| { | |
| "best_metric": 3.6932897567749023, | |
| "best_model_checkpoint": "models/GPT2_bigram_function_67/checkpoint-64200", | |
| "epoch": 10.0, | |
| "global_step": 64200, | |
| "is_hyper_param_search": false, | |
| "is_local_process_zero": true, | |
| "is_world_process_zero": true, | |
| "log_history": [ | |
| { | |
| "epoch": 0.16, | |
| "learning_rate": 1e-05, | |
| "loss": 7.7746, | |
| "step": 1000 | |
| }, | |
| { | |
| "epoch": 0.31, | |
| "learning_rate": 2e-05, | |
| "loss": 6.0981, | |
| "step": 2000 | |
| }, | |
| { | |
| "epoch": 0.47, | |
| "learning_rate": 3e-05, | |
| "loss": 5.6302, | |
| "step": 3000 | |
| }, | |
| { | |
| "epoch": 0.62, | |
| "learning_rate": 4e-05, | |
| "loss": 5.3269, | |
| "step": 4000 | |
| }, | |
| { | |
| "epoch": 0.78, | |
| "learning_rate": 5e-05, | |
| "loss": 5.0912, | |
| "step": 5000 | |
| }, | |
| { | |
| "epoch": 0.93, | |
| "learning_rate": 6e-05, | |
| "loss": 4.8987, | |
| "step": 6000 | |
| }, | |
| { | |
| "epoch": 1.0, | |
| "eval_accuracy": 0.25722452632352255, | |
| "eval_loss": 4.68282413482666, | |
| "eval_runtime": 1.9503, | |
| "eval_samples_per_second": 604.536, | |
| "eval_steps_per_second": 5.128, | |
| "step": 6420 | |
| }, | |
| { | |
| "epoch": 1.09, | |
| "learning_rate": 7e-05, | |
| "loss": 4.7305, | |
| "step": 7000 | |
| }, | |
| { | |
| "epoch": 1.25, | |
| "learning_rate": 8e-05, | |
| "loss": 4.6029, | |
| "step": 8000 | |
| }, | |
| { | |
| "epoch": 1.4, | |
| "learning_rate": 9e-05, | |
| "loss": 4.5041, | |
| "step": 9000 | |
| }, | |
| { | |
| "epoch": 1.56, | |
| "learning_rate": 0.0001, | |
| "loss": 4.4242, | |
| "step": 10000 | |
| }, | |
| { | |
| "epoch": 1.71, | |
| "learning_rate": 9.81568265682657e-05, | |
| "loss": 4.3514, | |
| "step": 11000 | |
| }, | |
| { | |
| "epoch": 1.87, | |
| "learning_rate": 9.631180811808118e-05, | |
| "loss": 4.2909, | |
| "step": 12000 | |
| }, | |
| { | |
| "epoch": 2.0, | |
| "eval_accuracy": 0.2995532047043738, | |
| "eval_loss": 4.165227890014648, | |
| "eval_runtime": 1.9589, | |
| "eval_samples_per_second": 601.855, | |
| "eval_steps_per_second": 5.105, | |
| "step": 12840 | |
| }, | |
| { | |
| "epoch": 2.02, | |
| "learning_rate": 9.446863468634687e-05, | |
| "loss": 4.2347, | |
| "step": 13000 | |
| }, | |
| { | |
| "epoch": 2.18, | |
| "learning_rate": 9.262361623616237e-05, | |
| "loss": 4.1618, | |
| "step": 14000 | |
| }, | |
| { | |
| "epoch": 2.34, | |
| "learning_rate": 9.078044280442805e-05, | |
| "loss": 4.1326, | |
| "step": 15000 | |
| }, | |
| { | |
| "epoch": 2.49, | |
| "learning_rate": 8.893542435424355e-05, | |
| "loss": 4.1068, | |
| "step": 16000 | |
| }, | |
| { | |
| "epoch": 2.65, | |
| "learning_rate": 8.709225092250924e-05, | |
| "loss": 4.0814, | |
| "step": 17000 | |
| }, | |
