| { | |
| "best_metric": 0.710116982460022, | |
| "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": 1.7348554134368896, | |
| "learning_rate": 0.00019596774193548388, | |
| "loss": 1.02, | |
| "step": 10 | |
| }, | |
| { | |
| "epoch": 0.16, | |
| "grad_norm": 0.9956100583076477, | |
| "learning_rate": 0.00019193548387096775, | |
| "loss": 1.0036, | |
| "step": 20 | |
| }, | |
| { | |
| "epoch": 0.2, | |
| "eval_accuracy": 0.5, | |
| "eval_loss": 0.9849038124084473, | |
| "eval_runtime": 8.3244, | |
| "eval_samples_per_second": 26.428, | |
| "eval_steps_per_second": 3.364, | |
| "step": 25 | |
| }, | |
| { | |
| "epoch": 0.24, | |
| "grad_norm": 1.1454524993896484, | |
| "learning_rate": 0.00018790322580645164, | |
| "loss": 0.9917, | |
| "step": 30 | |
| }, | |
| { | |
| "epoch": 0.32, | |
| "grad_norm": 2.6789638996124268, | |
| "learning_rate": 0.00018387096774193548, | |
| "loss": 0.8835, | |
| "step": 40 | |
| }, | |
| { | |
| "epoch": 0.4, | |
| "grad_norm": 2.0885393619537354, | |
| "learning_rate": 0.00017983870967741935, | |
| "loss": 0.8051, | |
| "step": 50 | |
| }, | |
| { | |
| "epoch": 0.4, | |
| "eval_accuracy": 0.5545454545454546, | |
| "eval_loss": 0.9105865359306335, | |
| "eval_runtime": 8.5151, | |
| "eval_samples_per_second": 25.836, | |
| "eval_steps_per_second": 3.288, | |
| "step": 50 | |
| }, | |
| { | |
| "epoch": 0.48, | |
| "grad_norm": 1.2426408529281616, | |
| "learning_rate": 0.00017580645161290325, | |
| "loss": 0.7545, | |
| "step": 60 | |
| }, | |
| { | |
| "epoch": 0.56, | |
| "grad_norm": 3.0708746910095215, | |
| "learning_rate": 0.00017177419354838711, | |
| "loss": 0.8336, | |
| "step": 70 | |
| }, | |
| { | |
| "epoch": 0.6, | |
| "eval_accuracy": 0.5954545454545455, | |
| "eval_loss": 0.9004097580909729, | |
| "eval_runtime": 8.3678, | |
| "eval_samples_per_second": 26.291, | |
| "eval_steps_per_second": 3.346, | |
| "step": 75 | |
| }, | |
| { | |
| "epoch": 0.65, | |
| "grad_norm": 1.7403985261917114, | |
| "learning_rate": 0.0001681451612903226, | |
| "loss": 0.8614, | |
| "step": 80 | |
| }, | |
| { | |
| "epoch": 0.73, | |
| "grad_norm": 1.7942191362380981, | |
| "learning_rate": 0.00016411290322580646, | |
| "loss": 0.8141, | |
| "step": 90 | |
| }, | |
| { | |
| "epoch": 0.81, | |
| "grad_norm": 2.56789231300354, | |
| "learning_rate": 0.00016008064516129033, | |
| "loss": 0.786, | |
| "step": 100 | |
| }, | |
| { | |
| "epoch": 0.81, | |
| "eval_accuracy": 0.6454545454545455, | |
| "eval_loss": 0.7700986266136169, | |
| "eval_runtime": 8.4023, | |
| "eval_samples_per_second": 26.183, | |
| "eval_steps_per_second": 3.332, | |
| "step": 100 | |
| }, | |
| { | |
| "epoch": 0.89, | |
| "grad_norm": 2.265244722366333, | |
| "learning_rate": 0.0001560483870967742, | |
| "loss": 0.7118, | |
| "step": 110 | |
| }, | |
| { | |
| "epoch": 0.97, | |
| "grad_norm": 2.001025915145874, | |
| "learning_rate": 0.0001520161290322581, | |
| "loss": 0.7854, | |
| "step": 120 | |
| }, | |
| { | |
| "epoch": 1.01, | |
| "eval_accuracy": 0.6227272727272727, | |
| "eval_loss": 0.7561001777648926, | |
| "eval_runtime": 8.1474, | |
| "eval_samples_per_second": 27.002, | |
| "eval_steps_per_second": 3.437, | |
| "step": 125 | |
| }, | |
| { | |
| "epoch": 1.05, | |
