| { | |
| "best_metric": 0.7924718260765076, | |
| "best_model_checkpoint": "./vit-base-beans-demo-v5/checkpoint-200", | |
| "epoch": 4.0, | |
| "eval_steps": 100, | |
| "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": 2.0731146335601807, | |
| "learning_rate": 0.00019596774193548388, | |
| "loss": 1.5546, | |
| "step": 10 | |
| }, | |
| { | |
| "epoch": 0.16, | |
| "grad_norm": 1.9955084323883057, | |
| "learning_rate": 0.00019193548387096775, | |
| "loss": 1.3431, | |
| "step": 20 | |
| }, | |
| { | |
| "epoch": 0.24, | |
| "grad_norm": 1.4386166334152222, | |
| "learning_rate": 0.00018790322580645164, | |
| "loss": 1.1702, | |
| "step": 30 | |
| }, | |
| { | |
| "epoch": 0.32, | |
| "grad_norm": 2.1243135929107666, | |
| "learning_rate": 0.00018387096774193548, | |
| "loss": 1.1703, | |
| "step": 40 | |
| }, | |
| { | |
| "epoch": 0.4, | |
| "grad_norm": 1.4760087728500366, | |
| "learning_rate": 0.00017983870967741935, | |
| "loss": 1.0635, | |
| "step": 50 | |
| }, | |
| { | |
| "epoch": 0.48, | |
| "grad_norm": 1.3573176860809326, | |
| "learning_rate": 0.00017580645161290325, | |
| "loss": 1.0524, | |
| "step": 60 | |
| }, | |
| { | |
| "epoch": 0.56, | |
| "grad_norm": 3.1192474365234375, | |
| "learning_rate": 0.00017177419354838711, | |
| "loss": 1.0872, | |
| "step": 70 | |
| }, | |
| { | |
| "epoch": 0.65, | |
| "grad_norm": 2.4728469848632812, | |
| "learning_rate": 0.00016774193548387098, | |
| "loss": 0.87, | |
| "step": 80 | |
| }, | |
| { | |
| "epoch": 0.73, | |
| "grad_norm": 3.819819688796997, | |
| "learning_rate": 0.00016370967741935485, | |
| "loss": 1.2173, | |
| "step": 90 | |
| }, | |
| { | |
| "epoch": 0.81, | |
| "grad_norm": 1.585181713104248, | |
| "learning_rate": 0.00015967741935483872, | |
| "loss": 1.1711, | |
| "step": 100 | |
| }, | |
| { | |
| "epoch": 0.81, | |
| "eval_accuracy": 0.5981735159817352, | |
| "eval_loss": 1.0254943370819092, | |
| "eval_runtime": 7.7395, | |
| "eval_samples_per_second": 28.296, | |
| "eval_steps_per_second": 3.618, | |
| "step": 100 | |
| }, | |
| { | |
| "epoch": 0.89, | |
| "grad_norm": 1.5606015920639038, | |
| "learning_rate": 0.0001556451612903226, | |
| "loss": 0.9609, | |
| "step": 110 | |
| }, | |
| { | |
| "epoch": 0.97, | |
| "grad_norm": 2.0005290508270264, | |
| "learning_rate": 0.00015161290322580646, | |
| "loss": 0.8673, | |
| "step": 120 | |
| }, | |
| { | |
| "epoch": 1.05, | |
| "grad_norm": 1.7934025526046753, | |
| "learning_rate": 0.00014758064516129032, | |
| "loss": 0.7807, | |
| "step": 130 | |
| }, | |
| { | |
| "epoch": 1.13, | |
| "grad_norm": 2.0995261669158936, | |
| "learning_rate": 0.00014354838709677422, | |
| "loss": 0.7211, | |
| "step": 140 | |
| }, | |
| { | |
| "epoch": 1.21, | |
| "grad_norm": 1.2384763956069946, | |
| "learning_rate": 0.0001395161290322581, | |
| "loss": 0.8051, | |
| "step": 150 | |
| }, | |
| { | |
| "epoch": 1.29, | |
| "grad_norm": 3.001422643661499, | |
| "learning_rate": 0.00013548387096774193, | |
| "loss": 0.9052, | |
| "step": 160 | |
| }, | |
| { | |
| "epoch": 1.37, | |
| "grad_norm": 1.8868207931518555, | |
| "learning_rate": 0.0001314516129032258, | |
| "loss": 0.7217, | |
| "step": 170 | |
| }, | |
| { | |
| "epoch": 1.45, | |
