| { | |
| "best_metric": 0.25605687499046326, | |
| "best_model_checkpoint": "/scratch/skscla001/speech/results/whisper-medium-nyagen-combined-model/checkpoint-1000", | |
| "epoch": 4.255659121171771, | |
| "eval_steps": 200, | |
| "global_step": 1600, | |
| "is_hyper_param_search": false, | |
| "is_local_process_zero": true, | |
| "is_world_process_zero": true, | |
| "log_history": [ | |
| { | |
| "epoch": 0.06657789613848203, | |
| "grad_norm": 56.093265533447266, | |
| "learning_rate": 4.2000000000000006e-07, | |
| "loss": 5.267, | |
| "step": 25 | |
| }, | |
| { | |
| "epoch": 0.13315579227696406, | |
| "grad_norm": 34.71885681152344, | |
| "learning_rate": 9.200000000000001e-07, | |
| "loss": 4.1657, | |
| "step": 50 | |
| }, | |
| { | |
| "epoch": 0.19973368841544606, | |
| "grad_norm": 31.17400360107422, | |
| "learning_rate": 1.42e-06, | |
| "loss": 3.1786, | |
| "step": 75 | |
| }, | |
| { | |
| "epoch": 0.2663115845539281, | |
| "grad_norm": 30.04043960571289, | |
| "learning_rate": 1.9200000000000003e-06, | |
| "loss": 2.1171, | |
| "step": 100 | |
| }, | |
| { | |
| "epoch": 0.33288948069241014, | |
| "grad_norm": 21.41140365600586, | |
| "learning_rate": 2.42e-06, | |
| "loss": 1.7518, | |
| "step": 125 | |
| }, | |
| { | |
| "epoch": 0.3994673768308921, | |
| "grad_norm": 20.285459518432617, | |
| "learning_rate": 2.92e-06, | |
| "loss": 1.3976, | |
| "step": 150 | |
| }, | |
| { | |
| "epoch": 0.46604527296937415, | |
| "grad_norm": 19.031597137451172, | |
| "learning_rate": 3.4200000000000007e-06, | |
| "loss": 1.2585, | |
| "step": 175 | |
| }, | |
| { | |
| "epoch": 0.5326231691078562, | |
| "grad_norm": 22.678789138793945, | |
| "learning_rate": 3.920000000000001e-06, | |
| "loss": 1.1602, | |
| "step": 200 | |
| }, | |
| { | |
| "epoch": 0.5326231691078562, | |
| "eval_loss": 0.5182062983512878, | |
| "eval_runtime": 169.8362, | |
| "eval_samples_per_second": 2.096, | |
| "eval_steps_per_second": 0.524, | |
| "eval_wer": 0.3746898263027295, | |
| "step": 200 | |
| }, | |
| { | |
| "epoch": 0.5992010652463382, | |
| "grad_norm": 19.633346557617188, | |
| "learning_rate": 4.42e-06, | |
| "loss": 0.976, | |
| "step": 225 | |
| }, | |
| { | |
| "epoch": 0.6657789613848203, | |
| "grad_norm": 22.2548828125, | |
| "learning_rate": 4.92e-06, | |
| "loss": 0.9464, | |
| "step": 250 | |
| }, | |
| { | |
| "epoch": 0.7323568575233023, | |
| "grad_norm": 19.41946029663086, | |
| "learning_rate": 5.420000000000001e-06, | |
| "loss": 0.8527, | |
| "step": 275 | |
| }, | |
| { | |
| "epoch": 0.7989347536617842, | |
| "grad_norm": 20.135112762451172, | |
| "learning_rate": 5.92e-06, | |
| "loss": 0.8483, | |
| "step": 300 | |
| }, | |
| { | |
| "epoch": 0.8655126498002663, | |
| "grad_norm": 17.5079345703125, | |
| "learning_rate": 6.42e-06, | |
| "loss": 0.9001, | |
