| { | |
| "best_metric": null, | |
| "best_model_checkpoint": null, | |
| "epoch": 60.0, | |
| "eval_steps": 500, | |
| "global_step": 7500, | |
| "is_hyper_param_search": false, | |
| "is_local_process_zero": true, | |
| "is_world_process_zero": true, | |
| "log_history": [ | |
| { | |
| "epoch": 4.0, | |
| "grad_norm": 1.0876199007034302, | |
| "learning_rate": 5e-05, | |
| "loss": 0.1167, | |
| "step": 500 | |
| }, | |
| { | |
| "epoch": 4.0, | |
| "eval_cer": 0.03410275077441366, | |
| "eval_loss": 0.07084234058856964, | |
| "eval_runtime": 30.3562, | |
| "eval_samples_per_second": 49.347, | |
| "eval_steps_per_second": 6.193, | |
| "eval_wer": 0.10315320595475325, | |
| "step": 500 | |
| }, | |
| { | |
| "epoch": 8.0, | |
| "grad_norm": 4.202278137207031, | |
| "learning_rate": 0.0001, | |
| "loss": 0.1195, | |
| "step": 1000 | |
| }, | |
| { | |
| "epoch": 8.0, | |
| "eval_cer": 0.036258261601929964, | |
| "eval_loss": 0.07872834801673889, | |
| "eval_runtime": 30.3905, | |
| "eval_samples_per_second": 49.292, | |
| "eval_steps_per_second": 6.186, | |
| "eval_wer": 0.10690423162583519, | |
| "step": 1000 | |
| }, | |
| { | |
| "epoch": 12.0, | |
| "grad_norm": 1.2584967613220215, | |
| "learning_rate": 9.230769230769232e-05, | |
| "loss": 0.1259, | |
| "step": 1500 | |
| }, | |
| { | |
| "epoch": 12.0, | |
| "eval_cer": 0.036558034630922304, | |
| "eval_loss": 0.08242663741111755, | |
| "eval_runtime": 30.1208, | |
| "eval_samples_per_second": 49.733, | |
| "eval_steps_per_second": 6.242, | |
| "eval_wer": 0.10708006095416715, | |
| "step": 1500 | |
| }, | |
| { | |
| "epoch": 16.0, | |
| "grad_norm": 0.9028881788253784, | |
| "learning_rate": 8.461538461538461e-05, | |
| "loss": 0.1151, | |
| "step": 2000 | |
| }, | |
| { | |
| "epoch": 16.0, | |
| "eval_cer": 0.03672933350463221, | |
| "eval_loss": 0.08765117079019547, | |
| "eval_runtime": 30.2298, | |
| "eval_samples_per_second": 49.554, | |
| "eval_steps_per_second": 6.219, | |
| "eval_wer": 0.10795920759582699, | |
| "step": 2000 | |
| }, | |
| { | |
| "epoch": 20.0, | |
| "grad_norm": 1.2381340265274048, | |
| "learning_rate": 7.693846153846154e-05, | |
| "loss": 0.1082, | |
| "step": 2500 | |
| }, | |
| { | |
| "epoch": 20.0, | |
| "eval_cer": 0.03688635747219962, | |
| "eval_loss": 0.08228232711553574, | |
| "eval_runtime": 30.1272, | |
| "eval_samples_per_second": 49.723, | |
| "eval_steps_per_second": 6.24, | |
| "eval_wer": 0.10819364670026961, | |
| "step": 2500 | |
| }, | |
| { | |
| "epoch": 24.0, | |
| "grad_norm": 1.054206371307373, | |
| "learning_rate": 6.924615384615386e-05, | |
| "loss": 0.0975, | |
| "step": 3000 | |
| }, | |
| { | |
| "epoch": 24.0, | |
| "eval_cer": 0.03688635747219962, | |
| "eval_loss": 0.09040121734142303, | |
| "eval_runtime": 30.3039, | |
| "eval_samples_per_second": 49.433, | |
| "eval_steps_per_second": 6.204, | |
| "eval_wer": 0.1080764271480483, | |
| "step": 3000 | |
| }, | |
| { | |
| "epoch": 28.0, | |
| "grad_norm": 0.7761673331260681, | |
| "learning_rate": 6.156923076923077e-05, | |
| "loss": 0.1123, | |
| "step": 3500 | |
| }, | |
| { | |
| "epoch": 28.0, | |
| "eval_cer": 0.036800708035344666, | |
| "eval_loss": 0.08693404495716095, | |
| "eval_runtime": 30.3844, | |
| "eval_samples_per_second": 49.302, | |
| "eval_steps_per_second": 6.187, | |
| "eval_wer": 0.10778337826749501, | |
| "step": 3500 | |
| }, | |
| { | |
| "epoch": 32.0, | |
| "grad_norm": 1.127631425857544, | |
| "learning_rate": 5.387692307692308e-05, | |
| "loss": 0.0912, | |
| "step": 4000 | |
| }, | |
| { | |
| "epoch": 32.0, | |
| "eval_cer": 0.03721468031347694, | |
| "eval_loss": 0.08634809404611588, | |
| "eval_runtime": 30.4442, | |
| "eval_samples_per_second": 49.205, | |
| "eval_steps_per_second": 6.175, | |
