| { | |
| "best_metric": null, | |
| "best_model_checkpoint": null, | |
| "epoch": 75.0, | |
| "eval_steps": 100, | |
| "global_step": 2100, | |
| "is_hyper_param_search": false, | |
| "is_local_process_zero": true, | |
| "is_world_process_zero": true, | |
| "log_history": [ | |
| { | |
| "epoch": 3.57, | |
| "learning_rate": 2.5e-05, | |
| "loss": 8.1262, | |
| "step": 100 | |
| }, | |
| { | |
| "epoch": 3.57, | |
| "eval_loss": 4.0765790939331055, | |
| "eval_runtime": 21.8341, | |
| "eval_samples_per_second": 18.091, | |
| "eval_steps_per_second": 0.595, | |
| "eval_wer": 1.0, | |
| "step": 100 | |
| }, | |
| { | |
| "epoch": 7.14, | |
| "learning_rate": 5e-05, | |
| "loss": 3.5989, | |
| "step": 200 | |
| }, | |
| { | |
| "epoch": 7.14, | |
| "eval_loss": 3.226550340652466, | |
| "eval_runtime": 21.5714, | |
| "eval_samples_per_second": 18.311, | |
| "eval_steps_per_second": 0.603, | |
| "eval_wer": 1.0, | |
| "step": 200 | |
| }, | |
| { | |
| "epoch": 10.71, | |
| "learning_rate": 7.5e-05, | |
| "loss": 3.0472, | |
| "step": 300 | |
| }, | |
| { | |
| "epoch": 10.71, | |
| "eval_loss": 2.8889501094818115, | |
| "eval_runtime": 21.7358, | |
| "eval_samples_per_second": 18.173, | |
| "eval_steps_per_second": 0.598, | |
| "eval_wer": 1.0, | |
| "step": 300 | |
| }, | |
| { | |
| "epoch": 14.29, | |
| "learning_rate": 0.0001, | |
| "loss": 1.7327, | |
| "step": 400 | |
| }, | |
| { | |
| "epoch": 14.29, | |
| "eval_loss": 0.6693879961967468, | |
| "eval_runtime": 22.0147, | |
| "eval_samples_per_second": 17.943, | |
| "eval_steps_per_second": 0.591, | |
| "eval_wer": 0.6622969506982046, | |
| "step": 400 | |
| }, | |
| { | |
| "epoch": 17.86, | |
| "learning_rate": 0.000125, | |
| "loss": 0.4239, | |
| "step": 500 | |
| }, | |
| { | |
| "epoch": 17.86, | |
| "eval_loss": 0.4327137768268585, | |
| "eval_runtime": 21.6438, | |
| "eval_samples_per_second": 18.25, | |
| "eval_steps_per_second": 0.601, | |
| "eval_wer": 0.5483043602165859, | |
| "step": 500 | |
| }, | |
| { | |
| "epoch": 21.43, | |
| "learning_rate": 0.00015, | |
| "loss": 0.2337, | |
| "step": 600 | |
| }, | |
| { | |
| "epoch": 21.43, | |
| "eval_loss": 0.38920775055885315, | |
| "eval_runtime": 21.8228, | |
| "eval_samples_per_second": 18.1, | |
| "eval_steps_per_second": 0.596, | |
| "eval_wer": 0.5195212311199772, | |
| "step": 600 | |
| }, | |
| { | |
| "epoch": 25.0, | |
| "learning_rate": 0.000175, | |
| "loss": 0.1724, | |
| "step": 700 | |
| }, | |
| { | |
| "epoch": 25.0, | |
| "eval_loss": 0.391787052154541, | |
| "eval_runtime": 21.5976, | |
| "eval_samples_per_second": 18.289, | |
| "eval_steps_per_second": 0.602, | |
| "eval_wer": 0.5027073240239385, | |
| "step": 700 | |
| }, | |
| { | |
| "epoch": 28.57, | |