| "step": 325 | |
| }, | |
| { | |
| "epoch": 0.9320905459387483, | |
| "grad_norm": 21.997474670410156, | |
| "learning_rate": 6.92e-06, | |
| "loss": 0.761, | |
| "step": 350 | |
| }, | |
| { | |
| "epoch": 0.9986684420772304, | |
| "grad_norm": 20.054590225219727, | |
| "learning_rate": 7.420000000000001e-06, | |
| "loss": 0.7337, | |
| "step": 375 | |
| }, | |
| { | |
| "epoch": 1.0639147802929427, | |
| "grad_norm": 16.24369239807129, | |
| "learning_rate": 7.92e-06, | |
| "loss": 0.5456, | |
| "step": 400 | |
| }, | |
| { | |
| "epoch": 1.0639147802929427, | |
| "eval_loss": 0.3444797098636627, | |
| "eval_runtime": 166.2672, | |
| "eval_samples_per_second": 2.141, | |
| "eval_steps_per_second": 0.535, | |
| "eval_wer": 0.25516956162117455, | |
| "step": 400 | |
| }, | |
| { | |
| "epoch": 1.1304926764314247, | |
| "grad_norm": 17.575782775878906, | |
| "learning_rate": 8.42e-06, | |
| "loss": 0.4736, | |
| "step": 425 | |
| }, | |
| { | |
| "epoch": 1.1970705725699067, | |
| "grad_norm": 12.86981201171875, | |
| "learning_rate": 8.920000000000001e-06, | |
| "loss": 0.535, | |
| "step": 450 | |
| }, | |
| { | |
| "epoch": 1.2636484687083889, | |
| "grad_norm": 15.951074600219727, | |
| "learning_rate": 9.42e-06, | |
| "loss": 0.5233, | |
| "step": 475 | |
| }, | |
| { | |
| "epoch": 1.3302263648468708, | |
| "grad_norm": 18.161441802978516, | |
| "learning_rate": 9.920000000000002e-06, | |
| "loss": 0.5483, | |
| "step": 500 | |
| }, | |
| { | |
| "epoch": 1.3968042609853528, | |
| "grad_norm": 13.9922513961792, | |
| "learning_rate": 9.953333333333333e-06, | |
| "loss": 0.5045, | |
| "step": 525 | |
| }, | |
| { | |
| "epoch": 1.463382157123835, | |
| "grad_norm": 19.316190719604492, | |
| "learning_rate": 9.89777777777778e-06, | |
| "loss": 0.4916, | |
| "step": 550 | |
| }, | |
| { | |
| "epoch": 1.5299600532623168, | |
| "grad_norm": 14.279507637023926, | |
| "learning_rate": 9.842222222222223e-06, | |
| "loss": 0.5007, | |
| "step": 575 | |
| }, | |
| { | |
| "epoch": 1.596537949400799, | |
| "grad_norm": 15.150153160095215, | |
| "learning_rate": 9.786666666666667e-06, | |
| "loss": 0.5516, | |
| "step": 600 | |
| }, | |
| { | |
| "epoch": 1.596537949400799, | |
| "eval_loss": 0.29034528136253357, | |
| "eval_runtime": 169.9132, | |
| "eval_samples_per_second": 2.095, | |
| "eval_steps_per_second": 0.524, | |
| "eval_wer": 0.2413151364764268, | |
| "step": 600 | |
| }, | |
| { | |
| "epoch": 1.663115845539281, | |
| "grad_norm": 11.70125675201416, | |
| "learning_rate": 9.731111111111113e-06, | |
| "loss": 0.4788, | |
| "step": 625 | |
| }, | |
| { | |
| "epoch": 1.729693741677763, | |
| "grad_norm": 12.724228858947754, | |
| "learning_rate": 9.675555555555555e-06, | |
| "loss": 0.4391, | |
| "step": 650 | |
| }, | |
| { | |
| "epoch": 1.796271637816245, | |