| "grad_norm": 1.243610143661499, | |
| "learning_rate": 0.00014798387096774193, | |
| "loss": 0.5415, | |
| "step": 130 | |
| }, | |
| { | |
| "epoch": 1.13, | |
| "grad_norm": 1.9543741941452026, | |
| "learning_rate": 0.0001439516129032258, | |
| "loss": 0.5246, | |
| "step": 140 | |
| }, | |
| { | |
| "epoch": 1.21, | |
| "grad_norm": 1.4787832498550415, | |
| "learning_rate": 0.00013991935483870967, | |
| "loss": 0.4603, | |
| "step": 150 | |
| }, | |
| { | |
| "epoch": 1.21, | |
| "eval_accuracy": 0.6409090909090909, | |
| "eval_loss": 0.8104726672172546, | |
| "eval_runtime": 7.4392, | |
| "eval_samples_per_second": 29.573, | |
| "eval_steps_per_second": 3.764, | |
| "step": 150 | |
| }, | |
| { | |
| "epoch": 1.29, | |
| "grad_norm": 1.9429636001586914, | |
| "learning_rate": 0.00013588709677419357, | |
| "loss": 0.442, | |
| "step": 160 | |
| }, | |
| { | |
| "epoch": 1.37, | |
| "grad_norm": 2.558549642562866, | |
| "learning_rate": 0.00013185483870967743, | |
| "loss": 0.4934, | |
| "step": 170 | |
| }, | |
| { | |
| "epoch": 1.41, | |
| "eval_accuracy": 0.6181818181818182, | |
| "eval_loss": 0.874618411064148, | |
| "eval_runtime": 7.375, | |
| "eval_samples_per_second": 29.831, | |
| "eval_steps_per_second": 3.797, | |
| "step": 175 | |
| }, | |
| { | |
| "epoch": 1.45, | |
| "grad_norm": 3.0746328830718994, | |
| "learning_rate": 0.0001278225806451613, | |
| "loss": 0.5744, | |
| "step": 180 | |
| }, | |
| { | |
| "epoch": 1.53, | |
| "grad_norm": 2.0470433235168457, | |
| "learning_rate": 0.00012379032258064514, | |
| "loss": 0.5933, | |
| "step": 190 | |
| }, | |
| { | |
| "epoch": 1.61, | |
| "grad_norm": 2.0045182704925537, | |
| "learning_rate": 0.00011975806451612903, | |
| "loss": 0.5315, | |
| "step": 200 | |
| }, | |
| { | |
| "epoch": 1.61, | |
| "eval_accuracy": 0.6636363636363637, | |
| "eval_loss": 0.8266905546188354, | |
| "eval_runtime": 7.745, | |
| "eval_samples_per_second": 28.405, | |
| "eval_steps_per_second": 3.615, | |
| "step": 200 | |
| }, | |
| { | |
| "epoch": 1.69, | |
| "grad_norm": 2.0231032371520996, | |
| "learning_rate": 0.00011572580645161291, | |
| "loss": 0.5538, | |
| "step": 210 | |
| }, | |
| { | |
| "epoch": 1.77, | |
| "grad_norm": 2.176417589187622, | |
| "learning_rate": 0.00011169354838709678, | |
| "loss": 0.5251, | |
| "step": 220 | |
| }, | |
| { | |
| "epoch": 1.81, | |
| "eval_accuracy": 0.65, | |
| "eval_loss": 0.8585435152053833, | |
| "eval_runtime": 7.3957, | |
| "eval_samples_per_second": 29.747, | |
| "eval_steps_per_second": 3.786, | |
| "step": 225 | |
| }, | |
| { | |
| "epoch": 1.85, | |
| "grad_norm": 3.1730520725250244, | |
| "learning_rate": 0.00010806451612903225, | |
| "loss": 0.5472, | |
| "step": 230 | |
| }, | |
| { | |
| "epoch": 1.94, | |
| "grad_norm": 2.2670953273773193, | |
| "learning_rate": 0.00010403225806451614, | |
| "loss": 0.5488, | |
| "step": 240 | |
| }, | |
| { | |
| "epoch": 2.02, | |
| "grad_norm": 0.9752326607704163, | |
| "learning_rate": 0.0001, | |
| "loss": 0.4386, | |
| "step": 250 | |
| }, | |
| { | |
| "epoch": 2.02, | |
| "eval_accuracy": 0.6909090909090909, | |
| "eval_loss": 0.710116982460022, | |
| "eval_runtime": 8.4962, | |
| "eval_samples_per_second": 25.894, | |
