| { | |
| "best_metric": 91.12657677250978, | |
| "best_model_checkpoint": "./checkpoint-600", | |
| "epoch": 3.0025, | |
| "eval_steps": 200, | |
| "global_step": 1000, | |
| "is_hyper_param_search": false, | |
| "is_local_process_zero": true, | |
| "is_world_process_zero": true, | |
| "log_history": [ | |
| { | |
| "epoch": 0.03, | |
| "learning_rate": 6e-08, | |
| "loss": 0.6521, | |
| "step": 25 | |
| }, | |
| { | |
| "epoch": 0.05, | |
| "learning_rate": 1.2e-07, | |
| "loss": 0.6708, | |
| "step": 50 | |
| }, | |
| { | |
| "epoch": 0.07, | |
| "learning_rate": 1.825e-07, | |
| "loss": 0.6688, | |
| "step": 75 | |
| }, | |
| { | |
| "epoch": 0.1, | |
| "learning_rate": 2.45e-07, | |
| "loss": 0.7849, | |
| "step": 100 | |
| }, | |
| { | |
| "epoch": 0.12, | |
| "learning_rate": 3.0749999999999997e-07, | |
| "loss": 0.6421, | |
| "step": 125 | |
| }, | |
| { | |
| "epoch": 0.15, | |
| "learning_rate": 3.7e-07, | |
| "loss": 0.717, | |
| "step": 150 | |
| }, | |
| { | |
| "epoch": 0.17, | |
| "learning_rate": 4.325e-07, | |
| "loss": 0.786, | |
| "step": 175 | |
| }, | |
| { | |
| "epoch": 0.2, | |
| "learning_rate": 4.95e-07, | |
| "loss": 0.6689, | |
| "step": 200 | |
| }, | |
| { | |
| "epoch": 0.2, | |
| "eval_loss": 0.5854436159133911, | |
| "eval_runtime": 839.3873, | |
| "eval_samples_per_second": 12.659, | |
| "eval_steps_per_second": 0.397, | |
| "eval_wer": 91.63114397564158, | |
| "step": 200 | |
| }, | |
| { | |
| "epoch": 0.23, | |
| "learning_rate": 4.856249999999999e-07, | |
| "loss": 0.4781, | |
| "step": 225 | |
| }, | |
| { | |
| "epoch": 0.25, | |
| "learning_rate": 4.6999999999999995e-07, | |
| "loss": 0.5582, | |
| "step": 250 | |
| }, | |
| { | |
| "epoch": 0.28, | |
| "learning_rate": 4.5437499999999994e-07, | |
| "loss": 0.6674, | |
| "step": 275 | |
| }, | |
| { | |
| "epoch": 0.3, | |
| "learning_rate": 4.3874999999999993e-07, | |
| "loss": 0.606, | |
| "step": 300 | |
| }, | |
| { | |
| "epoch": 0.33, | |
| "learning_rate": 4.2312499999999997e-07, | |
| "loss": 0.6115, | |
| "step": 325 | |
| }, | |
| { | |
| "epoch": 1.02, | |
| "learning_rate": 4.0749999999999996e-07, | |
| "loss": 0.565, | |
| "step": 350 | |
| }, | |
| { | |
| "epoch": 1.04, | |
| "learning_rate": 3.9187499999999995e-07, | |
| "loss": 0.704, | |
| "step": 375 | |
| }, | |
| { | |
| "epoch": 1.07, | |
| "learning_rate": 3.7624999999999994e-07, | |
| "loss": 0.6314, | |
| "step": 400 | |
| }, | |
| { | |
| "epoch": 1.07, | |
| "eval_loss": 0.5791469216346741, | |
| "eval_runtime": 835.5799, | |
| "eval_samples_per_second": 12.717, | |
| "eval_steps_per_second": 0.399, | |
| "eval_wer": 91.17877337973032, | |
| "step": 400 | |
| }, | |
| { | |
| "epoch": 1.09, | |
| "learning_rate": 3.6062499999999994e-07, | |
| "loss": 0.7393, | |
| "step": 425 | |
| }, | |
| { | |
| "epoch": 1.12, | |
| "learning_rate": 3.45e-07, | |
| "loss": 0.6695, | |
| "step": 450 | |
| }, | |
| { | |
| "epoch": 1.14, | |
| "learning_rate": 3.2937499999999997e-07, | |
| "loss": 0.7078, | |
| "step": 475 | |
| }, | |
| { | |
| "epoch": 1.17, | |
| "learning_rate": 3.1374999999999996e-07, | |
| "loss": 0.7056, | |
| "step": 500 | |
| }, | |
| { | |
| "epoch": 1.19, | |
| "learning_rate": 2.9812499999999995e-07, | |
| "loss": 0.7321, | |
| "step": 525 | |
| }, | |
| { | |
| "epoch": 1.22, | |
| "learning_rate": 2.8249999999999994e-07, | |
| "loss": 0.5089, | |