| "grad_norm": 13.262675285339355, | |
| "learning_rate": 9.620000000000001e-06, | |
| "loss": 0.5023, | |
| "step": 675 | |
| }, | |
| { | |
| "epoch": 1.862849533954727, | |
| "grad_norm": 15.492055892944336, | |
| "learning_rate": 9.564444444444445e-06, | |
| "loss": 0.493, | |
| "step": 700 | |
| }, | |
| { | |
| "epoch": 1.929427430093209, | |
| "grad_norm": 13.981544494628906, | |
| "learning_rate": 9.508888888888889e-06, | |
| "loss": 0.4262, | |
| "step": 725 | |
| }, | |
| { | |
| "epoch": 1.996005326231691, | |
| "grad_norm": 10.144460678100586, | |
| "learning_rate": 9.453333333333335e-06, | |
| "loss": 0.4629, | |
| "step": 750 | |
| }, | |
| { | |
| "epoch": 2.0612516644474033, | |
| "grad_norm": 13.497632026672363, | |
| "learning_rate": 9.397777777777779e-06, | |
| "loss": 0.2155, | |
| "step": 775 | |
| }, | |
| { | |
| "epoch": 2.1278295605858855, | |
| "grad_norm": 7.0717878341674805, | |
| "learning_rate": 9.342222222222223e-06, | |
| "loss": 0.224, | |
| "step": 800 | |
| }, | |
| { | |
| "epoch": 2.1278295605858855, | |
| "eval_loss": 0.2817358374595642, | |
| "eval_runtime": 174.0748, | |
| "eval_samples_per_second": 2.045, | |
| "eval_steps_per_second": 0.511, | |
| "eval_wer": 0.23842018196856907, | |
| "step": 800 | |
| }, | |
| { | |
| "epoch": 2.1944074567243677, | |
| "grad_norm": 6.820193767547607, | |
| "learning_rate": 9.286666666666667e-06, | |
| "loss": 0.241, | |
| "step": 825 | |
| }, | |
| { | |
| "epoch": 2.2609853528628494, | |
| "grad_norm": 8.093194007873535, | |
| "learning_rate": 9.231111111111111e-06, | |
| "loss": 0.2301, | |
| "step": 850 | |
| }, | |
| { | |
| "epoch": 2.3275632490013316, | |
| "grad_norm": 6.863702774047852, | |
| "learning_rate": 9.175555555555557e-06, | |
| "loss": 0.2417, | |
| "step": 875 | |
| }, | |
| { | |
| "epoch": 2.3941411451398134, | |
| "grad_norm": 8.645722389221191, | |
| "learning_rate": 9.12e-06, | |
| "loss": 0.2456, | |
| "step": 900 | |
| }, | |
| { | |
| "epoch": 2.4607190412782955, | |
| "grad_norm": 6.348605632781982, | |
| "learning_rate": 9.064444444444447e-06, | |
| "loss": 0.2535, | |
| "step": 925 | |
| }, | |
| { | |
| "epoch": 2.5272969374167777, | |
| "grad_norm": 12.011576652526855, | |
| "learning_rate": 9.008888888888889e-06, | |
| "loss": 0.241, | |
| "step": 950 | |
| }, | |
| { | |
| "epoch": 2.5938748335552595, | |
| "grad_norm": 8.227922439575195, | |
| "learning_rate": 8.953333333333335e-06, | |
| "loss": 0.1779, | |
| "step": 975 | |
| }, | |
| { | |
| "epoch": 2.6604527296937417, | |
| "grad_norm": 10.178849220275879, | |
| "learning_rate": 8.897777777777779e-06, | |
| "loss": 0.2413, | |
| "step": 1000 | |
| }, | |
| { | |
| "epoch": 2.6604527296937417, | |
| "eval_loss": 0.25605687499046326, | |
| "eval_runtime": 166.4206, | |
| "eval_samples_per_second": 2.139, | |
