| { | |
| "best_metric": 0.5411643981933594, | |
| "best_model_checkpoint": "/scratch/skscla001/speech/results/whisper-medium-bemgen-balanced-model/checkpoint-1000", | |
| "epoch": 3.1683168316831685, | |
| "eval_steps": 200, | |
| "global_step": 1600, | |
| "is_hyper_param_search": false, | |
| "is_local_process_zero": true, | |
| "is_world_process_zero": true, | |
| "log_history": [ | |
| { | |
| "epoch": 0.04950495049504951, | |
| "grad_norm": 121.81847381591797, | |
| "learning_rate": 4.0000000000000003e-07, | |
| "loss": 11.6499, | |
| "step": 25 | |
| }, | |
| { | |
| "epoch": 0.09900990099009901, | |
| "grad_norm": 104.46240234375, | |
| "learning_rate": 9.000000000000001e-07, | |
| "loss": 9.6493, | |
| "step": 50 | |
| }, | |
| { | |
| "epoch": 0.1485148514851485, | |
| "grad_norm": 105.9942626953125, | |
| "learning_rate": 1.3800000000000001e-06, | |
| "loss": 7.7745, | |
| "step": 75 | |
| }, | |
| { | |
| "epoch": 0.19801980198019803, | |
| "grad_norm": 86.44505310058594, | |
| "learning_rate": 1.8800000000000002e-06, | |
| "loss": 6.0363, | |
| "step": 100 | |
| }, | |
| { | |
| "epoch": 0.24752475247524752, | |
| "grad_norm": 89.36949920654297, | |
| "learning_rate": 2.38e-06, | |
| "loss": 4.9289, | |
| "step": 125 | |
| }, | |
| { | |
| "epoch": 0.297029702970297, | |
| "grad_norm": 67.95636749267578, | |
| "learning_rate": 2.88e-06, | |
| "loss": 4.2784, | |
| "step": 150 | |
| }, | |
| { | |
| "epoch": 0.3465346534653465, | |
| "grad_norm": 71.39818572998047, | |
| "learning_rate": 3.3800000000000007e-06, | |
| "loss": 3.8983, | |
| "step": 175 | |
| }, | |
| { | |
| "epoch": 0.39603960396039606, | |
| "grad_norm": 77.7619400024414, | |
| "learning_rate": 3.88e-06, | |
| "loss": 3.7045, | |
| "step": 200 | |
| }, | |
| { | |
| "epoch": 0.39603960396039606, | |
| "eval_loss": 0.9162073135375977, | |
| "eval_runtime": 279.8105, | |
| "eval_samples_per_second": 1.801, | |
| "eval_steps_per_second": 0.901, | |
| "eval_wer": 0.6826695371367061, | |
| "step": 200 | |
| }, | |
| { | |
| "epoch": 0.44554455445544555, | |
| "grad_norm": 65.45413208007812, | |
| "learning_rate": 4.38e-06, | |
| "loss": 3.4515, | |
| "step": 225 | |
| }, | |
| { | |
| "epoch": 0.49504950495049505, | |
| "grad_norm": 83.95701599121094, | |
| "learning_rate": 4.880000000000001e-06, | |
| "loss": 3.3691, | |
| "step": 250 | |
| }, | |
| { | |
| "epoch": 0.5445544554455446, | |
| "grad_norm": 60.99998092651367, | |
| "learning_rate": 5.380000000000001e-06, | |
| "loss": 3.1637, | |
| "step": 275 | |
| }, | |
| { | |
| "epoch": 0.594059405940594, | |
| "grad_norm": 86.96224212646484, | |
| "learning_rate": 5.8800000000000005e-06, | |
| "loss": 3.015, | |
| "step": 300 | |
| }, | |
| { | |
| "epoch": 0.6435643564356436, | |
| "grad_norm": 75.70954132080078, | |
| "learning_rate": 6.380000000000001e-06, | |
| "loss": 2.9369, | |
| "step": 325 | |
| }, | |
| { | |
| "epoch": 0.693069306930693, | |
