| { | |
| "best_metric": null, | |
| "best_model_checkpoint": null, | |
| "epoch": 137.93103448275863, | |
| "global_step": 4000, | |
| "is_hyper_param_search": false, | |
| "is_local_process_zero": true, | |
| "is_world_process_zero": true, | |
| "log_history": [ | |
| { | |
| "epoch": 3.45, | |
| "learning_rate": 7.2e-06, | |
| "loss": 16.2031, | |
| "step": 100 | |
| }, | |
| { | |
| "epoch": 6.9, | |
| "learning_rate": 1.47e-05, | |
| "loss": 6.6774, | |
| "step": 200 | |
| }, | |
| { | |
| "epoch": 10.34, | |
| "learning_rate": 2.2199999999999998e-05, | |
| "loss": 4.7603, | |
| "step": 300 | |
| }, | |
| { | |
| "epoch": 13.79, | |
| "learning_rate": 2.97e-05, | |
| "loss": 3.9198, | |
| "step": 400 | |
| }, | |
| { | |
| "epoch": 17.24, | |
| "learning_rate": 3.7199999999999996e-05, | |
| "loss": 3.5558, | |
| "step": 500 | |
| }, | |
| { | |
| "epoch": 17.24, | |
| "eval_loss": 3.5216922760009766, | |
| "eval_runtime": 2.4797, | |
| "eval_samples_per_second": 108.882, | |
| "eval_steps_per_second": 3.629, | |
| "eval_wer": 0.9812865497076023, | |
| "step": 500 | |
| }, | |
| { | |
| "epoch": 20.69, | |
| "learning_rate": 4.4699999999999996e-05, | |
| "loss": 3.4799, | |
| "step": 600 | |
| }, | |
| { | |
| "epoch": 24.14, | |
| "learning_rate": 5.2199999999999995e-05, | |
| "loss": 3.4271, | |
| "step": 700 | |
| }, | |
| { | |
| "epoch": 27.59, | |
| "learning_rate": 5.97e-05, | |
| "loss": 3.2779, | |
| "step": 800 | |
| }, | |
| { | |
| "epoch": 31.03, | |
| "learning_rate": 6.72e-05, | |
| "loss": 2.9317, | |
| "step": 900 | |
| }, | |
| { | |
| "epoch": 34.48, | |
| "learning_rate": 7.47e-05, | |
| "loss": 2.2202, | |
| "step": 1000 | |
| }, | |
| { | |
| "epoch": 34.48, | |
| "eval_loss": 0.9991578459739685, | |
| "eval_runtime": 1.9665, | |
| "eval_samples_per_second": 137.299, | |
| "eval_steps_per_second": 4.577, | |
| "eval_wer": 0.2637958532695375, | |
| "step": 1000 | |
| }, | |
| { | |
| "epoch": 37.93, | |
| "learning_rate": 7.285074626865672e-05, | |
| "loss": 1.835, | |
| "step": 1100 | |
| }, | |
| { | |
| "epoch": 41.38, | |
| "learning_rate": 7.061194029850745e-05, | |
| "loss": 1.653, | |
| "step": 1200 | |
| }, | |
| { | |
| "epoch": 44.83, | |
| "learning_rate": 6.83731343283582e-05, | |
| "loss": 1.5243, | |
| "step": 1300 | |
| }, | |
| { | |
| "epoch": 48.28, | |
| "learning_rate": 6.613432835820895e-05, | |
| "loss": 1.4343, | |
| "step": 1400 | |
| }, | |
| { | |
| "epoch": 51.72, | |
| "learning_rate": 6.38955223880597e-05, | |
| "loss": 1.376, | |
| "step": 1500 | |
| }, | |
| { | |
| "epoch": 51.72, | |
| "eval_loss": 0.5396047830581665, | |
| "eval_runtime": 1.8956, | |
| "eval_samples_per_second": 142.435, | |
| "eval_steps_per_second": 4.748, | |
| "eval_wer": 0.1946836788942052, | |
| "step": 1500 | |
| }, | |
| { | |
| "epoch": 55.17, | |
| "learning_rate": 6.165671641791044e-05, | |
| "loss": 1.3168, | |
| "step": 1600 | |
| }, | |
| { | |
| "epoch": 58.62, | |
| "learning_rate": 5.941791044776119e-05, | |
| "loss": 1.2773, | |
| "step": 1700 | |
| }, | |
| { | |
| "epoch": 62.07, | |
| "learning_rate": 5.717910447761193e-05, | |
| "loss": 1.2455, | |
| "step": 1800 | |
| }, | |
| { | |
| "epoch": 65.52, | |
| "learning_rate": 5.494029850746268e-05, | |
| "loss": 1.1881, | |
| "step": 1900 | |
| }, | |
| { | |
| "epoch": 68.97, | |
| "learning_rate": 5.272388059701492e-05, | |
| "loss": 1.1565, | |
| "step": 2000 | |
| }, | |
| { | |
| "epoch": 68.97, | |
| "eval_loss": 0.4707144498825073, | |
| "eval_runtime": 1.8963, | |
| "eval_samples_per_second": 142.381, | |
| "eval_steps_per_second": 4.746, | |
| "eval_wer": 0.1773524720893142, | |
| "step": 2000 | |
| }, | |
| { | |
