| { | |
| "best_metric": 8.39864864864865, | |
| "best_model_checkpoint": "./checkpoint-10000", | |
| "epoch": 10.090817356205852, | |
| "global_step": 10000, | |
| "is_hyper_param_search": false, | |
| "is_local_process_zero": true, | |
| "is_world_process_zero": true, | |
| "log_history": [ | |
| { | |
| "epoch": 0.5, | |
| "learning_rate": 9.940000000000001e-06, | |
| "loss": 0.3462, | |
| "step": 500 | |
| }, | |
| { | |
| "epoch": 1.01, | |
| "learning_rate": 9.476842105263158e-06, | |
| "loss": 0.1691, | |
| "step": 1000 | |
| }, | |
| { | |
| "epoch": 1.01, | |
| "eval_cer": 6.350948206752441, | |
| "eval_loss": 0.1870781034231186, | |
| "eval_runtime": 4246.4995, | |
| "eval_samples_per_second": 1.084, | |
| "eval_steps_per_second": 0.068, | |
| "eval_wer": 10.173986486486486, | |
| "step": 1000 | |
| }, | |
| { | |
| "epoch": 1.51, | |
| "learning_rate": 8.950526315789474e-06, | |
| "loss": 0.097, | |
| "step": 1500 | |
| }, | |
| { | |
| "epoch": 2.02, | |
| "learning_rate": 8.42421052631579e-06, | |
| "loss": 0.0916, | |
| "step": 2000 | |
| }, | |
| { | |
| "epoch": 2.02, | |
| "eval_cer": 5.5499336320984085, | |
| "eval_loss": 0.16914600133895874, | |
| "eval_runtime": 4299.9346, | |
| "eval_samples_per_second": 1.071, | |
| "eval_steps_per_second": 0.067, | |
| "eval_wer": 8.97972972972973, | |
| "step": 2000 | |
| }, | |
| { | |
| "epoch": 2.52, | |
| "learning_rate": 7.897894736842106e-06, | |
| "loss": 0.0453, | |
| "step": 2500 | |
| }, | |
| { | |
| "epoch": 3.03, | |
| "learning_rate": 7.371578947368422e-06, | |
| "loss": 0.0452, | |
| "step": 3000 | |
| }, | |
| { | |
| "epoch": 3.03, | |
| "eval_cer": 5.58673167654519, | |
| "eval_loss": 0.19017226994037628, | |
| "eval_runtime": 4203.2916, | |
| "eval_samples_per_second": 1.095, | |
| "eval_steps_per_second": 0.069, | |
| "eval_wer": 8.98141891891892, | |
| "step": 3000 | |
| }, | |
| { | |
| "epoch": 3.53, | |
| "learning_rate": 6.845263157894737e-06, | |
| "loss": 0.0214, | |
| "step": 3500 | |
| }, | |
| { | |
| "epoch": 4.04, | |
| "learning_rate": 6.318947368421054e-06, | |
| "loss": 0.0213, | |
| "step": 4000 | |
| }, | |
| { | |
| "epoch": 4.04, | |
| "eval_cer": 5.653099578136705, | |
| "eval_loss": 0.20616890490055084, | |
| "eval_runtime": 4260.0725, | |
| "eval_samples_per_second": 1.081, | |
| "eval_steps_per_second": 0.068, | |
| "eval_wer": 8.9375, | |
| "step": 4000 | |
| }, | |
| { | |
| "epoch": 4.54, | |
| "learning_rate": 5.7926315789473685e-06, | |
| "loss": 0.0096, | |
| "step": 4500 | |
| }, | |
| { | |
| "epoch": 5.05, | |
| "learning_rate": 5.266315789473685e-06, | |
| "loss": 0.0096, | |
| "step": 5000 | |
| }, | |
| { | |
| "epoch": 5.05, | |
| "eval_cer": 5.620244181309221, | |
| "eval_loss": 0.22839371860027313, | |
| "eval_runtime": 4320.8285, | |
| "eval_samples_per_second": 1.066, | |
| "eval_steps_per_second": 0.067, | |
| "eval_wer": 8.733108108108107, | |
| "step": 5000 | |
| }, | |
| { | |
| "epoch": 5.55, | |
| "learning_rate": 4.74e-06, | |