| "grad_norm": 68.08271026611328, | |
| "learning_rate": 6.88e-06, | |
| "loss": 2.7672, | |
| "step": 350 | |
| }, | |
| { | |
| "epoch": 0.7425742574257426, | |
| "grad_norm": 59.94082260131836, | |
| "learning_rate": 7.3800000000000005e-06, | |
| "loss": 2.7812, | |
| "step": 375 | |
| }, | |
| { | |
| "epoch": 0.7920792079207921, | |
| "grad_norm": 63.15815353393555, | |
| "learning_rate": 7.88e-06, | |
| "loss": 2.687, | |
| "step": 400 | |
| }, | |
| { | |
| "epoch": 0.7920792079207921, | |
| "eval_loss": 0.6817564368247986, | |
| "eval_runtime": 278.5541, | |
| "eval_samples_per_second": 1.809, | |
| "eval_steps_per_second": 0.905, | |
| "eval_wer": 0.5351991388589882, | |
| "step": 400 | |
| }, | |
| { | |
| "epoch": 0.8415841584158416, | |
| "grad_norm": 66.49015808105469, | |
| "learning_rate": 8.380000000000001e-06, | |
| "loss": 2.6355, | |
| "step": 425 | |
| }, | |
| { | |
| "epoch": 0.8910891089108911, | |
| "grad_norm": 69.2128677368164, | |
| "learning_rate": 8.880000000000001e-06, | |
| "loss": 2.5594, | |
| "step": 450 | |
| }, | |
| { | |
| "epoch": 0.9405940594059405, | |
| "grad_norm": 82.29959106445312, | |
| "learning_rate": 9.38e-06, | |
| "loss": 2.6779, | |
| "step": 475 | |
| }, | |
| { | |
| "epoch": 0.9900990099009901, | |
| "grad_norm": 75.67899322509766, | |
| "learning_rate": 9.88e-06, | |
| "loss": 2.6395, | |
| "step": 500 | |
| }, | |
| { | |
| "epoch": 1.0396039603960396, | |
| "grad_norm": 74.15483093261719, | |
| "learning_rate": 9.957777777777779e-06, | |
| "loss": 1.9931, | |
| "step": 525 | |
| }, | |
| { | |
| "epoch": 1.0891089108910892, | |
| "grad_norm": 49.705169677734375, | |
| "learning_rate": 9.902222222222223e-06, | |
| "loss": 1.7258, | |
| "step": 550 | |
| }, | |
| { | |
| "epoch": 1.1386138613861387, | |
| "grad_norm": 37.47819519042969, | |
| "learning_rate": 9.846666666666668e-06, | |
| "loss": 1.8517, | |
| "step": 575 | |
| }, | |
| { | |
| "epoch": 1.188118811881188, | |
| "grad_norm": 50.20145034790039, | |
| "learning_rate": 9.791111111111112e-06, | |
| "loss": 1.7185, | |
| "step": 600 | |
| }, | |
| { | |
| "epoch": 1.188118811881188, | |
| "eval_loss": 0.6266342401504517, | |
| "eval_runtime": 282.9063, | |
| "eval_samples_per_second": 1.782, | |
| "eval_steps_per_second": 0.891, | |
| "eval_wer": 0.4988159311087191, | |
| "step": 600 | |
| }, | |
| { | |
| "epoch": 1.2376237623762376, | |
| "grad_norm": 77.02053833007812, | |
| "learning_rate": 9.735555555555556e-06, | |
| "loss": 1.8482, | |
| "step": 625 | |
| }, | |
| { | |
| "epoch": 1.2871287128712872, | |
| "grad_norm": 73.36985778808594, | |
| "learning_rate": 9.68e-06, | |
| "loss": 1.834, | |
| "step": 650 | |
| }, | |
| { | |
| "epoch": 1.3366336633663367, | |
| "grad_norm": 48.02408981323242, | |
| "learning_rate": 9.624444444444445e-06, | |
| "loss": 1.6974, | |
| "step": 675 | |
| }, | |
| { | |
| "epoch": 1.386138613861386, | |
