{ "best_metric": 1.0, "best_model_checkpoint": "vit-base-patch16-224-Trial008-YEL_STEM/checkpoint-13", "epoch": 50.0, "global_step": 50, "is_hyper_param_search": false, "is_local_process_zero": true, "is_world_process_zero": true, "log_history": [ { "epoch": 1.0, "learning_rate": 1e-05, "loss": 0.7185, "step": 1 }, { "epoch": 1.0, "eval_accuracy": 0.5925925925925926, "eval_loss": 0.7705853581428528, "eval_runtime": 0.1712, "eval_samples_per_second": 157.732, "eval_steps_per_second": 5.842, "step": 1 }, { "epoch": 2.0, "learning_rate": 2e-05, "loss": 0.7119, "step": 2 }, { "epoch": 2.0, "eval_accuracy": 0.5555555555555556, "eval_loss": 0.6863459944725037, "eval_runtime": 0.1662, "eval_samples_per_second": 162.481, "eval_steps_per_second": 6.018, "step": 2 }, { "epoch": 3.0, "learning_rate": 3e-05, "loss": 0.6949, "step": 3 }, { "epoch": 3.0, "eval_accuracy": 0.6296296296296297, "eval_loss": 0.6335163116455078, "eval_runtime": 0.1656, "eval_samples_per_second": 163.044, "eval_steps_per_second": 6.039, "step": 3 }, { "epoch": 4.0, "learning_rate": 4e-05, "loss": 0.6605, "step": 4 }, { "epoch": 4.0, "eval_accuracy": 0.7037037037037037, "eval_loss": 0.5916621088981628, "eval_runtime": 0.1676, "eval_samples_per_second": 161.123, "eval_steps_per_second": 5.968, "step": 4 }, { "epoch": 5.0, "learning_rate": 5e-05, "loss": 0.6505, "step": 5 }, { "epoch": 5.0, "eval_accuracy": 0.7407407407407407, "eval_loss": 0.5240380764007568, "eval_runtime": 0.1669, "eval_samples_per_second": 161.734, "eval_steps_per_second": 5.99, "step": 5 }, { "epoch": 6.0, "learning_rate": 4.888888888888889e-05, "loss": 0.59, "step": 6 }, { "epoch": 6.0, "eval_accuracy": 0.7037037037037037, "eval_loss": 0.5366202592849731, "eval_runtime": 0.1664, "eval_samples_per_second": 162.276, "eval_steps_per_second": 6.01, "step": 6 }, { "epoch": 7.0, "learning_rate": 4.7777777777777784e-05, "loss": 0.6172, "step": 7 }, { "epoch": 7.0, "eval_accuracy": 0.7407407407407407, "eval_loss": 0.460934042930603, "eval_runtime": 0.1699, "eval_samples_per_second": 158.909, "eval_steps_per_second": 5.886, "step": 7 }, { "epoch": 8.0, "learning_rate": 4.666666666666667e-05, "loss": 0.5515, "step": 8 }, { "epoch": 8.0, "eval_accuracy": 0.8888888888888888, "eval_loss": 0.36247193813323975, "eval_runtime": 0.1698, "eval_samples_per_second": 158.965, "eval_steps_per_second": 5.888, "step": 8 }, { "epoch": 9.0, "learning_rate": 4.555555555555556e-05, "loss": 0.5121, "step": 9 }, { "epoch": 9.0, "eval_accuracy": 0.8888888888888888, "eval_loss": 0.32386109232902527, "eval_runtime": 0.1742, "eval_samples_per_second": 154.975, "eval_steps_per_second": 5.74, "step": 9 }, { "epoch": 10.0, "learning_rate": 4.4444444444444447e-05, "loss": 0.5379, "step": 10 }, { "epoch": 10.0, "eval_accuracy": 0.8518518518518519, "eval_loss": 0.3688213527202606, "eval_runtime": 0.1675, "eval_samples_per_second": 161.227, "eval_steps_per_second": 5.971, "step": 10 }, { "epoch": 11.0, "learning_rate": 4.3333333333333334e-05, "loss": 0.4648, "step": 11 }, { "epoch": 11.0, "eval_accuracy": 0.8148148148148148, "eval_loss": 0.35447680950164795, "eval_runtime": 0.1732, "eval_samples_per_second": 155.859, "eval_steps_per_second": 5.773, "step": 11 }, { "epoch": 12.0, "learning_rate": 4.222222222222222e-05, "loss": 0.4653, "step": 12 }, { "epoch": 12.0, "eval_accuracy": 0.9259259259259259, "eval_loss": 0.24839486181735992, "eval_runtime": 0.1653, "eval_samples_per_second": 163.353, "eval_steps_per_second": 6.05, "step": 