| "loss": 0.0045, | |
| "step": 5500 | |
| }, | |
| { | |
| "epoch": 6.05, | |
| "learning_rate": 4.213684210526316e-06, | |
| "loss": 0.0041, | |
| "step": 6000 | |
| }, | |
| { | |
| "epoch": 6.05, | |
| "eval_cer": 5.300889724146089, | |
| "eval_loss": 0.2394615262746811, | |
| "eval_runtime": 4207.8926, | |
| "eval_samples_per_second": 1.094, | |
| "eval_steps_per_second": 0.068, | |
| "eval_wer": 8.505067567567567, | |
| "step": 6000 | |
| }, | |
| { | |
| "epoch": 6.56, | |
| "learning_rate": 3.687368421052632e-06, | |
| "loss": 0.0024, | |
| "step": 6500 | |
| }, | |
| { | |
| "epoch": 7.06, | |
| "learning_rate": 3.1610526315789474e-06, | |
| "loss": 0.0022, | |
| "step": 7000 | |
| }, | |
| { | |
| "epoch": 7.06, | |
| "eval_cer": 5.363972086054855, | |
| "eval_loss": 0.2534642517566681, | |
| "eval_runtime": 4214.0551, | |
| "eval_samples_per_second": 1.093, | |
| "eval_steps_per_second": 0.068, | |
| "eval_wer": 8.550675675675675, | |
| "step": 7000 | |
| }, | |
| { | |
| "epoch": 7.57, | |
| "learning_rate": 2.6347368421052633e-06, | |
| "loss": 0.0013, | |
| "step": 7500 | |
| }, | |
| { | |
| "epoch": 8.07, | |
| "learning_rate": 2.1084210526315792e-06, | |
| "loss": 0.001, | |
| "step": 8000 | |
| }, | |
| { | |
| "epoch": 8.07, | |
| "eval_cer": 5.379085568595497, | |
| "eval_loss": 0.2655620574951172, | |
| "eval_runtime": 4212.0826, | |
| "eval_samples_per_second": 1.093, | |
| "eval_steps_per_second": 0.068, | |
| "eval_wer": 8.555743243243242, | |
| "step": 8000 | |
| }, | |
| { | |
| "epoch": 8.58, | |
| "learning_rate": 1.5821052631578948e-06, | |
| "loss": 0.0007, | |
| "step": 8500 | |
| }, | |
| { | |
| "epoch": 9.08, | |
| "learning_rate": 1.0557894736842105e-06, | |
| "loss": 0.0006, | |
| "step": 9000 | |
| }, | |
| { | |
| "epoch": 9.08, | |
| "eval_cer": 5.273948298747552, | |
| "eval_loss": 0.2720721364021301, | |
| "eval_runtime": 4204.0827, | |
| "eval_samples_per_second": 1.095, | |
| "eval_steps_per_second": 0.069, | |
| "eval_wer": 8.403716216216216, | |
| "step": 9000 | |
| }, | |
| { | |
| "epoch": 9.59, | |
| "learning_rate": 5.294736842105263e-07, | |
| "loss": 0.0005, | |
| "step": 9500 | |
| }, | |
| { | |
| "epoch": 10.09, | |
| "learning_rate": 3.1578947368421054e-09, | |
| "loss": 0.0004, | |
| "step": 10000 | |
| }, | |
| { | |
| "epoch": 10.09, | |
| "eval_cer": 5.2581777082703605, | |
| "eval_loss": 0.2789727747440338, | |
| "eval_runtime": 4206.3116, | |
| "eval_samples_per_second": 1.095, | |
| "eval_steps_per_second": 0.068, | |
| "eval_wer": 8.39864864864865, | |
| "step": 10000 | |
| }, | |
| { | |
| "epoch": 10.09, | |
| "step": 10000, | |
| "total_flos": 3.265119220137984e+20, | |
| "train_loss": 0.04369454556107521, | |
| "train_runtime": 84974.4249, | |
| "train_samples_per_second": 3.766, | |
| "train_steps_per_second": 0.118 | |
| } | |
| ], | |
| "max_steps": 10000, | |
| "num_train_epochs": 11, | |
| "total_flos": 3.265119220137984e+20, | |
| "trial_name": null, | |
| "trial_params": null | |
| } | |