad_norm": 0.7994292974472046, | |
| "learning_rate": 1.4919354838709679e-05, | |
| "loss": 0.133, | |
| "step": 460 | |
| }, | |
| { | |
| "epoch": 3.79, | |
| "grad_norm": 0.4317198097705841, | |
| "learning_rate": 1.0887096774193549e-05, | |
| "loss": 0.1678, | |
| "step": 470 | |
| }, | |
| { | |
| "epoch": 3.83, | |
| "eval_accuracy": 0.7031963470319634, | |
| "eval_loss": 0.9787965416908264, | |
| "eval_runtime": 7.434, | |
| "eval_samples_per_second": 29.459, | |
| "eval_steps_per_second": 3.766, | |
| "step": 475 | |
| }, | |
| { | |
| "epoch": 3.87, | |
| "grad_norm": 0.2247464805841446, | |
| "learning_rate": 6.854838709677419e-06, | |
| "loss": 0.148, | |
| "step": 480 | |
| }, | |
| { | |
| "epoch": 3.95, | |
| "grad_norm": 0.28201037645339966, | |
| "learning_rate": 2.82258064516129e-06, | |
| "loss": 0.1406, | |
| "step": 490 | |
| }, | |
| { | |
| "epoch": 4.0, | |
| "step": 496, | |
| "total_flos": 6.10974224738132e+17, | |
| "train_loss": 0.601069396421794, | |
| "train_runtime": 693.9158, | |
| "train_samples_per_second": 11.362, | |
| "train_steps_per_second": 0.715 | |
| } | |
| ], | |
| "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 | |
| } | |