| "eval_steps_per_second": 3.296, | |
| "step": 250 | |
| }, | |
| { | |
| "epoch": 2.1, | |
| "grad_norm": 0.6374533176422119, | |
| "learning_rate": 9.596774193548387e-05, | |
| "loss": 0.241, | |
| "step": 260 | |
| }, | |
| { | |
| "epoch": 2.18, | |
| "grad_norm": 0.9005341529846191, | |
| "learning_rate": 9.193548387096774e-05, | |
| "loss": 0.2627, | |
| "step": 270 | |
| }, | |
| { | |
| "epoch": 2.22, | |
| "eval_accuracy": 0.6409090909090909, | |
| "eval_loss": 1.0041698217391968, | |
| "eval_runtime": 7.4719, | |
| "eval_samples_per_second": 29.443, | |
| "eval_steps_per_second": 3.747, | |
| "step": 275 | |
| }, | |
| { | |
| "epoch": 2.26, | |
| "grad_norm": 2.1455209255218506, | |
| "learning_rate": 8.790322580645162e-05, | |
| "loss": 0.1868, | |
| "step": 280 | |
| }, | |
| { | |
| "epoch": 2.34, | |
| "grad_norm": 0.3361818790435791, | |
| "learning_rate": 8.387096774193549e-05, | |
| "loss": 0.1521, | |
| "step": 290 | |
| }, | |
| { | |
| "epoch": 2.42, | |
| "grad_norm": 5.28161096572876, | |
| "learning_rate": 7.983870967741936e-05, | |
| "loss": 0.1524, | |
| "step": 300 | |
| }, | |
| { | |
| "epoch": 2.42, | |
| "eval_accuracy": 0.6545454545454545, | |
| "eval_loss": 0.9489237666130066, | |
| "eval_runtime": 8.3095, | |
| "eval_samples_per_second": 26.476, | |
| "eval_steps_per_second": 3.37, | |
| "step": 300 | |
| }, | |
| { | |
| "epoch": 2.5, | |
| "grad_norm": 2.760798215866089, | |
| "learning_rate": 7.580645161290323e-05, | |
| "loss": 0.1971, | |
| "step": 310 | |
| }, | |
| { | |
| "epoch": 2.58, | |
| "grad_norm": 1.529565453529358, | |
| "learning_rate": 7.177419354838711e-05, | |
| "loss": 0.1272, | |
| "step": 320 | |
| }, | |
| { | |
| "epoch": 2.62, | |
| "eval_accuracy": 0.65, | |
| "eval_loss": 1.066292405128479, | |
| "eval_runtime": 8.3186, | |
| "eval_samples_per_second": 26.447, | |
| "eval_steps_per_second": 3.366, | |
| "step": 325 | |
| }, | |
| { | |
| "epoch": 2.66, | |
| "grad_norm": 0.5179287791252136, | |
| "learning_rate": 6.774193548387096e-05, | |
| "loss": 0.1676, | |
| "step": 330 | |
| }, | |
| { | |
| "epoch": 2.74, | |
| "grad_norm": 1.027869701385498, | |
| "learning_rate": 6.370967741935485e-05, | |
| "loss": 0.1194, | |
| "step": 340 | |
| }, | |
| { | |
| "epoch": 2.82, | |
| "grad_norm": 1.2000805139541626, | |
| "learning_rate": 5.9677419354838715e-05, | |
| "loss": 0.186, | |
| "step": 350 | |
| }, | |
| { | |
| "epoch": 2.82, | |
| "eval_accuracy": 0.6545454545454545, | |
| "eval_loss": 1.0830607414245605, | |
| "eval_runtime": 8.4011, | |
| "eval_samples_per_second": 26.187, | |
| "eval_steps_per_second": 3.333, | |
| "step": 350 | |
| }, | |
| { | |
| "epoch": 2.9, | |
| "grad_norm": 3.6447389125823975, | |
| "learning_rate": 5.5645161290322576e-05, | |
| "loss": 0.2194, | |
| "step": 360 | |
| }, | |
| { | |
| "epoch": 2.98, | |
| "grad_norm": 4.36700439453125, | |
| "learning_rate": 5.161290322580645e-05, | |
| "loss": 0.1544, | |
| "step": 370 | |
| }, | |
| { | |
| "epoch": 3.02, | |
| "eval_accuracy": 0.6363636363636364, | |
| "eval_loss": 1.115280032157898, | |
| "eval_runtime": 7.5342, | |
| "eval_samples_per_second": 29.2, | |
| "eval_steps_per_second": 3.716, | |
| "step": 375 | |
| }, | |
| { | |
| "epoch": 3.06, | |
