| { | |
| "best_metric": null, | |
| "best_model_checkpoint": null, | |
| "epoch": 0.9990491283676703, | |
| "global_step": 788, | |
| "is_hyper_param_search": false, | |
| "is_local_process_zero": true, | |
| "is_world_process_zero": true, | |
| "log_history": [ | |
| { | |
| "epoch": 0.01, | |
| "learning_rate": 8.859191006777897e-06, | |
| "loss": 14.8557, | |
| "step": 10 | |
| }, | |
| { | |
| "epoch": 0.03, | |
| "learning_rate": 1e-05, | |
| "loss": 12.1896, | |
| "step": 20 | |
| }, | |
| { | |
| "epoch": 0.04, | |
| "learning_rate": 1e-05, | |
| "loss": 11.6323, | |
| "step": 30 | |
| }, | |
| { | |
| "epoch": 0.05, | |
| "learning_rate": 1e-05, | |
| "loss": 10.8152, | |
| "step": 40 | |
| }, | |
| { | |
| "epoch": 0.06, | |
| "learning_rate": 1e-05, | |
| "loss": 10.798, | |
| "step": 50 | |
| }, | |
| { | |
| "epoch": 0.08, | |
| "learning_rate": 1e-05, | |
| "loss": 10.755, | |
| "step": 60 | |
| }, | |
| { | |
| "epoch": 0.09, | |
| "learning_rate": 1e-05, | |
| "loss": 10.4725, | |
| "step": 70 | |
| }, | |
| { | |
| "epoch": 0.1, | |
| "learning_rate": 1e-05, | |
| "loss": 10.4207, | |
| "step": 80 | |
| }, | |
| { | |
| "epoch": 0.11, | |
| "learning_rate": 1e-05, | |
| "loss": 11.1635, | |
| "step": 90 | |
| }, | |
| { | |
| "epoch": 0.13, | |
| "learning_rate": 1e-05, | |
| "loss": 10.207, | |
| "step": 100 | |
| }, | |
| { | |
| "epoch": 0.14, | |
| "learning_rate": 1e-05, | |
| "loss": 10.2941, | |
| "step": 110 | |
| }, | |
| { | |
| "epoch": 0.15, | |
| "learning_rate": 1e-05, | |
| "loss": 9.577, | |
| "step": 120 | |
| }, | |
| { | |
| "epoch": 0.16, | |
| "learning_rate": 1e-05, | |
| "loss": 10.3416, | |
| "step": 130 | |
| }, | |
| { | |
| "epoch": 0.18, | |
| "learning_rate": 1e-05, | |
| "loss": 9.3784, | |
| "step": 140 | |
| }, | |
| { | |
| "epoch": 0.19, | |
| "learning_rate": 1e-05, | |
| "loss": 10.477, | |
| "step": 150 | |
| }, | |
| { | |
| "epoch": 0.2, | |
| "learning_rate": 1e-05, | |
| "loss": 9.5938, | |
| "step": 160 | |
| }, | |
| { | |
| "epoch": 0.22, | |
| "learning_rate": 1e-05, | |
| "loss": 9.6989, | |
| "step": 170 | |
| }, | |
| { | |
| "epoch": 0.23, | |
| "learning_rate": 1e-05, | |
| "loss": 9.4838, | |
| "step": 180 | |
| }, | |
| { | |
| "epoch": 0.24, | |
| "learning_rate": 1e-05, | |
| "loss": 9.1456, | |
| "step": 190 | |
| }, | |
| { | |
| "epoch": 0.25, | |
| "learning_rate": 1e-05, | |
| "loss": 9.9256, | |
| "step": 200 | |
| }, | |
| { | |
| "epoch": 0.27, | |
| "learning_rate": 1e-05, | |
| "loss": 9.5248, | |
| "step": 210 | |
| }, | |
| { | |
| "epoch": 0.28, | |
| "learning_rate": 1e-05, | |
| "loss": 8.8531, | |
| "step": 220 | |
| }, | |
| { | |
| "epoch": 0.29, | |
| "learning_rate": 1e-05, | |
| "loss": 8.4109, | |
| "step": 230 | |
| }, | |
| { | |
| "epoch": 0.3, | |
| "learning_rate": 1e-05, | |
| "loss": 8.7742, | |
| "step": 240 | |
| }, | |
| { | |
| "epoch": 0.32, | |
| "learning_rate": 1e-05, | |
| "loss": 8.8604, | |
