| { | |
| "best_global_step": 774, | |
| "best_metric": 0.3903301886792453, | |
| "best_model_checkpoint": "./whisper-sardu-checkpoints/checkpoint-774", | |
| "epoch": 7.0, | |
| "eval_steps": 100, | |
| "global_step": 903, | |
| "is_hyper_param_search": false, | |
| "is_local_process_zero": true, | |
| "is_world_process_zero": true, | |
| "log_history": [ | |
| { | |
| "epoch": 0.38910505836575876, | |
| "grad_norm": 11.551776885986328, | |
| "learning_rate": 1.3800000000000002e-05, | |
| "loss": 2.7016, | |
| "step": 50 | |
| }, | |
| { | |
| "epoch": 0.7782101167315175, | |
| "grad_norm": 8.857993125915527, | |
| "learning_rate": 2.88e-05, | |
| "loss": 0.9443, | |
| "step": 100 | |
| }, | |
| { | |
| "epoch": 1.0, | |
| "eval_loss": 0.5204761028289795, | |
| "eval_runtime": 546.4632, | |
| "eval_samples_per_second": 0.375, | |
| "eval_steps_per_second": 0.048, | |
| "eval_wer": 0.5660377358490566, | |
| "step": 129 | |
| }, | |
| { | |
| "epoch": 1.1634241245136188, | |
| "grad_norm": 5.102855205535889, | |
| "learning_rate": 2.9953507804049928e-05, | |
| "loss": 0.4937, | |
| "step": 150 | |
| }, | |
| { | |
| "epoch": 1.5525291828793775, | |
| "grad_norm": 5.129488945007324, | |
| "learning_rate": 2.979785942784645e-05, | |
| "loss": 0.3193, | |
| "step": 200 | |
| }, | |
| { | |
| "epoch": 1.9416342412451362, | |
| "grad_norm": 4.888234615325928, | |
| "learning_rate": 2.953384299136067e-05, | |
| "loss": 0.2984, | |
| "step": 250 | |
| }, | |
| { | |
| "epoch": 2.0, | |
| "eval_loss": 0.41961559653282166, | |
| "eval_runtime": 542.4905, | |
| "eval_samples_per_second": 0.378, | |
| "eval_steps_per_second": 0.048, | |
| "eval_wer": 0.5660377358490566, | |
| "step": 258 | |
| }, | |
| { | |
| "epoch": 2.3268482490272375, | |
| "grad_norm": 3.4571590423583984, | |
| "learning_rate": 2.9163391946470812e-05, | |
| "loss": 0.1334, | |
| "step": 300 | |
| }, | |
| { | |
| "epoch": 2.7159533073929962, | |
| "grad_norm": 3.795649528503418, | |
| "learning_rate": 2.8689219189684655e-05, | |
| "loss": 0.1045, | |
| "step": 350 | |
| }, | |
| { | |
| "epoch": 3.0, | |
| "eval_loss": 0.4368809163570404, | |
| "eval_runtime": 437.9643, | |
| "eval_samples_per_second": 0.468, | |
| "eval_steps_per_second": 0.059, | |
| "eval_wer": 0.42452830188679247, | |
| "step": 387 | |
| }, | |
| { | |
| "epoch": 3.1011673151750974, | |
| "grad_norm": 1.8026232719421387, | |
| "learning_rate": 2.811479719498683e-05, | |
| "loss": 0.1069, | |
| "step": 400 | |
| }, | |
| { | |
| "epoch": 3.490272373540856, | |
| "grad_norm": 3.5268654823303223, | |
| "learning_rate": 2.7444332584129128e-05, | |
| "loss": 0.0503, | |
| "step": 450 | |
| }, | |
| { | |
| "epoch": 3.8793774319066148, | |
| "grad_norm": 2.7765798568725586, | |
| "learning_rate": 2.6682735320591115e-05, | |
| "loss": 0.0516, | |
| "step": 500 | |
| }, | |
| { | |
| "epoch": 4.0, | |