| "grad_norm": 54.08694076538086, | |
| "learning_rate": 9.56888888888889e-06, | |
| "loss": 1.7489, | |
| "step": 700 | |
| }, | |
| { | |
| "epoch": 1.4356435643564356, | |
| "grad_norm": 39.281005859375, | |
| "learning_rate": 9.513333333333334e-06, | |
| "loss": 1.8013, | |
| "step": 725 | |
| }, | |
| { | |
| "epoch": 1.4851485148514851, | |
| "grad_norm": 43.782196044921875, | |
| "learning_rate": 9.457777777777778e-06, | |
| "loss": 1.7419, | |
| "step": 750 | |
| }, | |
| { | |
| "epoch": 1.5346534653465347, | |
| "grad_norm": 40.640323638916016, | |
| "learning_rate": 9.402222222222222e-06, | |
| "loss": 1.7863, | |
| "step": 775 | |
| }, | |
| { | |
| "epoch": 1.5841584158415842, | |
| "grad_norm": 43.944881439208984, | |
| "learning_rate": 9.346666666666666e-06, | |
| "loss": 1.7232, | |
| "step": 800 | |
| }, | |
| { | |
| "epoch": 1.5841584158415842, | |
| "eval_loss": 0.5673945546150208, | |
| "eval_runtime": 279.2797, | |
| "eval_samples_per_second": 1.805, | |
| "eval_steps_per_second": 0.902, | |
| "eval_wer": 0.4592034445640474, | |
| "step": 800 | |
| }, | |
| { | |
| "epoch": 1.6336633663366338, | |
| "grad_norm": 53.22232437133789, | |
| "learning_rate": 9.291111111111112e-06, | |
| "loss": 1.6903, | |
| "step": 825 | |
| }, | |
| { | |
| "epoch": 1.6831683168316833, | |
| "grad_norm": 41.854610443115234, | |
| "learning_rate": 9.235555555555556e-06, | |
| "loss": 1.656, | |
| "step": 850 | |
| }, | |
| { | |
| "epoch": 1.7326732673267327, | |
| "grad_norm": 38.988529205322266, | |
| "learning_rate": 9.180000000000002e-06, | |
| "loss": 1.5997, | |
| "step": 875 | |
| }, | |
| { | |
| "epoch": 1.7821782178217822, | |
| "grad_norm": 54.63161087036133, | |
| "learning_rate": 9.124444444444444e-06, | |
| "loss": 1.6233, | |
| "step": 900 | |
| }, | |
| { | |
| "epoch": 1.8316831683168315, | |
| "grad_norm": 41.60354232788086, | |
| "learning_rate": 9.06888888888889e-06, | |
| "loss": 1.4889, | |
| "step": 925 | |
| }, | |
| { | |
| "epoch": 1.881188118811881, | |
| "grad_norm": 42.82672119140625, | |
| "learning_rate": 9.013333333333334e-06, | |
| "loss": 1.6465, | |
| "step": 950 | |
| }, | |
| { | |
| "epoch": 1.9306930693069306, | |
| "grad_norm": 50.5968132019043, | |
| "learning_rate": 8.957777777777778e-06, | |
| "loss": 1.6738, | |
| "step": 975 | |
| }, | |
| { | |
| "epoch": 1.9801980198019802, | |
| "grad_norm": 46.50210952758789, | |
| "learning_rate": 8.902222222222224e-06, | |
| "loss": 1.6083, | |
| "step": 1000 | |
| }, | |
| { | |
| "epoch": 1.9801980198019802, | |
| "eval_loss": 0.5411643981933594, | |
| "eval_runtime": 277.778, | |
| "eval_samples_per_second": 1.814, | |
| "eval_steps_per_second": 0.907, | |
| "eval_wer": 0.4413347685683531, | |
| "step": 1000 | |
| }, | |
| { | |
| "epoch": 2.0297029702970297, | |
| "grad_norm": 25.421289443969727, | |
| "learning_rate": 8.846666666666668e-06, | |
| "loss": 1.1495, | |
| "step": 1025 | |