| "step": 250 | |
| }, | |
| { | |
| "epoch": 0.33, | |
| "learning_rate": 1e-05, | |
| "loss": 9.6211, | |
| "step": 260 | |
| }, | |
| { | |
| "epoch": 0.34, | |
| "learning_rate": 1e-05, | |
| "loss": 9.4113, | |
| "step": 270 | |
| }, | |
| { | |
| "epoch": 0.35, | |
| "learning_rate": 1e-05, | |
| "loss": 9.6674, | |
| "step": 280 | |
| }, | |
| { | |
| "epoch": 0.37, | |
| "learning_rate": 1e-05, | |
| "loss": 8.721, | |
| "step": 290 | |
| }, | |
| { | |
| "epoch": 0.38, | |
| "learning_rate": 1e-05, | |
| "loss": 8.8226, | |
| "step": 300 | |
| }, | |
| { | |
| "epoch": 0.39, | |
| "learning_rate": 1e-05, | |
| "loss": 9.2014, | |
| "step": 310 | |
| }, | |
| { | |
| "epoch": 0.41, | |
| "learning_rate": 1e-05, | |
| "loss": 8.5717, | |
| "step": 320 | |
| }, | |
| { | |
| "epoch": 0.42, | |
| "learning_rate": 1e-05, | |
| "loss": 9.2165, | |
| "step": 330 | |
| }, | |
| { | |
| "epoch": 0.43, | |
| "learning_rate": 1e-05, | |
| "loss": 7.5827, | |
| "step": 340 | |
| }, | |
| { | |
| "epoch": 0.44, | |
| "learning_rate": 1e-05, | |
| "loss": 8.3725, | |
| "step": 350 | |
| }, | |
| { | |
| "epoch": 0.46, | |
| "learning_rate": 1e-05, | |
| "loss": 8.4591, | |
| "step": 360 | |
| }, | |
| { | |
| "epoch": 0.47, | |
| "learning_rate": 1e-05, | |
| "loss": 8.382, | |
| "step": 370 | |
| }, | |
| { | |
| "epoch": 0.48, | |
| "learning_rate": 1e-05, | |
| "loss": 8.5894, | |
| "step": 380 | |
| }, | |
| { | |
| "epoch": 0.49, | |
| "learning_rate": 1e-05, | |
| "loss": 8.7015, | |
| "step": 390 | |
| }, | |
| { | |
| "epoch": 0.51, | |
| "learning_rate": 1e-05, | |
| "loss": 8.7413, | |
| "step": 400 | |
| }, | |
| { | |
| "epoch": 0.52, | |
| "learning_rate": 1e-05, | |
| "loss": 8.5898, | |
| "step": 410 | |
| }, | |
| { | |
| "epoch": 0.53, | |
| "learning_rate": 1e-05, | |
| "loss": 8.1689, | |
| "step": 420 | |
| }, | |
| { | |
| "epoch": 0.55, | |
| "learning_rate": 1e-05, | |
| "loss": 8.2817, | |
| "step": 430 | |
| }, | |
| { | |
| "epoch": 0.56, | |
| "learning_rate": 1e-05, | |
| "loss": 8.5049, | |
| "step": 440 | |
| }, | |
| { | |
| "epoch": 0.57, | |
| "learning_rate": 1e-05, | |
| "loss": 7.9563, | |
| "step": 450 | |
| }, | |
| { | |
| "epoch": 0.58, | |
| "learning_rate": 1e-05, | |
| "loss": 7.0652, | |
| "step": 460 | |
| }, | |
| { | |
| "epoch": 0.6, | |
| "learning_rate": 1e-05, | |
| "loss": 8.2745, | |
| "step": 470 | |
| }, | |
| { | |
| "epoch": 0.61, | |
| "learning_rate": 1e-05, | |
| "loss": 8.1843, | |
| "step": 480 | |
| }, | |
| { | |
| "epoch": 0.62, | |
| "learning_rate": 1e-05, | |
| "loss": 7.5988, | |
| "step": 490 | |
| }, | |
| { | |
| "epoch": 0.63, | |
| "learning_rate": 1e-05, | |
| "loss": 8.1985, | |
| "step": 500 | |
| }, | |
| { | |
| "epoch": 0.65, | |
| "learning_rate": 1e-05, | |
| "loss": 7.6444, | |
| "step": 510 | |
| }, | |
| { | |
| "epoch": 0.66, | |
| "learning_rate": 1e-05, | |
| "loss": 8.0742, | |
| "step": 520 | |
| }, | |
