| { | |
| "best_metric": 0.7622188329696655, | |
| "best_model_checkpoint": "./vit-base-renovation/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": 1.4599734544754028, | |
| "learning_rate": 0.00019596774193548388, | |
| "loss": 0.9874, | |
| "step": 10 | |
| }, | |
| { | |
| "epoch": 0.16, | |
| "grad_norm": 1.6008720397949219, | |
| "learning_rate": 0.00019193548387096775, | |
| "loss": 0.9737, | |
| "step": 20 | |
| }, | |
| { | |
| "epoch": 0.2, | |
| "eval_accuracy": 0.5045454545454545, | |
| "eval_loss": 1.0076260566711426, | |
| "eval_runtime": 8.1182, | |
| "eval_samples_per_second": 27.1, | |
| "eval_steps_per_second": 3.449, | |
| "step": 25 | |
| }, | |
| { | |
| "epoch": 0.24, | |
| "grad_norm": 1.4715768098831177, | |
| "learning_rate": 0.00018790322580645164, | |
| "loss": 1.0314, | |
| "step": 30 | |
| }, | |
| { | |
| "epoch": 0.32, | |
| "grad_norm": 3.563518762588501, | |
| "learning_rate": 0.00018387096774193548, | |
| "loss": 0.9846, | |
| "step": 40 | |
| }, | |
| { | |
| "epoch": 0.4, | |
| "grad_norm": 2.6104962825775146, | |
| "learning_rate": 0.00017983870967741935, | |
| "loss": 0.862, | |
| "step": 50 | |
| }, | |
| { | |
| "epoch": 0.4, | |
| "eval_accuracy": 0.5045454545454545, | |
| "eval_loss": 1.0219964981079102, | |
| "eval_runtime": 8.0163, | |
| "eval_samples_per_second": 27.444, | |
| "eval_steps_per_second": 3.493, | |
| "step": 50 | |
| }, | |
| { | |
| "epoch": 0.48, | |
| "grad_norm": 1.1494791507720947, | |
| "learning_rate": 0.00017580645161290325, | |
| "loss": 0.7918, | |
| "step": 60 | |
| }, | |
| { | |
| "epoch": 0.56, | |
| "grad_norm": 2.8406925201416016, | |
| "learning_rate": 0.00017177419354838711, | |
| "loss": 0.9064, | |
| "step": 70 | |
| }, | |
| { | |
| "epoch": 0.6, | |
| "eval_accuracy": 0.5590909090909091, | |
| "eval_loss": 0.9076008200645447, | |
| "eval_runtime": 7.2828, | |
| "eval_samples_per_second": 30.208, | |
| "eval_steps_per_second": 3.845, | |
| "step": 75 | |
| }, | |
| { | |
| "epoch": 0.65, | |
| "grad_norm": 1.1309199333190918, | |
| "learning_rate": 0.00016774193548387098, | |
| "loss": 0.944, | |
| "step": 80 | |
| }, | |
| { | |
| "epoch": 0.73, | |
| "grad_norm": 1.6623650789260864, | |
| "learning_rate": 0.00016370967741935485, | |
| "loss": 0.865, | |
| "step": 90 | |
| }, | |
| { | |
| "epoch": 0.81, | |
| "grad_norm": 2.2675681114196777, | |
| "learning_rate": 0.00015967741935483872, | |
| "loss": 0.8528, | |
| "step": 100 | |
| }, | |
| { | |
| "epoch": 0.81, | |
| "eval_accuracy": 0.65, | |
| "eval_loss": 0.8157132267951965, | |
| "eval_runtime": 7.8906, | |
| "eval_samples_per_second": 27.881, | |
| "eval_steps_per_second": 3.549, | |
| "step": 100 | |
| }, | |
| { | |
| "epoch": 0.89, | |
| "grad_norm": 2.036137819290161, | |
| "learning_rate": 0.0001556451612903226, | |
| "loss": 0.7899, | |
| "step": 110 | |
| }, | |
| { | |
| "epoch": 0.97, | |
| "grad_norm": 1.0894314050674438, | |
| "learning_rate": 0.00015161290322580646, | |
| "loss": 0.8848, | |
| "step": 120 | |
| }, | |
| { | |
| "epoch": 1.01, | |
| "eval_accuracy": 0.6272727272727273, | |
| "eval_loss": 0.8089344501495361, | |
| "eval_runtime": 7.2662, | |
| "eval_samples_per_second": 30.277, | |
| "eval_steps_per_second": 3.853, | |
| "step": 125 | |
| }, | |
| { | |
| "epoch": 1.05, | |