12 }, { "epoch": 13.0, "learning_rate": 4.111111111111111e-05, "loss": 0.4433, "step": 13 }, { "epoch": 13.0, "eval_accuracy": 1.0, "eval_loss": 0.19366493821144104, "eval_runtime": 0.1654, "eval_samples_per_second": 163.25, "eval_steps_per_second": 6.046, "step": 13 }, { "epoch": 14.0, "learning_rate": 4e-05, "loss": 0.4537, "step": 14 }, { "epoch": 14.0, "eval_accuracy": 0.9629629629629629, "eval_loss": 0.18937674164772034, "eval_runtime": 0.1655, "eval_samples_per_second": 163.171, "eval_steps_per_second": 6.043, "step": 14 }, { "epoch": 15.0, "learning_rate": 3.888888888888889e-05, "loss": 0.427, "step": 15 }, { "epoch": 15.0, "eval_accuracy": 0.8518518518518519, "eval_loss": 0.22488076984882355, "eval_runtime": 0.1664, "eval_samples_per_second": 162.218, "eval_steps_per_second": 6.008, "step": 15 }, { "epoch": 16.0, "learning_rate": 3.777777777777778e-05, "loss": 0.4154, "step": 16 }, { "epoch": 16.0, "eval_accuracy": 0.9629629629629629, "eval_loss": 0.15886831283569336, "eval_runtime": 0.1726, "eval_samples_per_second": 156.432, "eval_steps_per_second": 5.794, "step": 16 }, { "epoch": 17.0, "learning_rate": 3.6666666666666666e-05, "loss": 0.3895, "step": 17 }, { "epoch": 17.0, "eval_accuracy": 1.0, "eval_loss": 0.10410090535879135, "eval_runtime": 0.1666, "eval_samples_per_second": 162.095, "eval_steps_per_second": 6.004, "step": 17 }, { "epoch": 18.0, "learning_rate": 3.555555555555556e-05, "loss": 0.3994, "step": 18 }, { "epoch": 18.0, "eval_accuracy": 1.0, "eval_loss": 0.09158260375261307, "eval_runtime": 0.1684, "eval_samples_per_second": 160.318, "eval_steps_per_second": 5.938, "step": 18 }, { "epoch": 19.0, "learning_rate": 3.444444444444445e-05, "loss": 0.3692, "step": 19 }, { "epoch": 19.0, "eval_accuracy": 1.0, "eval_loss": 0.09137348085641861, "eval_runtime": 0.1719, "eval_samples_per_second": 157.043, "eval_steps_per_second": 5.816, "step": 19 }, { "epoch": 20.0, "learning_rate": 3.3333333333333335e-05, "loss": 0.3647, "step": 20 }, { "epoch": 20.0, "eval_accuracy": 0.8888888888888888, "eval_loss": 0.14642660319805145, "eval_runtime": 0.1664, "eval_samples_per_second": 162.284, "eval_steps_per_second": 6.011, "step": 20 }, { "epoch": 21.0, "learning_rate": 3.222222222222223e-05, "loss": 0.3789, "step": 21 }, { "epoch": 21.0, "eval_accuracy": 0.9259259259259259, "eval_loss": 0.15246732532978058, "eval_runtime": 0.1668, "eval_samples_per_second": 161.905, "eval_steps_per_second": 5.996, "step": 21 }, { "epoch": 22.0, "learning_rate": 3.111111111111111e-05, "loss": 0.3889, "step": 22 }, { "epoch": 22.0, "eval_accuracy": 1.0, "eval_loss": 0.09971807897090912, "eval_runtime": 0.1672, "eval_samples_per_second": 161.509, "eval_steps_per_second": 5.982, "step": 22 }, { "epoch": 23.0, "learning_rate": 3e-05, "loss": 0.3312, "step": 23 }, { "epoch": 23.0, "eval_accuracy": 1.0, "eval_loss": 0.0697513148188591, "eval_runtime": 0.1661, "eval_samples_per_second": 162.597, "eval_steps_per_second": 6.022, "step": 23 }, { "epoch": 24.0, "learning_rate": 2.8888888888888888e-05, "loss": 0.3653, "step": 24 }, { "epoch": 24.0, "eval_accuracy": 1.0, "eval_loss": 0.06495417654514313, "eval_runtime": 0.1655, "eval_samples_per_second": 163.146, "eval_steps_per_second": 6.042, "step": 24 }, { "epoch": 25.0, "learning_rate": 2.777777777777778e-05, "loss": 0.3499, "step": 25 }, { "epoch": 25.0, "eval_accuracy": 1.0, "eval_loss": 0.06258516013622284, "eval_runtime": 0.1682, "eval_samples_per_second": 160.522, "eval_steps_per_second": 