| "eval_loss": 0.4473298192024231, | |
| "eval_runtime": 439.5871, | |
| "eval_samples_per_second": 0.466, | |
| "eval_steps_per_second": 0.059, | |
| "eval_wer": 0.4033018867924528, | |
| "step": 516 | |
| }, | |
| { | |
| "epoch": 4.264591439688716, | |
| "grad_norm": 3.099731206893921, | |
| "learning_rate": 2.5835582752810806e-05, | |
| "loss": 0.0333, | |
| "step": 550 | |
| }, | |
| { | |
| "epoch": 4.653696498054475, | |
| "grad_norm": 1.8574504852294922, | |
| "learning_rate": 2.4909078770004833e-05, | |
| "loss": 0.0267, | |
| "step": 600 | |
| }, | |
| { | |
| "epoch": 5.0, | |
| "eval_loss": 0.4518275558948517, | |
| "eval_runtime": 439.9833, | |
| "eval_samples_per_second": 0.466, | |
| "eval_steps_per_second": 0.059, | |
| "eval_wer": 0.41509433962264153, | |
| "step": 645 | |
| }, | |
| { | |
| "epoch": 5.038910505836576, | |
| "grad_norm": 1.2988134622573853, | |
| "learning_rate": 2.3910008369689226e-05, | |
| "loss": 0.0315, | |
| "step": 650 | |
| }, | |
| { | |
| "epoch": 5.428015564202335, | |
| "grad_norm": 0.6612567901611328, | |
| "learning_rate": 2.2845687969613155e-05, | |
| "loss": 0.0168, | |
| "step": 700 | |
| }, | |
| { | |
| "epoch": 5.817120622568093, | |
| "grad_norm": 0.6520644426345825, | |
| "learning_rate": 2.1723911827982513e-05, | |
| "loss": 0.0177, | |
| "step": 750 | |
| }, | |
| { | |
| "epoch": 6.0, | |
| "eval_loss": 0.45992329716682434, | |
| "eval_runtime": 436.4339, | |
| "eval_samples_per_second": 0.47, | |
| "eval_steps_per_second": 0.06, | |
| "eval_wer": 0.3903301886792453, | |
| "step": 774 | |
| }, | |
| { | |
| "epoch": 6.202334630350195, | |
| "grad_norm": 1.1170883178710938, | |
| "learning_rate": 2.0552894964350105e-05, | |
| "loss": 0.0134, | |
| "step": 800 | |
| }, | |
| { | |
| "epoch": 6.591439688715953, | |
| "grad_norm": 1.3675938844680786, | |
| "learning_rate": 1.9341212999175674e-05, | |
| "loss": 0.009, | |
| "step": 850 | |
| }, | |
| { | |
| "epoch": 6.980544747081712, | |
| "grad_norm": 2.0311999320983887, | |
| "learning_rate": 1.8097739352624154e-05, | |
| "loss": 0.0087, | |
| "step": 900 | |
| }, | |
| { | |
| "epoch": 7.0, | |
| "eval_loss": 0.4803558588027954, | |
| "eval_runtime": 437.6131, | |
| "eval_samples_per_second": 0.468, | |
| "eval_steps_per_second": 0.059, | |
| "eval_wer": 0.41037735849056606, | |
| "step": 903 | |
| } | |
| ], | |
| "logging_steps": 50, | |
| "max_steps": 1935, | |
| "num_input_tokens_seen": 0, | |
| "num_train_epochs": 15, | |
| "save_steps": 500, | |
| "stateful_callbacks": { | |
| "TrainerControl": { | |
| "args": { | |
| "should_epoch_stop": false, | |
| "should_evaluate": false, | |
| "should_log": false, | |
| "should_save": true, | |
| "should_training_stop": false | |
| }, | |
| "attributes": {} | |
| } | |
| }, | |
| "total_flos": 4.1513009836032e+18, | |
| "train_batch_size": 8, | |
| "trial_name": null, | |
| "trial_params": null | |
| } | |