| "grad_norm": 1.2654739618301392, | |
| "learning_rate": 0.00014758064516129032, | |
| "loss": 0.6029, | |
| "step": 130 | |
| }, | |
| { | |
| "epoch": 1.13, | |
| "grad_norm": 2.1379916667938232, | |
| "learning_rate": 0.00014354838709677422, | |
| "loss": 0.6292, | |
| "step": 140 | |
| }, | |
| { | |
| "epoch": 1.21, | |
| "grad_norm": 3.7936744689941406, | |
| "learning_rate": 0.0001395161290322581, | |
| "loss": 0.6608, | |
| "step": 150 | |
| }, | |
| { | |
| "epoch": 1.21, | |
| "eval_accuracy": 0.6409090909090909, | |
| "eval_loss": 0.8614596128463745, | |
| "eval_runtime": 8.173, | |
| "eval_samples_per_second": 26.918, | |
| "eval_steps_per_second": 3.426, | |
| "step": 150 | |
| }, | |
| { | |
| "epoch": 1.29, | |
| "grad_norm": 1.3360966444015503, | |
| "learning_rate": 0.00013548387096774193, | |
| "loss": 0.539, | |
| "step": 160 | |
| }, | |
| { | |
| "epoch": 1.37, | |
| "grad_norm": 4.338820934295654, | |
| "learning_rate": 0.0001314516129032258, | |
| "loss": 0.6748, | |
| "step": 170 | |
| }, | |
| { | |
| "epoch": 1.41, | |
| "eval_accuracy": 0.6318181818181818, | |
| "eval_loss": 0.8425928354263306, | |
| "eval_runtime": 7.2996, | |
| "eval_samples_per_second": 30.139, | |
| "eval_steps_per_second": 3.836, | |
| "step": 175 | |
| }, | |
| { | |
| "epoch": 1.45, | |
| "grad_norm": 2.611848831176758, | |
| "learning_rate": 0.0001274193548387097, | |
| "loss": 0.6317, | |
| "step": 180 | |
| }, | |
| { | |
| "epoch": 1.53, | |
| "grad_norm": 2.4912662506103516, | |
| "learning_rate": 0.00012379032258064514, | |
| "loss": 0.6935, | |
| "step": 190 | |
| }, | |
| { | |
| "epoch": 1.61, | |
| "grad_norm": 2.0737533569335938, | |
| "learning_rate": 0.00011975806451612903, | |
| "loss": 0.6559, | |
| "step": 200 | |
| }, | |
| { | |
| "epoch": 1.61, | |
| "eval_accuracy": 0.6090909090909091, | |
| "eval_loss": 0.8427345752716064, | |
| "eval_runtime": 8.531, | |
| "eval_samples_per_second": 25.788, | |
| "eval_steps_per_second": 3.282, | |
| "step": 200 | |
| }, | |
| { | |
| "epoch": 1.69, | |
| "grad_norm": 2.8581531047821045, | |
| "learning_rate": 0.00011572580645161291, | |
| "loss": 0.5985, | |
| "step": 210 | |
| }, | |
| { | |
| "epoch": 1.77, | |
| "grad_norm": 1.7107830047607422, | |
| "learning_rate": 0.00011169354838709678, | |
| "loss": 0.5654, | |
| "step": 220 | |
| }, | |
| { | |
| "epoch": 1.81, | |
| "eval_accuracy": 0.6681818181818182, | |
| "eval_loss": 0.8266812562942505, | |
| "eval_runtime": 7.3084, | |
| "eval_samples_per_second": 30.102, | |
| "eval_steps_per_second": 3.831, | |
| "step": 225 | |
| }, | |
| { | |
| "epoch": 1.85, | |
| "grad_norm": 2.7065391540527344, | |
| "learning_rate": 0.00010766129032258066, | |
| "loss": 0.6948, | |
| "step": 230 | |
| }, | |
| { | |
| "epoch": 1.94, | |
| "grad_norm": 1.9006155729293823, | |
| "learning_rate": 0.00010362903225806453, | |
| "loss": 0.5896, | |
| "step": 240 | |
| }, | |
| { | |
| "epoch": 2.02, | |
| "grad_norm": 1.5013632774353027, | |
| "learning_rate": 9.95967741935484e-05, | |
| "loss": 0.5254, | |
| "step": 250 | |
| }, | |
| { | |
| "epoch": 2.02, | |
| "eval_accuracy": 0.6545454545454545, | |
| "eval_loss": 0.7622188329696655, | |
| "eval_runtime": 8.0247, | |
| "eval_samples_per_second": 27.415, | |
| "eval_steps_per_second": 3.489, | |
| "step": 250 | |
| }, | |
