{ "best_metric": 31.350262297574698, "best_model_checkpoint": "./whisper-medium-ckb/checkpoint-1830", "epoch": 6.004909983633388, "global_step": 1833, "is_hyper_param_search": false, "is_local_process_zero": true, "is_world_process_zero": true, "log_history": [ { "epoch": 0.01, "learning_rate": 4.2000000000000006e-07, "loss": 3.1622, "step": 25 }, { "epoch": 0.03, "learning_rate": 9.200000000000001e-07, "loss": 2.1678, "step": 50 }, { "epoch": 0.04, "learning_rate": 1.42e-06, "loss": 1.2419, "step": 75 }, { "epoch": 0.05, "learning_rate": 1.9200000000000003e-06, "loss": 0.7203, "step": 100 }, { "epoch": 0.07, "learning_rate": 2.42e-06, "loss": 0.5754, "step": 125 }, { "epoch": 0.08, "learning_rate": 2.92e-06, "loss": 0.4364, "step": 150 }, { "epoch": 0.1, "learning_rate": 3.4200000000000007e-06, "loss": 0.4198, "step": 175 }, { "epoch": 0.11, "learning_rate": 3.920000000000001e-06, "loss": 0.404, "step": 200 }, { "epoch": 0.12, "learning_rate": 4.42e-06, "loss": 0.3441, "step": 225 }, { "epoch": 0.14, "learning_rate": 4.92e-06, "loss": 0.3334, "step": 250 }, { "epoch": 0.15, "learning_rate": 5.420000000000001e-06, "loss": 0.2885, "step": 275 }, { "epoch": 0.16, "learning_rate": 5.92e-06, "loss": 0.274, "step": 300 }, { "epoch": 1.01, "learning_rate": 6.42e-06, "loss": 0.249, "step": 325 }, { "epoch": 1.03, "learning_rate": 6.92e-06, "loss": 0.27, "step": 350 }, { "epoch": 1.03, "eval_loss": 0.27222105860710144, "eval_runtime": 5782.8476, "eval_samples_per_second": 0.841, "eval_steps_per_second": 0.053, "eval_wer": 43.91773739831217, "step": 366 }, { "epoch": 1.04, "learning_rate": 7.420000000000001e-06, "loss": 0.2623, "step": 375 }, { "epoch": 1.05, "learning_rate": 7.92e-06, "loss": 0.2521, "step": 400 }, { "epoch": 1.07, "learning_rate": 8.42e-06, "loss": 0.258, "step": 425 }, { "epoch": 1.08, "learning_rate": 8.920000000000001e-06, "loss": 0.2201, "step": 450 }, { "epoch": 1.09, "learning_rate": 9.42e-06, "loss": 0.2285, "step": 475 }, { "epoch": 1.11, "learning_rate": 9.920000000000002e-06, "loss": 0.2239, "step": 500 }, { "epoch": 1.12, "learning_rate": 9.84246061515379e-06, "loss": 0.1928, "step": 525 }, { "epoch": 1.13, "learning_rate": 9.654913728432109e-06, "loss": 0.197, "step": 550 }, { "epoch": 1.15, "learning_rate": 9.467366841710428e-06, "loss": 0.1732, "step": 575 }, { "epoch": 1.16, "learning_rate": 9.279819954988747e-06, "loss": 0.1454, "step": 600 }, { "epoch": 2.01, "learning_rate": 9.092273068267068e-06, "loss": 0.142, "step": 625 }, { "epoch": 2.02, "learning_rate": 8.904726181545387e-06, "loss": 0.1486, "step": 650 }, { "epoch": 2.04, "learning_rate": 8.717179294823706e-06, "loss": 0.1446, "step": 675 }, { "epoch": 2.05, "learning_rate": 8.529632408102027e-06, "loss": 0.1329, "step": 700 }, { "epoch": 2.06, "learning_rate": 8.342085521380345e-06, "loss": 0.1498, "step": 725 }, { "epoch": 2.07, "eval_loss": 0.23542030155658722, "eval_runtime": 5576.1832, "eval_samples_per_second": 0.872, "eval_steps_per_second": 0.055, "eval_wer": 36.284497833193946, "step": 732 }, { "epoch": 2.08, "learning_rate": 8.154538634658665e-06, "loss": 0.1265, "step": 750 }, { "epoch": 2.09, "learning_rate": 7.966991747936984e-06, "loss": 0.1154, "step": 775 }, { "epoch": 2.1, "learning_rate": 7.779444861215303e-06, "loss": 0.1095, "step": 800 }, { "epoch": 2.12, "learning_rate": 7.591897974493624e-06, "loss": 0.1, "step": 825 }, { "epoch": 2.13, "learning_rate": 7.404351087771943e-06, "loss": 0.1, "step": 850 }, { "epoch": 2.15, "learning_rate": 7.216804201050263e-06, "loss": 0.0924, "step": 875 }, { "epoch": 2.16, "learning_rate": 7.0292573143285824e-06, "loss": 0.0773, "step": 900 }, { "epoch": 3.01, "learning_rate": 6.841710427606902e-06, "loss": 0.0687, "step": 925 }, { "epoch": 3.02, "learning_rate": 6.654163540885222e-06, "loss": 0.0778, "step": 950 }, { "epoch": 3.03, "learning_rate": 6.4666166541635414e-06, "loss": 0.0677, "step": 975 }, { "epoch": 