5.945, "step": 25 }, { "epoch": 26.0, "learning_rate": 2.6666666666666667e-05, "loss": 0.3602, "step": 26 }, { "epoch": 26.0, "eval_accuracy": 1.0, "eval_loss": 0.07324908673763275, "eval_runtime": 0.1705, "eval_samples_per_second": 158.4, "eval_steps_per_second": 5.867, "step": 26 }, { "epoch": 27.0, "learning_rate": 2.5555555555555554e-05, "loss": 0.3209, "step": 27 }, { "epoch": 27.0, "eval_accuracy": 1.0, "eval_loss": 0.06223699823021889, "eval_runtime": 0.1675, "eval_samples_per_second": 161.197, "eval_steps_per_second": 5.97, "step": 27 }, { "epoch": 28.0, "learning_rate": 2.4444444444444445e-05, "loss": 0.3, "step": 28 }, { "epoch": 28.0, "eval_accuracy": 1.0, "eval_loss": 0.054389819502830505, "eval_runtime": 0.1713, "eval_samples_per_second": 157.638, "eval_steps_per_second": 5.838, "step": 28 }, { "epoch": 29.0, "learning_rate": 2.3333333333333336e-05, "loss": 0.2738, "step": 29 }, { "epoch": 29.0, "eval_accuracy": 1.0, "eval_loss": 0.04480647295713425, "eval_runtime": 0.1694, "eval_samples_per_second": 159.352, "eval_steps_per_second": 5.902, "step": 29 }, { "epoch": 30.0, "learning_rate": 2.2222222222222223e-05, "loss": 0.3283, "step": 30 }, { "epoch": 30.0, "eval_accuracy": 1.0, "eval_loss": 0.04304824769496918, "eval_runtime": 0.1678, "eval_samples_per_second": 160.864, "eval_steps_per_second": 5.958, "step": 30 }, { "epoch": 31.0, "learning_rate": 2.111111111111111e-05, "loss": 0.3162, "step": 31 }, { "epoch": 31.0, "eval_accuracy": 1.0, "eval_loss": 0.040174148976802826, "eval_runtime": 0.168, "eval_samples_per_second": 160.743, "eval_steps_per_second": 5.953, "step": 31 }, { "epoch": 32.0, "learning_rate": 2e-05, "loss": 0.3411, "step": 32 }, { "epoch": 32.0, "eval_accuracy": 1.0, "eval_loss": 0.03942341357469559, "eval_runtime": 0.1668, "eval_samples_per_second": 161.85, "eval_steps_per_second": 5.994, "step": 32 }, { "epoch": 33.0, "learning_rate": 1.888888888888889e-05, "loss": 0.3195, "step": 33 }, { "epoch": 33.0, "eval_accuracy": 1.0, "eval_loss": 0.03807567059993744, "eval_runtime": 0.1706, "eval_samples_per_second": 158.261, "eval_steps_per_second": 5.862, "step": 33 }, { "epoch": 34.0, "learning_rate": 1.777777777777778e-05, "loss": 0.3111, "step": 34 }, { "epoch": 34.0, "eval_accuracy": 1.0, "eval_loss": 0.0349883958697319, "eval_runtime": 0.166, "eval_samples_per_second": 162.686, "eval_steps_per_second": 6.025, "step": 34 }, { "epoch": 35.0, "learning_rate": 1.6666666666666667e-05, "loss": 0.2816, "step": 35 }, { "epoch": 35.0, "eval_accuracy": 1.0, "eval_loss": 0.034967441111803055, "eval_runtime": 0.1673, "eval_samples_per_second": 161.39, "eval_steps_per_second": 5.977, "step": 35 }, { "epoch": 36.0, "learning_rate": 1.5555555555555555e-05, "loss": 0.2602, "step": 36 }, { "epoch": 36.0, "eval_accuracy": 1.0, "eval_loss": 0.03582083806395531, "eval_runtime": 0.1664, "eval_samples_per_second": 162.232, "eval_steps_per_second": 6.009, "step": 36 }, { "epoch": 37.0, "learning_rate": 1.4444444444444444e-05, "loss": 0.3128, "step": 37 }, { "epoch": 37.0, "eval_accuracy": 1.0, "eval_loss": 0.038833122700452805, "eval_runtime": 0.1722, "eval_samples_per_second": 156.817, "eval_steps_per_second": 5.808, "step": 37 }, { "epoch": 38.0, "learning_rate": 1.3333333333333333e-05, "loss": 0.326, "step": 38 }, { "epoch": 38.0, "eval_accuracy": 1.0, "eval_loss": 0.04981035739183426, "eval_runtime": 0.1656, "eval_samples_per_second": 163.003, "eval_steps_per_second": 6.037, "step": 38 }, { "epoch": 