| "learning_rate": 0.0002, | |
| "loss": 0.1392, | |
| "step": 800 | |
| }, | |
| { | |
| "epoch": 28.57, | |
| "eval_loss": 0.36863410472869873, | |
| "eval_runtime": 21.5216, | |
| "eval_samples_per_second": 18.354, | |
| "eval_steps_per_second": 0.604, | |
| "eval_wer": 0.4670846394984326, | |
| "step": 800 | |
| }, | |
| { | |
| "epoch": 32.14, | |
| "learning_rate": 0.00022500000000000002, | |
| "loss": 0.1196, | |
| "step": 900 | |
| }, | |
| { | |
| "epoch": 32.14, | |
| "eval_loss": 0.3849872052669525, | |
| "eval_runtime": 21.7525, | |
| "eval_samples_per_second": 18.159, | |
| "eval_steps_per_second": 0.598, | |
| "eval_wer": 0.45397549159304645, | |
| "step": 900 | |
| }, | |
| { | |
| "epoch": 35.71, | |
| "learning_rate": 0.00025, | |
| "loss": 0.1095, | |
| "step": 1000 | |
| }, | |
| { | |
| "epoch": 35.71, | |
| "eval_loss": 0.3586702346801758, | |
| "eval_runtime": 21.8949, | |
| "eval_samples_per_second": 18.041, | |
| "eval_steps_per_second": 0.594, | |
| "eval_wer": 0.4516956397834141, | |
| "step": 1000 | |
| }, | |
| { | |
| "epoch": 39.29, | |
| "learning_rate": 0.000275, | |
| "loss": 0.1026, | |
| "step": 1100 | |
| }, | |
| { | |
| "epoch": 39.29, | |
| "eval_loss": 0.36062636971473694, | |
| "eval_runtime": 21.8096, | |
| "eval_samples_per_second": 18.111, | |
| "eval_steps_per_second": 0.596, | |
| "eval_wer": 0.44029638073525224, | |
| "step": 1100 | |
| }, | |
| { | |
| "epoch": 42.86, | |
| "learning_rate": 0.0003, | |
| "loss": 0.0972, | |
| "step": 1200 | |
| }, | |
| { | |
| "epoch": 42.86, | |
| "eval_loss": 0.40814533829689026, | |
| "eval_runtime": 21.9118, | |
| "eval_samples_per_second": 18.027, | |
| "eval_steps_per_second": 0.593, | |
| "eval_wer": 0.43317184383015106, | |
| "step": 1200 | |
| }, | |
| { | |
| "epoch": 46.43, | |
| "learning_rate": 0.00032500000000000004, | |
| "loss": 0.0932, | |
| "step": 1300 | |
| }, | |
| { | |
| "epoch": 46.43, | |
| "eval_loss": 0.3904629945755005, | |
| "eval_runtime": 21.7505, | |
| "eval_samples_per_second": 18.16, | |
| "eval_steps_per_second": 0.598, | |
| "eval_wer": 0.426902251353662, | |
| "step": 1300 | |
| }, | |
| { | |
| "epoch": 50.0, | |
| "learning_rate": 0.00035, | |
| "loss": 0.0897, | |
| "step": 1400 | |
| }, | |
| { | |
| "epoch": 50.0, | |
| "eval_loss": 0.37547701597213745, | |
| "eval_runtime": 22.1102, | |
| "eval_samples_per_second": 17.865, | |
| "eval_steps_per_second": 0.588, | |
| "eval_wer": 0.42747221430607013, | |
| "step": 1400 | |
| }, | |
| { | |
| "epoch": 53.57, | |
| "learning_rate": 0.000375, | |
| "loss": 0.0846, | |
| "step": 1500 | |
| }, | |
| { | |
| "epoch": 53.57, | |
| "eval_loss": 0.36823779344558716, | |
| "eval_runtime": 21.896, | |
| "eval_samples_per_second": 18.04, | |