| { | |
| "epoch": 0.67, | |
| "learning_rate": 1e-05, | |
| "loss": 7.8268, | |
| "step": 530 | |
| }, | |
| { | |
| "epoch": 0.68, | |
| "learning_rate": 1e-05, | |
| "loss": 7.5661, | |
| "step": 540 | |
| }, | |
| { | |
| "epoch": 0.7, | |
| "learning_rate": 1e-05, | |
| "loss": 7.4674, | |
| "step": 550 | |
| }, | |
| { | |
| "epoch": 0.71, | |
| "learning_rate": 1e-05, | |
| "loss": 7.1126, | |
| "step": 560 | |
| }, | |
| { | |
| "epoch": 0.72, | |
| "learning_rate": 1e-05, | |
| "loss": 7.6766, | |
| "step": 570 | |
| }, | |
| { | |
| "epoch": 0.74, | |
| "learning_rate": 1e-05, | |
| "loss": 8.5397, | |
| "step": 580 | |
| }, | |
| { | |
| "epoch": 0.75, | |
| "learning_rate": 1e-05, | |
| "loss": 8.5168, | |
| "step": 590 | |
| }, | |
| { | |
| "epoch": 0.76, | |
| "learning_rate": 1e-05, | |
| "loss": 8.0453, | |
| "step": 600 | |
| }, | |
| { | |
| "epoch": 0.77, | |
| "learning_rate": 1e-05, | |
| "loss": 8.4384, | |
| "step": 610 | |
| }, | |
| { | |
| "epoch": 0.79, | |
| "learning_rate": 1e-05, | |
| "loss": 6.8534, | |
| "step": 620 | |
| }, | |
| { | |
| "epoch": 0.8, | |
| "learning_rate": 1e-05, | |
| "loss": 7.0237, | |
| "step": 630 | |
| }, | |
| { | |
| "epoch": 0.81, | |
| "learning_rate": 1e-05, | |
| "loss": 7.7834, | |
| "step": 640 | |
| }, | |
| { | |
| "epoch": 0.82, | |
| "learning_rate": 1e-05, | |
| "loss": 7.5957, | |
| "step": 650 | |
| }, | |
| { | |
| "epoch": 0.84, | |
| "learning_rate": 1e-05, | |
| "loss": 7.3629, | |
| "step": 660 | |
| }, | |
| { | |
| "epoch": 0.85, | |
| "learning_rate": 1e-05, | |
| "loss": 7.3123, | |
| "step": 670 | |
| }, | |
| { | |
| "epoch": 0.86, | |
| "learning_rate": 1e-05, | |
| "loss": 8.2057, | |
| "step": 680 | |
| }, | |
| { | |
| "epoch": 0.87, | |
| "learning_rate": 1e-05, | |
| "loss": 7.4482, | |
| "step": 690 | |
| }, | |
| { | |
| "epoch": 0.89, | |
| "learning_rate": 1e-05, | |
| "loss": 8.1283, | |
| "step": 700 | |
| }, | |
| { | |
| "epoch": 0.9, | |
| "learning_rate": 1e-05, | |
| "loss": 8.3028, | |
| "step": 710 | |
| }, | |
| { | |
| "epoch": 0.91, | |
| "learning_rate": 1e-05, | |
| "loss": 7.2104, | |
| "step": 720 | |
| }, | |
| { | |
| "epoch": 0.93, | |
| "learning_rate": 1e-05, | |
| "loss": 9.4241, | |
| "step": 730 | |
| }, | |
| { | |
| "epoch": 0.94, | |
| "learning_rate": 1e-05, | |
| "loss": 8.0627, | |
| "step": 740 | |
| }, | |
| { | |
| "epoch": 0.95, | |
| "learning_rate": 1e-05, | |
| "loss": 8.662, | |
| "step": 750 | |
| }, | |
| { | |
| "epoch": 0.96, | |
| "learning_rate": 1e-05, | |
| "loss": 7.3603, | |
| "step": 760 | |
| }, | |
| { | |
| "epoch": 0.98, | |
| "learning_rate": 1e-05, | |
| "loss": 7.3165, | |
| "step": 770 | |
| }, | |
| { | |
| "epoch": 0.99, | |
| "learning_rate": 1e-05, | |
| "loss": 7.4981, | |
| "step": 780 | |
| } | |
| ], | |
| "max_steps": 788, | |
| "num_train_epochs": 1, | |
| "total_flos": 0.0, | |
| "trial_name": null, | |
| "trial_params": null | |
| } | |