| }, | |
| { | |
| "epoch": 2.0792079207920793, | |
| "grad_norm": 34.11124038696289, | |
| "learning_rate": 8.791111111111112e-06, | |
| "loss": 0.8415, | |
| "step": 1050 | |
| }, | |
| { | |
| "epoch": 2.128712871287129, | |
| "grad_norm": 28.867008209228516, | |
| "learning_rate": 8.735555555555556e-06, | |
| "loss": 0.7942, | |
| "step": 1075 | |
| }, | |
| { | |
| "epoch": 2.1782178217821784, | |
| "grad_norm": 28.74831199645996, | |
| "learning_rate": 8.68e-06, | |
| "loss": 0.894, | |
| "step": 1100 | |
| }, | |
| { | |
| "epoch": 2.227722772277228, | |
| "grad_norm": 31.9539794921875, | |
| "learning_rate": 8.624444444444446e-06, | |
| "loss": 0.9315, | |
| "step": 1125 | |
| }, | |
| { | |
| "epoch": 2.2772277227722775, | |
| "grad_norm": 25.5980224609375, | |
| "learning_rate": 8.56888888888889e-06, | |
| "loss": 0.8735, | |
| "step": 1150 | |
| }, | |
| { | |
| "epoch": 2.3267326732673266, | |
| "grad_norm": 30.295211791992188, | |
| "learning_rate": 8.513333333333335e-06, | |
| "loss": 0.7829, | |
| "step": 1175 | |
| }, | |
| { | |
| "epoch": 2.376237623762376, | |
| "grad_norm": 38.251556396484375, | |
| "learning_rate": 8.457777777777778e-06, | |
| "loss": 0.7643, | |
| "step": 1200 | |
| }, | |
| { | |
| "epoch": 2.376237623762376, | |
| "eval_loss": 0.565162181854248, | |
| "eval_runtime": 302.1807, | |
| "eval_samples_per_second": 1.668, | |
| "eval_steps_per_second": 0.834, | |
| "eval_wer": 0.4279870828848224, | |
| "step": 1200 | |
| }, | |
| { | |
| "epoch": 2.4257425742574257, | |
| "grad_norm": 22.155136108398438, | |
| "learning_rate": 8.402222222222223e-06, | |
| "loss": 0.8875, | |
| "step": 1225 | |
| }, | |
| { | |
| "epoch": 2.4752475247524752, | |
| "grad_norm": 29.48054313659668, | |
| "learning_rate": 8.346666666666668e-06, | |
| "loss": 0.8138, | |
| "step": 1250 | |
| }, | |
| { | |
| "epoch": 2.5247524752475248, | |
| "grad_norm": 31.146007537841797, | |
| "learning_rate": 8.291111111111112e-06, | |
| "loss": 0.8036, | |
| "step": 1275 | |
| }, | |
| { | |
| "epoch": 2.5742574257425743, | |
| "grad_norm": 39.311256408691406, | |
| "learning_rate": 8.235555555555557e-06, | |
| "loss": 0.8389, | |
| "step": 1300 | |
| }, | |
| { | |
| "epoch": 2.623762376237624, | |
| "grad_norm": 21.874267578125, | |
| "learning_rate": 8.18e-06, | |
| "loss": 0.8243, | |
| "step": 1325 | |
| }, | |
| { | |
| "epoch": 2.6732673267326734, | |
| "grad_norm": 27.84779167175293, | |
| "learning_rate": 8.124444444444445e-06, | |
| "loss": 1.0702, | |
| "step": 1350 | |
| }, | |
| { | |
| "epoch": 2.7227722772277225, | |
| "grad_norm": 41.43334197998047, | |
| "learning_rate": 8.06888888888889e-06, | |
| "loss": 0.8892, | |
| "step": 1375 | |
| }, | |
| { | |
| "epoch": 2.772277227722772, | |
| "grad_norm": 36.145965576171875, | |
| "learning_rate": 8.013333333333333e-06, | |
| "loss": 0.8362, | |
| "step": 1400 | |
| }, | |
| { | |
| "epoch": 2.772277227722772, | |