| "eval_steps_per_second": 0.535, | |
| "eval_wer": 0.19520264681555005, | |
| "step": 1000 | |
| }, | |
| { | |
| "epoch": 2.7270306258322234, | |
| "grad_norm": 8.071171760559082, | |
| "learning_rate": 8.842222222222223e-06, | |
| "loss": 0.1885, | |
| "step": 1025 | |
| }, | |
| { | |
| "epoch": 2.7936085219707056, | |
| "grad_norm": 10.7284574508667, | |
| "learning_rate": 8.786666666666668e-06, | |
| "loss": 0.2392, | |
| "step": 1050 | |
| }, | |
| { | |
| "epoch": 2.860186418109188, | |
| "grad_norm": 6.85145902633667, | |
| "learning_rate": 8.73111111111111e-06, | |
| "loss": 0.2225, | |
| "step": 1075 | |
| }, | |
| { | |
| "epoch": 2.92676431424767, | |
| "grad_norm": 11.033931732177734, | |
| "learning_rate": 8.675555555555556e-06, | |
| "loss": 0.2241, | |
| "step": 1100 | |
| }, | |
| { | |
| "epoch": 2.993342210386152, | |
| "grad_norm": 7.744363307952881, | |
| "learning_rate": 8.62e-06, | |
| "loss": 0.1949, | |
| "step": 1125 | |
| }, | |
| { | |
| "epoch": 3.0585885486018642, | |
| "grad_norm": 7.1467084884643555, | |
| "learning_rate": 8.564444444444445e-06, | |
| "loss": 0.1003, | |
| "step": 1150 | |
| }, | |
| { | |
| "epoch": 3.125166444740346, | |
| "grad_norm": 4.110116004943848, | |
| "learning_rate": 8.50888888888889e-06, | |
| "loss": 0.0878, | |
| "step": 1175 | |
| }, | |
| { | |
| "epoch": 3.191744340878828, | |
| "grad_norm": 4.550497531890869, | |
| "learning_rate": 8.453333333333334e-06, | |
| "loss": 0.1036, | |
| "step": 1200 | |
| }, | |
| { | |
| "epoch": 3.191744340878828, | |
| "eval_loss": 0.2583100497722626, | |
| "eval_runtime": 166.1164, | |
| "eval_samples_per_second": 2.143, | |
| "eval_steps_per_second": 0.536, | |
| "eval_wer": 0.19044665012406947, | |
| "step": 1200 | |
| }, | |
| { | |
| "epoch": 3.2583222370173104, | |
| "grad_norm": 5.678071975708008, | |
| "learning_rate": 8.397777777777778e-06, | |
| "loss": 0.0985, | |
| "step": 1225 | |
| }, | |
| { | |
| "epoch": 3.324900133155792, | |
| "grad_norm": 6.956425189971924, | |
| "learning_rate": 8.342222222222222e-06, | |
| "loss": 0.0963, | |
| "step": 1250 | |
| }, | |
| { | |
| "epoch": 3.3914780292942743, | |
| "grad_norm": 5.823480606079102, | |
| "learning_rate": 8.286666666666668e-06, | |
| "loss": 0.1072, | |
| "step": 1275 | |
| }, | |
| { | |
| "epoch": 3.458055925432756, | |
| "grad_norm": 3.2603001594543457, | |
| "learning_rate": 8.231111111111112e-06, | |
| "loss": 0.0844, | |
| "step": 1300 | |
| }, | |
| { | |
| "epoch": 3.5246338215712383, | |
| "grad_norm": 3.718132257461548, | |
| "learning_rate": 8.175555555555556e-06, | |
| "loss": 0.1595, | |
| "step": 1325 | |
| }, | |
| { | |
| "epoch": 3.5912117177097205, | |
| "grad_norm": 7.773962497711182, | |
| "learning_rate": 8.120000000000002e-06, | |
| "loss": 0.1304, | |
| "step": 1350 | |
| }, | |
| { | |