| "grad_norm": 1.7666105031967163, | |
| "learning_rate": 0.0001274193548387097, | |
| "loss": 0.728, | |
| "step": 180 | |
| }, | |
| { | |
| "epoch": 1.53, | |
| "grad_norm": 2.378178358078003, | |
| "learning_rate": 0.00012338709677419356, | |
| "loss": 0.7242, | |
| "step": 190 | |
| }, | |
| { | |
| "epoch": 1.61, | |
| "grad_norm": 2.4882402420043945, | |
| "learning_rate": 0.00011935483870967743, | |
| "loss": 0.7083, | |
| "step": 200 | |
| }, | |
| { | |
| "epoch": 1.61, | |
| "eval_accuracy": 0.6575342465753424, | |
| "eval_loss": 0.7924718260765076, | |
| "eval_runtime": 7.9712, | |
| "eval_samples_per_second": 27.474, | |
| "eval_steps_per_second": 3.513, | |
| "step": 200 | |
| }, | |
| { | |
| "epoch": 1.69, | |
| "grad_norm": 3.762314796447754, | |
| "learning_rate": 0.00011532258064516131, | |
| "loss": 0.6794, | |
| "step": 210 | |
| }, | |
| { | |
| "epoch": 1.77, | |
| "grad_norm": 1.8527878522872925, | |
| "learning_rate": 0.00011129032258064515, | |
| "loss": 0.5424, | |
| "step": 220 | |
| }, | |
| { | |
| "epoch": 1.85, | |
| "grad_norm": 2.1447601318359375, | |
| "learning_rate": 0.00010725806451612903, | |
| "loss": 0.7186, | |
| "step": 230 | |
| }, | |
| { | |
| "epoch": 1.94, | |
| "grad_norm": 2.2273943424224854, | |
| "learning_rate": 0.0001032258064516129, | |
| "loss": 0.7786, | |
| "step": 240 | |
| }, | |
| { | |
| "epoch": 2.02, | |
| "grad_norm": 1.1381219625473022, | |
| "learning_rate": 9.919354838709678e-05, | |
| "loss": 0.5616, | |
| "step": 250 | |
| }, | |
| { | |
| "epoch": 2.1, | |
| "grad_norm": 1.41716730594635, | |
| "learning_rate": 9.516129032258065e-05, | |
| "loss": 0.3615, | |
| "step": 260 | |
| }, | |
| { | |
| "epoch": 2.18, | |
| "grad_norm": 1.193400263786316, | |
| "learning_rate": 9.112903225806452e-05, | |
| "loss": 0.3589, | |
| "step": 270 | |
| }, | |
| { | |
| "epoch": 2.26, | |
| "grad_norm": 1.1476421356201172, | |
| "learning_rate": 8.709677419354839e-05, | |
| "loss": 0.302, | |
| "step": 280 | |
| }, | |
| { | |
| "epoch": 2.34, | |
| "grad_norm": 2.02689790725708, | |
| "learning_rate": 8.306451612903227e-05, | |
| "loss": 0.4161, | |
| "step": 290 | |
| }, | |
| { | |
| "epoch": 2.42, | |
| "grad_norm": 1.2040166854858398, | |
| "learning_rate": 7.903225806451613e-05, | |
| "loss": 0.2479, | |
| "step": 300 | |
| }, | |
| { | |
| "epoch": 2.42, | |
| "eval_accuracy": 0.6940639269406392, | |
| "eval_loss": 0.871184766292572, | |
| "eval_runtime": 7.1874, | |
| "eval_samples_per_second": 30.47, | |
| "eval_steps_per_second": 3.896, | |
| "step": 300 | |
| }, | |
| { | |
| "epoch": 2.5, | |
| "grad_norm": 4.465162754058838, | |
| "learning_rate": 7.500000000000001e-05, | |
| "loss": 0.2949, | |
| "step": 310 | |
| }, | |
| { | |
| "epoch": 2.58, | |
| "grad_norm": 2.3733272552490234, | |
| "learning_rate": 7.096774193548388e-05, | |
| "loss": 0.2599, | |
| "step": 320 | |
| }, | |
| { | |
| "epoch": 2.66, | |
| "grad_norm": 3.527358293533325, | |
| "learning_rate": 6.693548387096774e-05, | |
| "loss": 0.4337, | |
| "step": 330 | |
| }, | |
| { | |
| "epoch": 2.74, | |
| "grad_norm": 2.8637776374816895, | |
| "learning_rate": 6.290322580645161e-05, | |
| "loss": 0.3137, | |
| "step": 340 | |