| "eval_loss": 0.5454630851745605, | |
| "eval_runtime": 277.6054, | |
| "eval_samples_per_second": 1.816, | |
| "eval_steps_per_second": 0.908, | |
| "eval_wer": 0.40516684607104414, | |
| "step": 1400 | |
| }, | |
| { | |
| "epoch": 2.8217821782178216, | |
| "grad_norm": 82.11780548095703, | |
| "learning_rate": 7.957777777777779e-06, | |
| "loss": 0.7491, | |
| "step": 1425 | |
| }, | |
| { | |
| "epoch": 2.871287128712871, | |
| "grad_norm": 28.929828643798828, | |
| "learning_rate": 7.902222222222223e-06, | |
| "loss": 0.9076, | |
| "step": 1450 | |
| }, | |
| { | |
| "epoch": 2.9207920792079207, | |
| "grad_norm": 29.58544921875, | |
| "learning_rate": 7.846666666666667e-06, | |
| "loss": 0.8593, | |
| "step": 1475 | |
| }, | |
| { | |
| "epoch": 2.9702970297029703, | |
| "grad_norm": 36.51881790161133, | |
| "learning_rate": 7.791111111111111e-06, | |
| "loss": 0.7739, | |
| "step": 1500 | |
| }, | |
| { | |
| "epoch": 3.01980198019802, | |
| "grad_norm": 13.50954532623291, | |
| "learning_rate": 7.735555555555557e-06, | |
| "loss": 0.6666, | |
| "step": 1525 | |
| }, | |
| { | |
| "epoch": 3.0693069306930694, | |
| "grad_norm": 15.712291717529297, | |
| "learning_rate": 7.680000000000001e-06, | |
| "loss": 0.3365, | |
| "step": 1550 | |
| }, | |
| { | |
| "epoch": 3.118811881188119, | |
| "grad_norm": 22.804317474365234, | |
| "learning_rate": 7.624444444444445e-06, | |
| "loss": 0.3306, | |
| "step": 1575 | |
| }, | |
| { | |
| "epoch": 3.1683168316831685, | |
| "grad_norm": 22.823379516601562, | |
| "learning_rate": 7.56888888888889e-06, | |
| "loss": 0.422, | |
| "step": 1600 | |
| }, | |
| { | |
| "epoch": 3.1683168316831685, | |
| "eval_loss": 0.577109694480896, | |
| "eval_runtime": 282.3788, | |
| "eval_samples_per_second": 1.785, | |
| "eval_steps_per_second": 0.892, | |
| "eval_wer": 0.3991388589881593, | |
| "step": 1600 | |
| }, | |
| { | |
| "epoch": 3.1683168316831685, | |
| "step": 1600, | |
| "total_flos": 1.306068100153344e+19, | |
| "train_loss": 2.189091945886612, | |
| "train_runtime": 5496.1435, | |
| "train_samples_per_second": 7.278, | |
| "train_steps_per_second": 0.91 | |
| } | |
| ], | |
| "logging_steps": 25, | |
| "max_steps": 5000, | |
| "num_input_tokens_seen": 0, | |
| "num_train_epochs": 10, | |
| "save_steps": 200, | |
| "stateful_callbacks": { | |
| "EarlyStoppingCallback": { | |
| "args": { | |
| "early_stopping_patience": 3, | |
| "early_stopping_threshold": 0.0 | |
| }, | |
| "attributes": { | |
| "early_stopping_patience_counter": 3 | |
| } | |
| }, | |
| "TrainerControl": { | |
| "args": { | |
| "should_epoch_stop": false, | |
| "should_evaluate": false, | |
| "should_log": false, | |
| "should_save": true, | |
| "should_training_stop": true | |
| }, | |
| "attributes": {} | |
| } | |
| }, | |
| "total_flos": 1.306068100153344e+19, | |
| "train_batch_size": 2, | |
| "trial_name": null, | |
| "trial_params": null | |
| } | |