{ "best_metric": 8.312998407517968, "best_model_checkpoint": "./fine-tuned-whisper/openai/whisper-small/lr_1e-05_warmup_500_epochs_0_batch_32_grad_acc_1_max_steps_10000/dataset_all/2023-09-09_04-06-40/checkpoint-1500", "epoch": 1.9280205655526992, "eval_steps": 500, "global_step": 1500, "is_hyper_param_search": false, "is_local_process_zero": true, "is_world_process_zero": true, "log_history": [ { "epoch": 0.13, "learning_rate": 1.94e-06, "loss": 1.8115, "step": 100 }, { "epoch": 0.26, "learning_rate": 3.94e-06, "loss": 0.6627, "step": 200 }, { "epoch": 0.39, "learning_rate": 5.94e-06, "loss": 0.5014, "step": 300 }, { "epoch": 0.51, "learning_rate": 7.94e-06, "loss": 0.3889, "step": 400 }, { "epoch": 0.64, "learning_rate": 9.940000000000001e-06, "loss": 0.3855, "step": 500 }, { "epoch": 0.64, "eval_loss": 0.23195046186447144, "eval_runtime": 720.118, "eval_samples_per_second": 5.241, "eval_steps_per_second": 0.328, "eval_wer": 9.69718621140653, "step": 500 }, { "epoch": 0.77, "learning_rate": 9.897894736842107e-06, "loss": 0.3682, "step": 600 }, { "epoch": 0.9, "learning_rate": 9.79263157894737e-06, "loss": 0.3516, "step": 700 }, { "epoch": 1.03, "learning_rate": 9.687368421052632e-06, "loss": 0.3261, "step": 800 }, { "epoch": 1.16, "learning_rate": 9.582105263157897e-06, "loss": 0.2881, "step": 900 }, { "epoch": 1.29, "learning_rate": 9.476842105263158e-06, "loss": 0.2926, "step": 1000 }, { "epoch": 1.29, "eval_loss": 0.21654152870178223, "eval_runtime": 710.9642, "eval_samples_per_second": 5.308, "eval_steps_per_second": 0.332, "eval_wer": 8.38929222348033, "step": 1000 }, { "epoch": 1.41, "learning_rate": 9.371578947368421e-06, "loss": 0.2908, "step": 1100 }, { "epoch": 1.54, "learning_rate": 9.266315789473685e-06, "loss": 0.2792, "step": 1200 }, { "epoch": 1.67, "learning_rate": 9.161052631578948e-06, "loss": 0.2854, "step": 1300 }, { "epoch": 1.8, "learning_rate": 9.055789473684211e-06, "loss": 0.2913, "step": 1400 }, { "epoch": 1.93, "learning_rate": 8.950526315789474e-06, "loss": 0.297, "step": 1500 }, { "epoch": 1.93, "eval_loss": 0.21313807368278503, "eval_runtime": 710.8884, "eval_samples_per_second": 5.309, "eval_steps_per_second": 0.332, "eval_wer": 8.312998407517968, "step": 1500 } ], "logging_steps": 100, "max_steps": 10000, "num_train_epochs": 13, "save_steps": 500, "total_flos": 1.384719326355456e+19, "trial_name": null, "trial_params": null }