| { | |
| "epoch": 2.8, | |
| "learning_rate": 8.524723247232472e-05, | |
| "loss": 4.0586, | |
| "step": 18000 | |
| }, | |
| { | |
| "epoch": 2.96, | |
| "learning_rate": 8.340405904059041e-05, | |
| "loss": 4.0369, | |
| "step": 19000 | |
| }, | |
| { | |
| "epoch": 3.0, | |
| "eval_accuracy": 0.3198292961471419, | |
| "eval_loss": 3.976323127746582, | |
| "eval_runtime": 1.9592, | |
| "eval_samples_per_second": 601.767, | |
| "eval_steps_per_second": 5.104, | |
| "step": 19260 | |
| }, | |
| { | |
| "epoch": 3.12, | |
| "learning_rate": 8.155904059040591e-05, | |
| "loss": 3.9736, | |
| "step": 20000 | |
| }, | |
| { | |
| "epoch": 3.27, | |
| "learning_rate": 7.971586715867159e-05, | |
| "loss": 3.9515, | |
| "step": 21000 | |
| }, | |
| { | |
| "epoch": 3.43, | |
| "learning_rate": 7.787084870848709e-05, | |
| "loss": 3.9425, | |
| "step": 22000 | |
| }, | |
| { | |
| "epoch": 3.58, | |
| "learning_rate": 7.602767527675278e-05, | |
| "loss": 3.9328, | |
| "step": 23000 | |
| }, | |
| { | |
| "epoch": 3.74, | |
| "learning_rate": 7.418265682656826e-05, | |
| "loss": 3.9232, | |
| "step": 24000 | |
| }, | |
| { | |
| "epoch": 3.89, | |
| "learning_rate": 7.233948339483395e-05, | |
| "loss": 3.9112, | |
| "step": 25000 | |
| }, | |
| { | |
| "epoch": 4.0, | |
| "eval_accuracy": 0.33227144316884055, | |
| "eval_loss": 3.8776960372924805, | |
| "eval_runtime": 1.9587, | |
| "eval_samples_per_second": 601.943, | |
| "eval_steps_per_second": 5.106, | |
| "step": 25680 | |
| }, | |
| { | |
| "epoch": 4.05, | |
| "learning_rate": 7.049446494464945e-05, | |
| "loss": 3.8778, | |
| "step": 26000 | |
| }, | |
| { | |
| "epoch": 4.21, | |
| "learning_rate": 6.865129151291513e-05, | |
| "loss": 3.8334, | |
| "step": 27000 | |
| }, | |
| { | |
| "epoch": 4.36, | |
| "learning_rate": 6.680627306273063e-05, | |
| "loss": 3.8312, | |
| "step": 28000 | |
| }, | |
| { | |
| "epoch": 4.52, | |
| "learning_rate": 6.496309963099632e-05, | |
| "loss": 3.8288, | |
| "step": 29000 | |
| }, | |
| { | |
| "epoch": 4.67, | |
| "learning_rate": 6.311808118081182e-05, | |
| "loss": 3.8254, | |
| "step": 30000 | |
| }, | |
| { | |
| "epoch": 4.83, | |
| "learning_rate": 6.12749077490775e-05, | |
| "loss": 3.8183, | |
| "step": 31000 | |
| }, | |
| { | |
| "epoch": 4.98, | |
| "learning_rate": 5.942988929889299e-05, | |
| "loss": 3.8154, | |
| "step": 32000 | |
| }, | |
| { | |
| "epoch": 5.0, | |
| "eval_accuracy": 0.34073985026680825, | |
| "eval_loss": 3.8141610622406006, | |
| "eval_runtime": 1.9556, | |
| "eval_samples_per_second": 602.877, | |
| "eval_steps_per_second": 5.113, | |
| "step": 32100 | |
| }, | |
| { | |
| "epoch": 5.14, | |
| "learning_rate": 5.7584870848708496e-05, | |
| "loss": 3.7508, | |
| "step": 33000 | |
| }, | |
| { | |
