{ "best_metric": 65.95253342648115, "best_model_checkpoint": "./whisper-small-tr-cp2/checkpoint-5000", "epoch": 2.2143489813994686, "eval_steps": 2500, "global_step": 5000, "is_hyper_param_search": false, "is_local_process_zero": true, "is_world_process_zero": true, "log_history": [ { "epoch": 0.04, "learning_rate": 2e-05, "loss": 0.0599, "step": 100 }, { "epoch": 0.09, "learning_rate": 3.9800000000000005e-05, "loss": 0.0901, "step": 200 }, { "epoch": 0.13, "learning_rate": 5.9800000000000003e-05, "loss": 0.111, "step": 300 }, { "epoch": 0.18, "learning_rate": 7.98e-05, "loss": 0.1622, "step": 400 }, { "epoch": 0.22, "learning_rate": 9.98e-05, "loss": 0.2083, "step": 500 }, { "epoch": 0.27, "learning_rate": 9.89578947368421e-05, "loss": 0.2504, "step": 600 }, { "epoch": 0.31, "learning_rate": 9.790526315789475e-05, "loss": 0.2681, "step": 700 }, { "epoch": 0.35, "learning_rate": 9.685263157894737e-05, "loss": 0.2753, "step": 800 }, { "epoch": 0.4, "learning_rate": 9.582105263157896e-05, "loss": 0.2486, "step": 900 }, { "epoch": 0.44, "learning_rate": 9.476842105263159e-05, "loss": 0.274, "step": 1000 }, { "epoch": 0.49, "learning_rate": 9.371578947368421e-05, "loss": 0.2653, "step": 1100 }, { "epoch": 0.53, "learning_rate": 9.266315789473685e-05, "loss": 0.2603, "step": 1200 }, { "epoch": 0.58, "learning_rate": 9.161052631578948e-05, "loss": 0.2613, "step": 1300 }, { "epoch": 0.62, "learning_rate": 9.05578947368421e-05, "loss": 0.2598, "step": 1400 }, { "epoch": 0.66, "learning_rate": 8.950526315789473e-05, "loss": 0.2643, "step": 1500 }, { "epoch": 0.71, "learning_rate": 8.845263157894737e-05, "loss": 0.2488, "step": 1600 }, { "epoch": 0.75, "learning_rate": 8.740000000000001e-05, "loss": 0.2534, "step": 1700 }, { "epoch": 0.8, "learning_rate": 8.634736842105264e-05, "loss": 0.2556, "step": 1800 }, { "epoch": 0.84, "learning_rate": 8.529473684210528e-05, "loss": 0.2604, "step": 1900 }, { "epoch": 0.89, "learning_rate": 8.42421052631579e-05, "loss": 0.2553, "step": 2000 }, { "epoch": 0.93, "learning_rate": 8.318947368421053e-05, "loss": 0.2579, "step": 2100 }, { "epoch": 0.97, "learning_rate": 8.213684210526315e-05, "loss": 0.2637, "step": 2200 }, { "epoch": 1.02, "learning_rate": 8.108421052631579e-05, "loss": 0.2063, "step": 2300 }, { "epoch": 1.06, "learning_rate": 8.003157894736842e-05, "loss": 0.1555, "step": 2400 }, { "epoch": 1.11, "learning_rate": 7.897894736842106e-05, "loss": 0.1555, "step": 2500 }, { "epoch": 1.11, "eval_loss": 0.7550089359283447, "eval_runtime": 3798.4679, "eval_samples_per_second": 2.67, "eval_steps_per_second": 0.334, "eval_wer": 95.60846868875046, "step": 2500 }, { "epoch": 1.15, "learning_rate": 7.79263157894737e-05, "loss": 0.1618, "step": 2600 }, { "epoch": 1.2, "learning_rate": 7.687368421052632e-05, "loss": 0.154, "step": 2700 }, { "epoch": 1.24, "learning_rate": 7.582105263157895e-05, "loss": 0.1636, "step": 2800 }, { "epoch": 1.28, "learning_rate": 7.476842105263159e-05, "loss": 0.1661, "step": 2900 }, { "epoch": 1.33, "learning_rate": 7.371578947368422e-05, "loss": 0.164, "step": 3000 }, { "epoch": 1.37, "learning_rate": 7.266315789473684e-05, "loss": 0.1716, "step": 3100 }, { "epoch": 1.42, "learning_rate": 7.161052631578947e-05, "loss": 0.1621, "step": 3200 }, { "epoch": 1.46, "learning_rate": 7.055789473684211e-05, "loss": 0.1598, "step": 3300 }, { "epoch": 1.51, "learning_rate": 6.950526315789475e-05, "loss": 0.1665, "step": 3400 }, { "epoch": 1.55, "learning_rate": 6.845263157894737e-05, "loss": 0.1657, "step": 3500 }, { "epoch": 1.59, "learning_rate": 6.740000000000001e-05, "loss": 0.1597, "step": 3600 }, { "epoch": 1.64, "learning_rate": 6.634736842105264e-05, "loss": 0.1626, "step": 3700 }, { "epoch": 1.68, "learning_rate": 6.529473684210526e-05, "loss": 0.1782, "step": 3800 }, { "epoch": 1.73, "learning_rate": 6.424210526315789e-05, "loss": 0.1664, "step": 3900 }, { "epoch": 1.77, "learning_rate": 6.318947368421053e-05, "loss": 0.1617, "step": 4000 }, { "epoch": 1.82, "learning_rate": 6.213684210526316e-05, "loss": 0.151, "step": 4100 }, { "epoch": 1.86, "learning_rate": 6.10842105263158e-05, "loss": 0.1661, "step": 4200 }, { "epoch": 1.9, "learning_rate": 6.003157894736843e-05, "loss": 0.1612, "step": 4300 }, { "epoch": 1.95, "learning_rate": 5.8978947368421054e-05, "loss": 0.1525, "step": 4400 }, { "epoch": 1.99, "learning_rate": 5.7926315789473686e-05, "loss": 0.1598, "step": 4500 }, { "epoch": 2.04, "learning_rate": 5.6873684210526326e-05, "loss": 0.0891, "step": 4600 }, { "epoch": 2.08, "learning_rate": 5.582105263157895e-05, "loss": 0.0874, "step": 4700 }, { "epoch": 2.13, "learning_rate": 5.476842105263158e-05, "loss": 0.0833, "step": 4800 }, { "epoch": 2.17, "learning_rate": 5.371578947368421e-05, "loss": 0.0854, "step": 4900 }, { "epoch": 2.21, "learning_rate": 5.266315789473685e-05, "loss": 0.0731, "step": 5000 }, { "epoch": 2.21, "eval_loss": 0.6640641689300537, "eval_runtime": 4603.4147, "eval_samples_per_second": 2.203, "eval_steps_per_second": 0.275, "eval_wer": 65.95253342648115, "step": 5000 } ], "logging_steps": 100, "max_steps": 10000, "num_train_epochs": 5, "save_steps": 2500, "total_flos": 2.308510051319808e+19, "trial_name": null, "trial_params": null }