| "epoch": 72.41, | |
| "learning_rate": 5.048507462686567e-05, | |
| "loss": 1.1295, | |
| "step": 2100 | |
| }, | |
| { | |
| "epoch": 75.86, | |
| "learning_rate": 4.8268656716417906e-05, | |
| "loss": 1.0983, | |
| "step": 2200 | |
| }, | |
| { | |
| "epoch": 79.31, | |
| "learning_rate": 4.6029850746268655e-05, | |
| "loss": 1.0716, | |
| "step": 2300 | |
| }, | |
| { | |
| "epoch": 82.76, | |
| "learning_rate": 4.37910447761194e-05, | |
| "loss": 1.0582, | |
| "step": 2400 | |
| }, | |
| { | |
| "epoch": 86.21, | |
| "learning_rate": 4.155223880597015e-05, | |
| "loss": 1.0112, | |
| "step": 2500 | |
| }, | |
| { | |
| "epoch": 86.21, | |
| "eval_loss": 0.46029242873191833, | |
| "eval_runtime": 1.9007, | |
| "eval_samples_per_second": 142.05, | |
| "eval_steps_per_second": 4.735, | |
| "eval_wer": 0.17533227006911217, | |
| "step": 2500 | |
| }, | |
| { | |
| "epoch": 89.66, | |
| "learning_rate": 3.931343283582089e-05, | |
| "loss": 0.9962, | |
| "step": 2600 | |
| }, | |
| { | |
| "epoch": 93.1, | |
| "learning_rate": 3.707462686567164e-05, | |
| "loss": 0.9746, | |
| "step": 2700 | |
| }, | |
| { | |
| "epoch": 96.55, | |
| "learning_rate": 3.483582089552238e-05, | |
| "loss": 0.9592, | |
| "step": 2800 | |
| }, | |
| { | |
| "epoch": 100.0, | |
| "learning_rate": 3.259701492537313e-05, | |
| "loss": 0.9364, | |
| "step": 2900 | |
| }, | |
| { | |
| "epoch": 103.45, | |
| "learning_rate": 3.0358208955223878e-05, | |
| "loss": 0.9387, | |
| "step": 3000 | |
| }, | |
| { | |
| "epoch": 103.45, | |
| "eval_loss": 0.4466392397880554, | |
| "eval_runtime": 1.9017, | |
| "eval_samples_per_second": 141.977, | |
| "eval_steps_per_second": 4.733, | |
| "eval_wer": 0.1721424774056353, | |
| "step": 3000 | |
| }, | |
| { | |
| "epoch": 106.9, | |
| "learning_rate": 2.811940298507462e-05, | |
| "loss": 0.9223, | |
| "step": 3100 | |
| }, | |
| { | |
| "epoch": 110.34, | |
| "learning_rate": 2.588059701492537e-05, | |
| "loss": 0.8839, | |
| "step": 3200 | |
| }, | |
| { | |
| "epoch": 113.79, | |
| "learning_rate": 2.364179104477612e-05, | |
| "loss": 0.8845, | |
| "step": 3300 | |
| }, | |
| { | |
| "epoch": 117.24, | |
| "learning_rate": 2.1402985074626863e-05, | |
| "loss": 0.8604, | |
| "step": 3400 | |
| }, | |
| { | |
| "epoch": 120.69, | |
| "learning_rate": 1.9164179104477612e-05, | |
| "loss": 0.8526, | |
| "step": 3500 | |
| }, | |
| { | |
| "epoch": 120.69, | |
| "eval_loss": 0.43604278564453125, | |
| "eval_runtime": 1.8975, | |
| "eval_samples_per_second": 142.29, | |
| "eval_steps_per_second": 4.743, | |
| "eval_wer": 0.17044125465178098, | |
| "step": 3500 | |
| }, | |
| { | |
| "epoch": 124.14, | |
| "learning_rate": 1.6925373134328355e-05, | |
| "loss": 0.8311, | |
| "step": 3600 | |
| }, | |
| { | |
| "epoch": 127.59, | |
| "learning_rate": 1.4686567164179104e-05, | |
| "loss": 0.832, | |
| "step": 3700 | |
| }, | |
| { | |
| "epoch": 131.03, | |
| "learning_rate": 1.244776119402985e-05, | |
| "loss": 0.812, | |
| "step": 3800 | |
| }, | |
| { | |
| "epoch": 134.48, | |
| "learning_rate": 1.0208955223880597e-05, | |
| "loss": 0.8016, | |
| "step": 3900 | |
| }, | |
| { | |
| "epoch": 137.93, | |
| "learning_rate": 7.970149253731343e-06, | |
| "loss": 0.8018, | |
| "step": 4000 | |
| }, | |
| { | |
| "epoch": 137.93, | |
| "eval_loss": 0.437362939119339, | |
| "eval_runtime": 1.9635, | |
| "eval_samples_per_second": 137.506, | |
| "eval_steps_per_second": 4.584, | |
| "eval_wer": 0.1720361509835194, | |
| "step": 4000 | |
| } | |
| ], | |
| "max_steps": 4350, | |
| "num_train_epochs": 150, | |
| "total_flos": 1.7217444822080553e+19, | |
| "trial_name": null, | |
| "trial_params": null | |
| } | |