| "epoch": 5.3, | |
| "learning_rate": 5.574169741697417e-05, | |
| "loss": 3.7485, | |
| "step": 34000 | |
| }, | |
| { | |
| "epoch": 5.45, | |
| "learning_rate": 5.389667896678967e-05, | |
| "loss": 3.7497, | |
| "step": 35000 | |
| }, | |
| { | |
| "epoch": 5.61, | |
| "learning_rate": 5.205350553505536e-05, | |
| "loss": 3.7491, | |
| "step": 36000 | |
| }, | |
| { | |
| "epoch": 5.76, | |
| "learning_rate": 5.020848708487085e-05, | |
| "loss": 3.7468, | |
| "step": 37000 | |
| }, | |
| { | |
| "epoch": 5.92, | |
| "learning_rate": 4.836531365313653e-05, | |
| "loss": 3.7425, | |
| "step": 38000 | |
| }, | |
| { | |
| "epoch": 6.0, | |
| "eval_accuracy": 0.3456485878196523, | |
| "eval_loss": 3.772080659866333, | |
| "eval_runtime": 1.9713, | |
| "eval_samples_per_second": 598.088, | |
| "eval_steps_per_second": 5.073, | |
| "step": 38520 | |
| }, | |
| { | |
| "epoch": 6.07, | |
| "learning_rate": 4.652029520295203e-05, | |
| "loss": 3.7086, | |
| "step": 39000 | |
| }, | |
| { | |
| "epoch": 6.23, | |
| "learning_rate": 4.4677121771217715e-05, | |
| "loss": 3.6799, | |
| "step": 40000 | |
| }, | |
| { | |
| "epoch": 6.39, | |
| "learning_rate": 4.283210332103321e-05, | |
| "loss": 3.6856, | |
| "step": 41000 | |
| }, | |
| { | |
| "epoch": 6.54, | |
| "learning_rate": 4.098892988929889e-05, | |
| "loss": 3.6867, | |
| "step": 42000 | |
| }, | |
| { | |
| "epoch": 6.7, | |
| "learning_rate": 3.9143911439114395e-05, | |
| "loss": 3.6849, | |
| "step": 43000 | |
| }, | |
| { | |
| "epoch": 6.85, | |
| "learning_rate": 3.7300738007380073e-05, | |
| "loss": 3.6848, | |
| "step": 44000 | |
| }, | |
| { | |
| "epoch": 7.0, | |
| "eval_accuracy": 0.3501632906573701, | |
| "eval_loss": 3.739995002746582, | |
| "eval_runtime": 1.9567, | |
| "eval_samples_per_second": 602.556, | |
| "eval_steps_per_second": 5.111, | |
| "step": 44940 | |
| }, | |
| { | |
| "epoch": 7.01, | |
| "learning_rate": 3.545571955719557e-05, | |
| "loss": 3.6786, | |
| "step": 45000 | |
| }, | |
| { | |
| "epoch": 7.17, | |
| "learning_rate": 3.3612546125461256e-05, | |
| "loss": 3.6233, | |
| "step": 46000 | |
| }, | |
| { | |
| "epoch": 7.32, | |
| "learning_rate": 3.1767527675276754e-05, | |
| "loss": 3.6291, | |
| "step": 47000 | |
| }, | |
| { | |
| "epoch": 7.48, | |
| "learning_rate": 2.9924354243542435e-05, | |
| "loss": 3.6318, | |
| "step": 48000 | |
| }, | |
| { | |
| "epoch": 7.63, | |
| "learning_rate": 2.8079335793357936e-05, | |
| "loss": 3.6314, | |
| "step": 49000 | |
| }, | |
| { | |
| "epoch": 7.79, | |
| "learning_rate": 2.6236162361623618e-05, | |
| "loss": 3.6327, | |
| "step": 50000 | |
| }, | |
| { | |
| "epoch": 7.94, | |
| "learning_rate": 2.4391143911439115e-05, | |
| "loss": 3.6311, | |
| "step": 51000 | |
| }, | |
| { | |
| "epoch": 8.0, | |