| "grad_norm": 0.32941731810569763, | |
| "learning_rate": 4.7580645161290326e-05, | |
| "loss": 0.0438, | |
| "step": 380 | |
| }, | |
| { | |
| "epoch": 3.15, | |
| "grad_norm": 0.40441280603408813, | |
| "learning_rate": 4.3548387096774194e-05, | |
| "loss": 0.0432, | |
| "step": 390 | |
| }, | |
| { | |
| "epoch": 3.23, | |
| "grad_norm": 0.3993894159793854, | |
| "learning_rate": 3.951612903225806e-05, | |
| "loss": 0.0803, | |
| "step": 400 | |
| }, | |
| { | |
| "epoch": 3.23, | |
| "eval_accuracy": 0.6409090909090909, | |
| "eval_loss": 1.039893388748169, | |
| "eval_runtime": 7.6391, | |
| "eval_samples_per_second": 28.799, | |
| "eval_steps_per_second": 3.665, | |
| "step": 400 | |
| }, | |
| { | |
| "epoch": 3.31, | |
| "grad_norm": 0.0989086702466011, | |
| "learning_rate": 3.548387096774194e-05, | |
| "loss": 0.0298, | |
| "step": 410 | |
| }, | |
| { | |
| "epoch": 3.39, | |
| "grad_norm": 0.30518239736557007, | |
| "learning_rate": 3.1451612903225806e-05, | |
| "loss": 0.041, | |
| "step": 420 | |
| }, | |
| { | |
| "epoch": 3.43, | |
| "eval_accuracy": 0.6818181818181818, | |
| "eval_loss": 1.0910764932632446, | |
| "eval_runtime": 7.8435, | |
| "eval_samples_per_second": 28.049, | |
| "eval_steps_per_second": 3.57, | |
| "step": 425 | |
| }, | |
| { | |
| "epoch": 3.47, | |
| "grad_norm": 0.383977472782135, | |
| "learning_rate": 2.7419354838709678e-05, | |
| "loss": 0.0702, | |
| "step": 430 | |
| }, | |
| { | |
| "epoch": 3.55, | |
| "grad_norm": 0.12995119392871857, | |
| "learning_rate": 2.338709677419355e-05, | |
| "loss": 0.0211, | |
| "step": 440 | |
| }, | |
| { | |
| "epoch": 3.63, | |
| "grad_norm": 0.09603479504585266, | |
| "learning_rate": 1.935483870967742e-05, | |
| "loss": 0.0685, | |
| "step": 450 | |
| }, | |
| { | |
| "epoch": 3.63, | |
| "eval_accuracy": 0.6590909090909091, | |
| "eval_loss": 1.1889867782592773, | |
| "eval_runtime": 7.401, | |
| "eval_samples_per_second": 29.726, | |
| "eval_steps_per_second": 3.783, | |
| "step": 450 | |
| }, | |
| { | |
| "epoch": 3.71, | |
| "grad_norm": 0.13990958034992218, | |
| "learning_rate": 1.5322580645161292e-05, | |
| "loss": 0.0406, | |
| "step": 460 | |
| }, | |
| { | |
| "epoch": 3.79, | |
| "grad_norm": 0.11549140512943268, | |
| "learning_rate": 1.129032258064516e-05, | |
| "loss": 0.0475, | |
| "step": 470 | |
| }, | |
| { | |
| "epoch": 3.83, | |
| "eval_accuracy": 0.6681818181818182, | |
| "eval_loss": 1.1724765300750732, | |
| "eval_runtime": 8.3023, | |
| "eval_samples_per_second": 26.499, | |
| "eval_steps_per_second": 3.373, | |
| "step": 475 | |
| }, | |
| { | |
| "epoch": 3.87, | |
| "grad_norm": 2.7816131114959717, | |
| "learning_rate": 7.258064516129033e-06, | |
| "loss": 0.0376, | |
| "step": 480 | |
| }, | |
| { | |
| "epoch": 3.95, | |
| "grad_norm": 0.09355681389570236, | |
| "learning_rate": 3.225806451612903e-06, | |
| "loss": 0.0452, | |
| "step": 490 | |
| }, | |
| { | |
| "epoch": 4.0, | |
| "step": 496, | |
| "total_flos": 6.137420584742093e+17, | |
| "train_loss": 0.3987359667737638, | |
| "train_runtime": 862.5404, | |
| "train_samples_per_second": 9.182, | |
| "train_steps_per_second": 0.575 | |
| } | |
| ], | |
| "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 | |
| } | |