| "eval_wer": 0.10825225647638026, | |
| "step": 4000 | |
| }, | |
| { | |
| "epoch": 36.0, | |
| "grad_norm": 0.5615821480751038, | |
| "learning_rate": 4.618461538461539e-05, | |
| "loss": 0.0829, | |
| "step": 4500 | |
| }, | |
| { | |
| "epoch": 36.0, | |
| "eval_cer": 0.03658658444320729, | |
| "eval_loss": 0.08522398769855499, | |
| "eval_runtime": 30.4454, | |
| "eval_samples_per_second": 49.203, | |
| "eval_steps_per_second": 6.175, | |
| "eval_wer": 0.1080764271480483, | |
| "step": 4500 | |
| }, | |
| { | |
| "epoch": 40.0, | |
| "grad_norm": 0.9052166938781738, | |
| "learning_rate": 3.849230769230769e-05, | |
| "loss": 0.0824, | |
| "step": 5000 | |
| }, | |
| { | |
| "epoch": 40.0, | |
| "eval_cer": 0.03664368406777725, | |
| "eval_loss": 0.08616316318511963, | |
| "eval_runtime": 30.3545, | |
| "eval_samples_per_second": 49.35, | |
| "eval_steps_per_second": 6.193, | |
| "eval_wer": 0.10760754893916305, | |
| "step": 5000 | |
| }, | |
| { | |
| "epoch": 44.0, | |
| "grad_norm": 0.9279913902282715, | |
| "learning_rate": 3.08e-05, | |
| "loss": 0.0841, | |
| "step": 5500 | |
| }, | |
| { | |
| "epoch": 44.0, | |
| "eval_cer": 0.03681498294148716, | |
| "eval_loss": 0.08898201584815979, | |
| "eval_runtime": 30.4415, | |
| "eval_samples_per_second": 49.209, | |
| "eval_steps_per_second": 6.176, | |
| "eval_wer": 0.10778337826749501, | |
| "step": 5500 | |
| }, | |
| { | |
| "epoch": 48.0, | |
| "grad_norm": 0.42622652649879456, | |
| "learning_rate": 2.3107692307692308e-05, | |
| "loss": 0.0768, | |
| "step": 6000 | |
| }, | |
| { | |
| "epoch": 48.0, | |
| "eval_cer": 0.036358185944927415, | |
| "eval_loss": 0.08734322339296341, | |
| "eval_runtime": 30.364, | |
| "eval_samples_per_second": 49.335, | |
| "eval_steps_per_second": 6.192, | |
| "eval_wer": 0.10719728050638846, | |
| "step": 6000 | |
| }, | |
| { | |
| "epoch": 52.0, | |
| "grad_norm": 0.6296827793121338, | |
| "learning_rate": 1.5415384615384616e-05, | |
| "loss": 0.0692, | |
| "step": 6500 | |
| }, | |
| { | |
| "epoch": 52.0, | |
| "eval_cer": 0.036243986695787475, | |
| "eval_loss": 0.08950727432966232, | |
| "eval_runtime": 30.3163, | |
| "eval_samples_per_second": 49.412, | |
| "eval_steps_per_second": 6.201, | |
| "eval_wer": 0.10643535341694994, | |
| "step": 6500 | |
| }, | |
| { | |
| "epoch": 56.0, | |
| "grad_norm": 0.5045118927955627, | |
| "learning_rate": 7.723076923076922e-06, | |
| "loss": 0.0714, | |
| "step": 7000 | |
| }, | |
| { | |
| "epoch": 56.0, | |
| "eval_cer": 0.036358185944927415, | |
| "eval_loss": 0.08889112621545792, | |
| "eval_runtime": 30.18, | |
| "eval_samples_per_second": 49.636, | |
| "eval_steps_per_second": 6.229, | |
| "eval_wer": 0.1064939631930606, | |
| "step": 7000 | |
| }, | |
| { | |
| "epoch": 60.0, | |
| "grad_norm": 1.5385302305221558, | |
| "learning_rate": 3.076923076923077e-08, | |
| "loss": 0.0649, | |
| "step": 7500 | |
| }, | |
| { | |
| "epoch": 60.0, | |
| "eval_cer": 0.036301086320357445, | |
| "eval_loss": 0.08928388357162476, | |
| "eval_runtime": 30.1433, | |
| "eval_samples_per_second": 49.696, | |
| "eval_steps_per_second": 6.237, | |
| "eval_wer": 0.10655257296917126, | |
| "step": 7500 | |
| } | |
| ], | |
| "logging_steps": 500, | |
| "max_steps": 7500, | |
| "num_input_tokens_seen": 0, | |
| "num_train_epochs": 60, | |
| "save_steps": 500, | |
| "stateful_callbacks": { | |
| "TrainerControl": { | |
| "args": { | |
| "should_epoch_stop": false, | |
| "should_evaluate": false, | |
| "should_log": false, | |
| "should_save": true, | |
| "should_training_stop": true | |
| }, | |
| "attributes": {} | |
| } | |
| }, | |
| "total_flos": 1.3987580856546666e+19, | |
| "train_batch_size": 48, | |
| "trial_name": null, | |
| "trial_params": null | |
| } | |