| { | |
| "epoch": 2.1, | |
| "grad_norm": 3.849762439727783, | |
| "learning_rate": 9.556451612903226e-05, | |
| "loss": 0.3218, | |
| "step": 260 | |
| }, | |
| { | |
| "epoch": 2.18, | |
| "grad_norm": 1.2341018915176392, | |
| "learning_rate": 9.153225806451613e-05, | |
| "loss": 0.2778, | |
| "step": 270 | |
| }, | |
| { | |
| "epoch": 2.22, | |
| "eval_accuracy": 0.6636363636363637, | |
| "eval_loss": 0.9480502009391785, | |
| "eval_runtime": 7.5019, | |
| "eval_samples_per_second": 29.326, | |
| "eval_steps_per_second": 3.732, | |
| "step": 275 | |
| }, | |
| { | |
| "epoch": 2.26, | |
| "grad_norm": 2.990760087966919, | |
| "learning_rate": 8.75e-05, | |
| "loss": 0.3018, | |
| "step": 280 | |
| }, | |
| { | |
| "epoch": 2.34, | |
| "grad_norm": 3.7092368602752686, | |
| "learning_rate": 8.346774193548388e-05, | |
| "loss": 0.3696, | |
| "step": 290 | |
| }, | |
| { | |
| "epoch": 2.42, | |
| "grad_norm": 4.312985420227051, | |
| "learning_rate": 7.943548387096774e-05, | |
| "loss": 0.309, | |
| "step": 300 | |
| }, | |
| { | |
| "epoch": 2.42, | |
| "eval_accuracy": 0.6409090909090909, | |
| "eval_loss": 0.8998208045959473, | |
| "eval_runtime": 8.0465, | |
| "eval_samples_per_second": 27.341, | |
| "eval_steps_per_second": 3.48, | |
| "step": 300 | |
| }, | |
| { | |
| "epoch": 2.5, | |
| "grad_norm": 3.9428765773773193, | |
| "learning_rate": 7.540322580645162e-05, | |
| "loss": 0.4113, | |
| "step": 310 | |
| }, | |
| { | |
| "epoch": 2.58, | |
| "grad_norm": 1.6300395727157593, | |
| "learning_rate": 7.137096774193549e-05, | |
| "loss": 0.2396, | |
| "step": 320 | |
| }, | |
| { | |
| "epoch": 2.62, | |
| "eval_accuracy": 0.6409090909090909, | |
| "eval_loss": 0.9171214699745178, | |
| "eval_runtime": 7.6722, | |
| "eval_samples_per_second": 28.675, | |
| "eval_steps_per_second": 3.65, | |
| "step": 325 | |
| }, | |
| { | |
| "epoch": 2.66, | |
| "grad_norm": 1.3225387334823608, | |
| "learning_rate": 6.733870967741935e-05, | |
| "loss": 0.3451, | |
| "step": 330 | |
| }, | |
| { | |
| "epoch": 2.74, | |
| "grad_norm": 2.227588176727295, | |
| "learning_rate": 6.330645161290322e-05, | |
| "loss": 0.206, | |
| "step": 340 | |
| }, | |
| { | |
| "epoch": 2.82, | |
| "grad_norm": 3.282959222793579, | |
| "learning_rate": 5.9274193548387104e-05, | |
| "loss": 0.2773, | |
| "step": 350 | |
| }, | |
| { | |
| "epoch": 2.82, | |
| "eval_accuracy": 0.6090909090909091, | |
| "eval_loss": 1.0582060813903809, | |
| "eval_runtime": 7.9345, | |
| "eval_samples_per_second": 27.727, | |
| "eval_steps_per_second": 3.529, | |
| "step": 350 | |
| }, | |
| { | |
| "epoch": 2.9, | |
| "grad_norm": 4.272079944610596, | |
| "learning_rate": 5.5241935483870966e-05, | |
| "loss": 0.4049, | |
| "step": 360 | |
| }, | |
| { | |
| "epoch": 2.98, | |
| "grad_norm": 4.28529167175293, | |
| "learning_rate": 5.120967741935484e-05, | |
| "loss": 0.2516, | |
| "step": 370 | |
| }, | |
| { | |
| "epoch": 3.02, | |
| "eval_accuracy": 0.6454545454545455, | |
| "eval_loss": 0.9362202286720276, | |
| "eval_runtime": 7.9239, | |
| "eval_samples_per_second": 27.764, | |
| "eval_steps_per_second": 3.534, | |
| "step": 375 | |
| }, | |
| { | |
| "epoch": 3.06, | |
| "grad_norm": 0.6417772173881531, | |
| "learning_rate": 4.7177419354838716e-05, | |
| "loss": 0.1439, | |
| "step": 380 | |