| "step": 550 | |
| }, | |
| { | |
| "epoch": 1.24, | |
| "learning_rate": 2.66875e-07, | |
| "loss": 0.502, | |
| "step": 575 | |
| }, | |
| { | |
| "epoch": 1.27, | |
| "learning_rate": 2.5125e-07, | |
| "loss": 0.653, | |
| "step": 600 | |
| }, | |
| { | |
| "epoch": 1.27, | |
| "eval_loss": 0.5759375691413879, | |
| "eval_runtime": 857.6045, | |
| "eval_samples_per_second": 12.39, | |
| "eval_steps_per_second": 0.388, | |
| "eval_wer": 91.12657677250978, | |
| "step": 600 | |
| }, | |
| { | |
| "epoch": 1.29, | |
| "learning_rate": 2.35625e-07, | |
| "loss": 0.5852, | |
| "step": 625 | |
| }, | |
| { | |
| "epoch": 1.32, | |
| "learning_rate": 2.1999999999999998e-07, | |
| "loss": 0.5856, | |
| "step": 650 | |
| }, | |
| { | |
| "epoch": 2.01, | |
| "learning_rate": 2.04375e-07, | |
| "loss": 0.5815, | |
| "step": 675 | |
| }, | |
| { | |
| "epoch": 2.04, | |
| "learning_rate": 1.8875e-07, | |
| "loss": 0.706, | |
| "step": 700 | |
| }, | |
| { | |
| "epoch": 2.06, | |
| "learning_rate": 1.7312499999999998e-07, | |
| "loss": 0.587, | |
| "step": 725 | |
| }, | |
| { | |
| "epoch": 2.08, | |
| "learning_rate": 1.575e-07, | |
| "loss": 0.7652, | |
| "step": 750 | |
| }, | |
| { | |
| "epoch": 2.11, | |
| "learning_rate": 1.41875e-07, | |
| "loss": 0.6404, | |
| "step": 775 | |
| }, | |
| { | |
| "epoch": 2.13, | |
| "learning_rate": 1.2624999999999998e-07, | |
| "loss": 0.699, | |
| "step": 800 | |
| }, | |
| { | |
| "epoch": 2.13, | |
| "eval_loss": 0.5749017596244812, | |
| "eval_runtime": 839.408, | |
| "eval_samples_per_second": 12.659, | |
| "eval_steps_per_second": 0.397, | |
| "eval_wer": 91.20487168334058, | |
| "step": 800 | |
| }, | |
| { | |
| "epoch": 2.16, | |
| "learning_rate": 1.10625e-07, | |
| "loss": 0.6785, | |
| "step": 825 | |
| }, | |
| { | |
| "epoch": 2.19, | |
| "learning_rate": 9.499999999999999e-08, | |
| "loss": 0.7806, | |
| "step": 850 | |
| }, | |
| { | |
| "epoch": 2.21, | |
| "learning_rate": 7.9375e-08, | |
| "loss": 0.5384, | |
| "step": 875 | |
| }, | |
| { | |
| "epoch": 2.23, | |
| "learning_rate": 6.375e-08, | |
| "loss": 0.4755, | |
| "step": 900 | |
| }, | |
| { | |
| "epoch": 2.26, | |
| "learning_rate": 4.8124999999999996e-08, | |
| "loss": 0.623, | |
| "step": 925 | |
| }, | |
| { | |
| "epoch": 2.29, | |
| "learning_rate": 3.25e-08, | |
| "loss": 0.6052, | |
| "step": 950 | |
| }, | |
| { | |
| "epoch": 2.31, | |
| "learning_rate": 1.6875e-08, | |
| "loss": 0.5875, | |
| "step": 975 | |
| }, | |
| { | |
| "epoch": 3.0, | |
| "learning_rate": 1.25e-09, | |
| "loss": 0.5613, | |
| "step": 1000 | |
| }, | |
| { | |
| "epoch": 3.0, | |
| "eval_loss": 0.5744211077690125, | |
| "eval_runtime": 835.9541, | |
| "eval_samples_per_second": 12.711, | |
| "eval_steps_per_second": 0.398, | |
| "eval_wer": 91.15267507612005, | |
| "step": 1000 | |
| }, | |
| { | |
| "epoch": 3.0, | |
| "step": 1000, | |
| "total_flos": 1.57354066132992e+18, | |
| "train_loss": 0.640758861541748, | |
| "train_runtime": 6192.7747, | |
| "train_samples_per_second": 10.335, | |
| "train_steps_per_second": 0.161 | |
| } | |
| ], | |
| "logging_steps": 25, | |
| "max_steps": 1000, | |
| "num_input_tokens_seen": 0, | |
| "num_train_epochs": 9223372036854775807, | |
| "save_steps": 200, | |
| "total_flos": 1.57354066132992e+18, | |
| "train_batch_size": 32, | |
| "trial_name": null, | |
| "trial_params": null | |
| } | |