| }, | |
| { | |
| "epoch": 2.82, | |
| "grad_norm": 4.198659896850586, | |
| "learning_rate": 5.887096774193549e-05, | |
| "loss": 0.3192, | |
| "step": 350 | |
| }, | |
| { | |
| "epoch": 2.9, | |
| "grad_norm": 4.358771324157715, | |
| "learning_rate": 5.4838709677419355e-05, | |
| "loss": 0.3796, | |
| "step": 360 | |
| }, | |
| { | |
| "epoch": 2.98, | |
| "grad_norm": 3.1412558555603027, | |
| "learning_rate": 5.080645161290323e-05, | |
| "loss": 0.2622, | |
| "step": 370 | |
| }, | |
| { | |
| "epoch": 3.06, | |
| "grad_norm": 1.4319863319396973, | |
| "learning_rate": 4.67741935483871e-05, | |
| "loss": 0.2016, | |
| "step": 380 | |
| }, | |
| { | |
| "epoch": 3.15, | |
| "grad_norm": 0.5633509755134583, | |
| "learning_rate": 4.2741935483870973e-05, | |
| "loss": 0.1276, | |
| "step": 390 | |
| }, | |
| { | |
| "epoch": 3.23, | |
| "grad_norm": 0.31856054067611694, | |
| "learning_rate": 3.870967741935484e-05, | |
| "loss": 0.127, | |
| "step": 400 | |
| }, | |
| { | |
| "epoch": 3.23, | |
| "eval_accuracy": 0.6940639269406392, | |
| "eval_loss": 0.8440136313438416, | |
| "eval_runtime": 7.6031, | |
| "eval_samples_per_second": 28.804, | |
| "eval_steps_per_second": 3.683, | |
| "step": 400 | |
| }, | |
| { | |
| "epoch": 3.31, | |
| "grad_norm": 0.2085038274526596, | |
| "learning_rate": 3.467741935483872e-05, | |
| "loss": 0.1289, | |
| "step": 410 | |
| }, | |
| { | |
| "epoch": 3.39, | |
| "grad_norm": 2.471668004989624, | |
| "learning_rate": 3.0645161290322585e-05, | |
| "loss": 0.1495, | |
| "step": 420 | |
| }, | |
| { | |
| "epoch": 3.47, | |
| "grad_norm": 1.026688575744629, | |
| "learning_rate": 2.661290322580645e-05, | |
| "loss": 0.1389, | |
| "step": 430 | |
| }, | |
| { | |
| "epoch": 3.55, | |
| "grad_norm": 4.408120155334473, | |
| "learning_rate": 2.258064516129032e-05, | |
| "loss": 0.1799, | |
| "step": 440 | |
| }, | |
| { | |
| "epoch": 3.63, | |
| "grad_norm": 1.7386049032211304, | |
| "learning_rate": 1.8548387096774193e-05, | |
| "loss": 0.0727, | |
| "step": 450 | |
| }, | |
| { | |
| "epoch": 3.71, | |
| "grad_norm": 0.9766908884048462, | |
| "learning_rate": 1.4516129032258066e-05, | |
| "loss": 0.1062, | |
| "step": 460 | |
| }, | |
| { | |
| "epoch": 3.79, | |
| "grad_norm": 0.2413896918296814, | |
| "learning_rate": 1.0483870967741936e-05, | |
| "loss": 0.1303, | |
| "step": 470 | |
| }, | |
| { | |
| "epoch": 3.87, | |
| "grad_norm": 1.249718427658081, | |
| "learning_rate": 6.451612903225806e-06, | |
| "loss": 0.1025, | |
| "step": 480 | |
| }, | |
| { | |
| "epoch": 3.95, | |
| "grad_norm": 3.0604677200317383, | |
| "learning_rate": 2.4193548387096776e-06, | |
| "loss": 0.1515, | |
| "step": 490 | |
| }, | |
| { | |
| "epoch": 4.0, | |
| "step": 496, | |
| "total_flos": 6.10974224738132e+17, | |
| "train_loss": 0.575864625193419, | |
| "train_runtime": 402.9871, | |
| "train_samples_per_second": 19.564, | |
| "train_steps_per_second": 1.231 | |
| } | |
| ], | |
| "logging_steps": 10, | |
| "max_steps": 496, | |
| "num_input_tokens_seen": 0, | |
| "num_train_epochs": 4, | |
| "save_steps": 100, | |
| "total_flos": 6.10974224738132e+17, | |
| "train_batch_size": 16, | |
| "trial_name": null, | |
| "trial_params": null | |
| } | |