3.05, "learning_rate": 6.279069767441861e-06, "loss": 0.0723, "step": 1000 }, { "epoch": 3.06, "learning_rate": 6.091522880720181e-06, "loss": 0.0732, "step": 1025 }, { "epoch": 3.08, "learning_rate": 5.9039759939984996e-06, "loss": 0.062, "step": 1050 }, { "epoch": 3.09, "learning_rate": 5.7164291072768195e-06, "loss": 0.0538, "step": 1075 }, { "epoch": 3.1, "eval_loss": 0.24223530292510986, "eval_runtime": 5591.471, "eval_samples_per_second": 0.87, "eval_steps_per_second": 0.054, "eval_wer": 33.65391925796396, "step": 1098 }, { "epoch": 3.1, "learning_rate": 5.5288822205551395e-06, "loss": 0.0549, "step": 1100 }, { "epoch": 3.12, "learning_rate": 5.3413353338334586e-06, "loss": 0.0473, "step": 1125 }, { "epoch": 3.13, "learning_rate": 5.1537884471117785e-06, "loss": 0.0491, "step": 1150 }, { "epoch": 3.14, "learning_rate": 4.9662415603900985e-06, "loss": 0.0421, "step": 1175 }, { "epoch": 3.16, "learning_rate": 4.7786946736684176e-06, "loss": 0.0343, "step": 1200 }, { "epoch": 4.0, "learning_rate": 4.591147786946737e-06, "loss": 0.0336, "step": 1225 }, { "epoch": 4.02, "learning_rate": 4.403600900225057e-06, "loss": 0.0316, "step": 1250 }, { "epoch": 4.03, "learning_rate": 4.2160540135033765e-06, "loss": 0.0324, "step": 1275 }, { "epoch": 4.05, "learning_rate": 4.028507126781696e-06, "loss": 0.031, "step": 1300 }, { "epoch": 4.06, "learning_rate": 3.840960240060016e-06, "loss": 0.031, "step": 1325 }, { "epoch": 4.07, "learning_rate": 3.6534133533383347e-06, "loss": 0.028, "step": 1350 }, { "epoch": 4.09, "learning_rate": 3.4658664666166546e-06, "loss": 0.0218, "step": 1375 }, { "epoch": 4.1, "learning_rate": 3.278319579894974e-06, "loss": 0.0221, "step": 1400 }, { "epoch": 4.11, "learning_rate": 3.0907726931732933e-06, "loss": 0.0183, "step": 1425 }, { "epoch": 4.13, "learning_rate": 2.903225806451613e-06, "loss": 0.0168, "step": 1450 }, { "epoch": 4.14, "eval_loss": 0.27173367142677307, "eval_runtime": 5547.7291, "eval_samples_per_second": 0.876, "eval_steps_per_second": 0.055, "eval_wer": 32.384246939861626, "step": 1464 }, { "epoch": 4.14, "learning_rate": 2.7156789197299327e-06, "loss": 0.0175, "step": 1475 }, { "epoch": 4.15, "learning_rate": 2.5281320330082523e-06, "loss": 0.0121, "step": 1500 }, { "epoch": 5.0, "learning_rate": 2.3405851462865718e-06, "loss": 0.0103, "step": 1525 }, { "epoch": 5.02, "learning_rate": 2.1530382595648913e-06, "loss": 0.0098, "step": 1550 }, { "epoch": 5.03, "learning_rate": 1.965491372843211e-06, "loss": 0.0124, "step": 1575 }, { "epoch": 5.04, "learning_rate": 1.7779444861215306e-06, "loss": 0.0102, "step": 1600 }, { "epoch": 5.06, "learning_rate": 1.5903975993998503e-06, "loss": 0.0101, "step": 1625 }, { "epoch": 5.07, "learning_rate": 1.4028507126781696e-06, "loss": 0.0081, "step": 1650 }, { "epoch": 5.08, "learning_rate": 1.2153038259564891e-06, "loss": 0.0076, "step": 1675 }, { "epoch": 5.1, "learning_rate": 1.0277569392348089e-06, "loss": 0.0076, "step": 1700 }, { "epoch": 5.11, "learning_rate": 8.402100525131284e-07, "loss": 0.0055, "step": 1725 }, { "epoch": 5.13, "learning_rate": 6.52663165791448e-07, "loss": 0.0046, "step": 1750 }, { "epoch": 5.14, "learning_rate": 4.651162790697675e-07, "loss": 0.005, "step": 1775 }, { "epoch": 5.15, "learning_rate": 2.7756939234808706e-07, "loss": 0.0031, "step": 1800 }, { "epoch": 6.0, "learning_rate": 9.002250562640661e-08, "loss": 0.0025, "step": 1825 }, { "epoch": 6.0, "eval_loss": 0.2976975440979004, "eval_runtime": 5536.8386, "eval_samples_per_second": 0.878, "eval_steps_per_second": 0.055, "eval_wer": 31.350262297574698, "step": 1830 }, { "epoch": 6.0, "step": 1833, "total_flos": 5.977887064326144e+19, "train_loss": 0.2124484316768522, "train_runtime": 44455.4113, "train_samples_per_second": 1.319, "train_steps_per_second": 0.041 } ], "max_steps": 1833, "num_train_epochs": 9223372036854775807, "total_flos": 5.977887064326144e+19, "trial_name": null, "trial_params": null }