| "eval_steps_per_second": 0.594, | |
| "eval_wer": 0.42091764035337703, | |
| "step": 1500 | |
| }, | |
| { | |
| "epoch": 57.14, | |
| "learning_rate": 0.0004, | |
| "loss": 0.0854, | |
| "step": 1600 | |
| }, | |
| { | |
| "epoch": 57.14, | |
| "eval_loss": 0.3795730471611023, | |
| "eval_runtime": 21.5795, | |
| "eval_samples_per_second": 18.304, | |
| "eval_steps_per_second": 0.602, | |
| "eval_wer": 0.41635793673411226, | |
| "step": 1600 | |
| }, | |
| { | |
| "epoch": 60.71, | |
| "learning_rate": 0.000425, | |
| "loss": 0.0845, | |
| "step": 1700 | |
| }, | |
| { | |
| "epoch": 60.71, | |
| "eval_loss": 0.35857513546943665, | |
| "eval_runtime": 21.7949, | |
| "eval_samples_per_second": 18.124, | |
| "eval_steps_per_second": 0.596, | |
| "eval_wer": 0.3941293815901966, | |
| "step": 1700 | |
| }, | |
| { | |
| "epoch": 64.29, | |
| "learning_rate": 0.00045000000000000004, | |
| "loss": 0.0854, | |
| "step": 1800 | |
| }, | |
| { | |
| "epoch": 64.29, | |
| "eval_loss": 0.376005083322525, | |
| "eval_runtime": 21.5391, | |
| "eval_samples_per_second": 18.339, | |
| "eval_steps_per_second": 0.604, | |
| "eval_wer": 0.41664291821031635, | |
| "step": 1800 | |
| }, | |
| { | |
| "epoch": 67.86, | |
| "learning_rate": 0.000475, | |
| "loss": 0.0846, | |
| "step": 1900 | |
| }, | |
| { | |
| "epoch": 67.86, | |
| "eval_loss": 0.3710671663284302, | |
| "eval_runtime": 21.5817, | |
| "eval_samples_per_second": 18.303, | |
| "eval_steps_per_second": 0.602, | |
| "eval_wer": 0.41208321459105157, | |
| "step": 1900 | |
| }, | |
| { | |
| "epoch": 71.43, | |
| "learning_rate": 0.0005, | |
| "loss": 0.0827, | |
| "step": 2000 | |
| }, | |
| { | |
| "epoch": 71.43, | |
| "eval_loss": 0.40105244517326355, | |
| "eval_runtime": 21.6021, | |
| "eval_samples_per_second": 18.285, | |
| "eval_steps_per_second": 0.602, | |
| "eval_wer": 0.4160729552579082, | |
| "step": 2000 | |
| }, | |
| { | |
| "epoch": 75.0, | |
| "learning_rate": 0.0, | |
| "loss": 0.0665, | |
| "step": 2100 | |
| }, | |
| { | |
| "epoch": 75.0, | |
| "eval_loss": 0.3478299379348755, | |
| "eval_runtime": 21.8418, | |
| "eval_samples_per_second": 18.085, | |
| "eval_steps_per_second": 0.595, | |
| "eval_wer": 0.36534625249358793, | |
| "step": 2100 | |
| }, | |
| { | |
| "epoch": 75.0, | |
| "step": 2100, | |
| "total_flos": 5.419723804611084e+19, | |
| "train_loss": 0.8885498528253465, | |
| "train_runtime": 26470.8294, | |
| "train_samples_per_second": 10.064, | |
| "train_steps_per_second": 0.079 | |
| } | |
| ], | |
| "logging_steps": 100, | |
| "max_steps": 2100, | |
| "num_train_epochs": 75, | |
| "save_steps": 100, | |
| "total_flos": 5.419723804611084e+19, | |
| "trial_name": null, | |
| "trial_params": null | |
| } | |