| }, | |
| { | |
| "epoch": 3.15, | |
| "grad_norm": 0.6277107000350952, | |
| "learning_rate": 4.3145161290322584e-05, | |
| "loss": 0.1053, | |
| "step": 390 | |
| }, | |
| { | |
| "epoch": 3.23, | |
| "grad_norm": 4.842502593994141, | |
| "learning_rate": 3.911290322580645e-05, | |
| "loss": 0.1578, | |
| "step": 400 | |
| }, | |
| { | |
| "epoch": 3.23, | |
| "eval_accuracy": 0.6772727272727272, | |
| "eval_loss": 0.9263622760772705, | |
| "eval_runtime": 8.037, | |
| "eval_samples_per_second": 27.373, | |
| "eval_steps_per_second": 3.484, | |
| "step": 400 | |
| }, | |
| { | |
| "epoch": 3.31, | |
| "grad_norm": 3.4976587295532227, | |
| "learning_rate": 3.508064516129033e-05, | |
| "loss": 0.1233, | |
| "step": 410 | |
| }, | |
| { | |
| "epoch": 3.39, | |
| "grad_norm": 0.3740604519844055, | |
| "learning_rate": 3.1048387096774195e-05, | |
| "loss": 0.0979, | |
| "step": 420 | |
| }, | |
| { | |
| "epoch": 3.43, | |
| "eval_accuracy": 0.6772727272727272, | |
| "eval_loss": 0.946960985660553, | |
| "eval_runtime": 7.5496, | |
| "eval_samples_per_second": 29.141, | |
| "eval_steps_per_second": 3.709, | |
| "step": 425 | |
| }, | |
| { | |
| "epoch": 3.47, | |
| "grad_norm": 0.2739836275577545, | |
| "learning_rate": 2.7016129032258064e-05, | |
| "loss": 0.083, | |
| "step": 430 | |
| }, | |
| { | |
| "epoch": 3.55, | |
| "grad_norm": 2.7154436111450195, | |
| "learning_rate": 2.2983870967741935e-05, | |
| "loss": 0.085, | |
| "step": 440 | |
| }, | |
| { | |
| "epoch": 3.63, | |
| "grad_norm": 1.2921925783157349, | |
| "learning_rate": 1.8951612903225807e-05, | |
| "loss": 0.0836, | |
| "step": 450 | |
| }, | |
| { | |
| "epoch": 3.63, | |
| "eval_accuracy": 0.6681818181818182, | |
| "eval_loss": 0.9940941333770752, | |
| "eval_runtime": 8.0024, | |
| "eval_samples_per_second": 27.492, | |
| "eval_steps_per_second": 3.499, | |
| "step": 450 | |
| }, | |
| { | |
| "epoch": 3.71, | |
| "grad_norm": 0.7346110939979553, | |
| "learning_rate": 1.4919354838709679e-05, | |
| "loss": 0.0661, | |
| "step": 460 | |
| }, | |
| { | |
| "epoch": 3.79, | |
| "grad_norm": 0.540484607219696, | |
| "learning_rate": 1.0887096774193549e-05, | |
| "loss": 0.126, | |
| "step": 470 | |
| }, | |
| { | |
| "epoch": 3.83, | |
| "eval_accuracy": 0.6863636363636364, | |
| "eval_loss": 0.9761253595352173, | |
| "eval_runtime": 7.4163, | |
| "eval_samples_per_second": 29.665, | |
| "eval_steps_per_second": 3.775, | |
| "step": 475 | |
| }, | |
| { | |
| "epoch": 3.87, | |
| "grad_norm": 1.4092530012130737, | |
| "learning_rate": 6.854838709677419e-06, | |
| "loss": 0.0657, | |
| "step": 480 | |
| }, | |
| { | |
| "epoch": 3.95, | |
| "grad_norm": 1.824102759361267, | |
| "learning_rate": 3.225806451612903e-06, | |
| "loss": 0.075, | |
| "step": 490 | |
| }, | |
| { | |
| "epoch": 4.0, | |
| "step": 496, | |
| "total_flos": 6.137420584742093e+17, | |
| "train_loss": 0.4817413741181935, | |
| "train_runtime": 646.4292, | |
| "train_samples_per_second": 12.252, | |
| "train_steps_per_second": 0.767 | |
| } | |
| ], | |
| "logging_steps": 10, | |
| "max_steps": 496, | |
| "num_input_tokens_seen": 0, | |
| "num_train_epochs": 4, | |
| "save_steps": 25, | |
| "total_flos": 6.137420584742093e+17, | |
| "train_batch_size": 16, | |
| "trial_name": null, | |
| "trial_params": null | |
| } | |