39.0, "learning_rate": 1.2222222222222222e-05, "loss": 0.3228, "step": 39 }, { "epoch": 39.0, "eval_accuracy": 1.0, "eval_loss": 0.0702390968799591, "eval_runtime": 0.1726, "eval_samples_per_second": 156.474, "eval_steps_per_second": 5.795, "step": 39 }, { "epoch": 40.0, "learning_rate": 1.1111111111111112e-05, "loss": 0.3073, "step": 40 }, { "epoch": 40.0, "eval_accuracy": 0.9629629629629629, "eval_loss": 0.07822608947753906, "eval_runtime": 0.1726, "eval_samples_per_second": 156.457, "eval_steps_per_second": 5.795, "step": 40 }, { "epoch": 41.0, "learning_rate": 1e-05, "loss": 0.3266, "step": 41 }, { "epoch": 41.0, "eval_accuracy": 0.9629629629629629, "eval_loss": 0.07207214087247849, "eval_runtime": 0.1667, "eval_samples_per_second": 161.996, "eval_steps_per_second": 6.0, "step": 41 }, { "epoch": 42.0, "learning_rate": 8.88888888888889e-06, "loss": 0.3546, "step": 42 }, { "epoch": 42.0, "eval_accuracy": 1.0, "eval_loss": 0.05794690176844597, "eval_runtime": 0.1682, "eval_samples_per_second": 160.517, "eval_steps_per_second": 5.945, "step": 42 }, { "epoch": 43.0, "learning_rate": 7.777777777777777e-06, "loss": 0.2832, "step": 43 }, { "epoch": 43.0, "eval_accuracy": 1.0, "eval_loss": 0.04870154336094856, "eval_runtime": 0.1649, "eval_samples_per_second": 163.704, "eval_steps_per_second": 6.063, "step": 43 }, { "epoch": 44.0, "learning_rate": 6.666666666666667e-06, "loss": 0.2872, "step": 44 }, { "epoch": 44.0, "eval_accuracy": 1.0, "eval_loss": 0.04281274601817131, "eval_runtime": 0.1667, "eval_samples_per_second": 162.007, "eval_steps_per_second": 6.0, "step": 44 }, { "epoch": 45.0, "learning_rate": 5.555555555555556e-06, "loss": 0.2699, "step": 45 }, { "epoch": 45.0, "eval_accuracy": 1.0, "eval_loss": 0.03951677307486534, "eval_runtime": 0.1669, "eval_samples_per_second": 161.806, "eval_steps_per_second": 5.993, "step": 45 }, { "epoch": 46.0, "learning_rate": 4.444444444444445e-06, "loss": 0.3002, "step": 46 }, { "epoch": 46.0, "eval_accuracy": 1.0, "eval_loss": 0.03907058387994766, "eval_runtime": 0.1677, "eval_samples_per_second": 161.001, "eval_steps_per_second": 5.963, "step": 46 }, { "epoch": 47.0, "learning_rate": 3.3333333333333333e-06, "loss": 0.327, "step": 47 }, { "epoch": 47.0, "eval_accuracy": 1.0, "eval_loss": 0.03901884704828262, "eval_runtime": 0.168, "eval_samples_per_second": 160.751, "eval_steps_per_second": 5.954, "step": 47 }, { "epoch": 48.0, "learning_rate": 2.2222222222222225e-06, "loss": 0.2746, "step": 48 }, { "epoch": 48.0, "eval_accuracy": 1.0, "eval_loss": 0.0386575423181057, "eval_runtime": 0.1697, "eval_samples_per_second": 159.095, "eval_steps_per_second": 5.892, "step": 48 }, { "epoch": 49.0, "learning_rate": 1.1111111111111112e-06, "loss": 0.2781, "step": 49 }, { "epoch": 49.0, "eval_accuracy": 1.0, "eval_loss": 0.03857966884970665, "eval_runtime": 0.1664, "eval_samples_per_second": 162.219, "eval_steps_per_second": 6.008, "step": 49 }, { "epoch": 50.0, "learning_rate": 0.0, "loss": 0.2925, "step": 50 }, { "epoch": 50.0, "eval_accuracy": 1.0, "eval_loss": 0.03861464560031891, "eval_runtime": 0.1753, "eval_samples_per_second": 153.994, "eval_steps_per_second": 5.703, "step": 50 }, { "epoch": 50.0, "step": 50, "total_flos": 9.10530877971456e+17, "train_loss": 0.3985473644733429, "train_runtime": 860.9181, "train_samples_per_second": 13.648, "train_steps_per_second": 0.058 } ], "max_steps": 50, "num_train_epochs": 50, "total_flos": 9.10530877971456e+17, "trial_name": null, "trial_params": null }