| "epoch": 3.6577896138482027, | |
| "grad_norm": 8.074212074279785, | |
| "learning_rate": 8.064444444444444e-06, | |
| "loss": 0.0951, | |
| "step": 1375 | |
| }, | |
| { | |
| "epoch": 3.7243675099866844, | |
| "grad_norm": 3.417116165161133, | |
| "learning_rate": 8.00888888888889e-06, | |
| "loss": 0.1135, | |
| "step": 1400 | |
| }, | |
| { | |
| "epoch": 3.7243675099866844, | |
| "eval_loss": 0.2636599540710449, | |
| "eval_runtime": 173.0821, | |
| "eval_samples_per_second": 2.057, | |
| "eval_steps_per_second": 0.514, | |
| "eval_wer": 0.2119520264681555, | |
| "step": 1400 | |
| }, | |
| { | |
| "epoch": 3.790945406125166, | |
| "grad_norm": 3.8932766914367676, | |
| "learning_rate": 7.953333333333334e-06, | |
| "loss": 0.1156, | |
| "step": 1425 | |
| }, | |
| { | |
| "epoch": 3.8575233022636484, | |
| "grad_norm": 5.214465618133545, | |
| "learning_rate": 7.897777777777778e-06, | |
| "loss": 0.1334, | |
| "step": 1450 | |
| }, | |
| { | |
| "epoch": 3.9241011984021306, | |
| "grad_norm": 4.3109331130981445, | |
| "learning_rate": 7.842222222222224e-06, | |
| "loss": 0.1106, | |
| "step": 1475 | |
| }, | |
| { | |
| "epoch": 3.9906790945406128, | |
| "grad_norm": 5.095810413360596, | |
| "learning_rate": 7.786666666666666e-06, | |
| "loss": 0.1038, | |
| "step": 1500 | |
| }, | |
| { | |
| "epoch": 4.055925432756325, | |
| "grad_norm": 3.1624755859375, | |
| "learning_rate": 7.731111111111112e-06, | |
| "loss": 0.0806, | |
| "step": 1525 | |
| }, | |
| { | |
| "epoch": 4.1225033288948065, | |
| "grad_norm": 6.265628814697266, | |
| "learning_rate": 7.675555555555556e-06, | |
| "loss": 0.0626, | |
| "step": 1550 | |
| }, | |
| { | |
| "epoch": 4.189081225033289, | |
| "grad_norm": 3.3597617149353027, | |
| "learning_rate": 7.620000000000001e-06, | |
| "loss": 0.0594, | |
| "step": 1575 | |
| }, | |
| { | |
| "epoch": 4.255659121171771, | |
| "grad_norm": 6.631519794464111, | |
| "learning_rate": 7.564444444444446e-06, | |
| "loss": 0.057, | |
| "step": 1600 | |
| }, | |
| { | |
| "epoch": 4.255659121171771, | |
| "eval_loss": 0.273118257522583, | |
| "eval_runtime": 172.5497, | |
| "eval_samples_per_second": 2.063, | |
| "eval_steps_per_second": 0.516, | |
| "eval_wer": 0.20967741935483872, | |
| "step": 1600 | |
| }, | |
| { | |
| "epoch": 4.255659121171771, | |
| "step": 1600, | |
| "total_flos": 1.303516587884544e+19, | |
| "train_loss": 0.610494866669178, | |
| "train_runtime": 3998.2703, | |
| "train_samples_per_second": 10.004, | |
| "train_steps_per_second": 1.251 | |
| } | |
| ], | |
| "logging_steps": 25, | |
| "max_steps": 5000, | |
| "num_input_tokens_seen": 0, | |
| "num_train_epochs": 14, | |
| "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": 1.303516587884544e+19, | |
| "train_batch_size": 4, | |
| "trial_name": null, | |
| "trial_params": null | |
| } | |