| "eval_accuracy": 0.3533756753688232, | |
| "eval_loss": 3.716982364654541, | |
| "eval_runtime": 1.956, | |
| "eval_samples_per_second": 602.749, | |
| "eval_steps_per_second": 5.112, | |
| "step": 51360 | |
| }, | |
| { | |
| "epoch": 8.1, | |
| "learning_rate": 2.2546125461254613e-05, | |
| "loss": 3.597, | |
| "step": 52000 | |
| }, | |
| { | |
| "epoch": 8.26, | |
| "learning_rate": 2.0702952029520294e-05, | |
| "loss": 3.5812, | |
| "step": 53000 | |
| }, | |
| { | |
| "epoch": 8.41, | |
| "learning_rate": 1.8857933579335792e-05, | |
| "loss": 3.5837, | |
| "step": 54000 | |
| }, | |
| { | |
| "epoch": 8.57, | |
| "learning_rate": 1.7014760147601477e-05, | |
| "loss": 3.5859, | |
| "step": 55000 | |
| }, | |
| { | |
| "epoch": 8.72, | |
| "learning_rate": 1.5169741697416975e-05, | |
| "loss": 3.5863, | |
| "step": 56000 | |
| }, | |
| { | |
| "epoch": 8.88, | |
| "learning_rate": 1.3326568265682656e-05, | |
| "loss": 3.5851, | |
| "step": 57000 | |
| }, | |
| { | |
| "epoch": 9.0, | |
| "eval_accuracy": 0.35644781043590923, | |
| "eval_loss": 3.7016589641571045, | |
| "eval_runtime": 1.9641, | |
| "eval_samples_per_second": 600.29, | |
| "eval_steps_per_second": 5.092, | |
| "step": 57780 | |
| }, | |
| { | |
| "epoch": 9.03, | |
| "learning_rate": 1.1481549815498155e-05, | |
| "loss": 3.5738, | |
| "step": 58000 | |
| }, | |
| { | |
| "epoch": 9.19, | |
| "learning_rate": 9.638376383763839e-06, | |
| "loss": 3.5446, | |
| "step": 59000 | |
| }, | |
| { | |
| "epoch": 9.35, | |
| "learning_rate": 7.793357933579336e-06, | |
| "loss": 3.545, | |
| "step": 60000 | |
| }, | |
| { | |
| "epoch": 9.5, | |
| "learning_rate": 5.9501845018450185e-06, | |
| "loss": 3.5453, | |
| "step": 61000 | |
| }, | |
| { | |
| "epoch": 9.66, | |
| "learning_rate": 4.105166051660517e-06, | |
| "loss": 3.5478, | |
| "step": 62000 | |
| }, | |
| { | |
| "epoch": 9.81, | |
| "learning_rate": 2.2601476014760147e-06, | |
| "loss": 3.5458, | |
| "step": 63000 | |
| }, | |
| { | |
| "epoch": 9.97, | |
| "learning_rate": 4.1697416974169745e-07, | |
| "loss": 3.5457, | |
| "step": 64000 | |
| }, | |
| { | |
| "epoch": 10.0, | |
| "eval_accuracy": 0.3575364148183767, | |
| "eval_loss": 3.6932897567749023, | |
| "eval_runtime": 1.9569, | |
| "eval_samples_per_second": 602.483, | |
| "eval_steps_per_second": 5.11, | |
| "step": 64200 | |
| }, | |
| { | |
| "epoch": 10.0, | |
| "step": 64200, | |
| "total_flos": 5.367983505408e+17, | |
| "train_loss": 4.0210844616280905, | |
| "train_runtime": 28109.1932, | |
| "train_samples_per_second": 292.346, | |
| "train_steps_per_second": 2.284 | |
| } | |
| ], | |
| "max_steps": 64200, | |
| "num_train_epochs": 10, | |
| "total_flos": 5.367983505408e+17, | |
| "trial_name": null, | |
| "trial_params": null | |
| } | |