| { | |
| "best_metric": 0.17163243889808655, | |
| "best_model_checkpoint": "/scratch/skscla001/speech/results/mms-1b-nyagen-balanced-model/checkpoint-900", | |
| "epoch": 3.314917127071823, | |
| "eval_steps": 100, | |
| "global_step": 1200, | |
| "is_hyper_param_search": false, | |
| "is_local_process_zero": true, | |
| "is_world_process_zero": true, | |
| "log_history": [ | |
| { | |
| "epoch": 0.27624309392265195, | |
| "grad_norm": 4.279232025146484, | |
| "learning_rate": 0.00028799999999999995, | |
| "loss": 7.181, | |
| "step": 100 | |
| }, | |
| { | |
| "epoch": 0.27624309392265195, | |
| "eval_loss": 0.6055206656455994, | |
| "eval_runtime": 16.1135, | |
| "eval_samples_per_second": 10.488, | |
| "eval_steps_per_second": 2.669, | |
| "eval_wer": 0.5246717971933001, | |
| "step": 100 | |
| }, | |
| { | |
| "epoch": 0.5524861878453039, | |
| "grad_norm": 2.128457546234131, | |
| "learning_rate": 0.00029735130111524163, | |
| "loss": 0.5071, | |
| "step": 200 | |
| }, | |
| { | |
| "epoch": 0.5524861878453039, | |
| "eval_loss": 0.24523495137691498, | |
| "eval_runtime": 15.4101, | |
| "eval_samples_per_second": 10.967, | |
| "eval_steps_per_second": 2.79, | |
| "eval_wer": 0.3594386600271616, | |
| "step": 200 | |
| }, | |
| { | |
| "epoch": 0.8287292817679558, | |
| "grad_norm": 1.3095550537109375, | |
| "learning_rate": 0.0002945631970260223, | |
| "loss": 0.3791, | |
| "step": 300 | |
| }, | |
| { | |
| "epoch": 0.8287292817679558, | |
| "eval_loss": 0.21587276458740234, | |
| "eval_runtime": 15.3962, | |
| "eval_samples_per_second": 10.977, | |
| "eval_steps_per_second": 2.793, | |
| "eval_wer": 0.3232231779085559, | |
| "step": 300 | |
| }, | |
| { | |
| "epoch": 1.1049723756906078, | |
| "grad_norm": 4.117229461669922, | |
| "learning_rate": 0.0002917750929368029, | |
| "loss": 0.3464, | |
| "step": 400 | |
| }, | |
| { | |
| "epoch": 1.1049723756906078, | |
| "eval_loss": 0.20587413012981415, | |
| "eval_runtime": 15.3374, | |
| "eval_samples_per_second": 11.019, | |
| "eval_steps_per_second": 2.804, | |
| "eval_wer": 0.3046627433227705, | |
| "step": 400 | |
| }, | |
| { | |
| "epoch": 1.3812154696132597, | |
| "grad_norm": 1.0849329233169556, | |
| "learning_rate": 0.0002889869888475836, | |
| "loss": 0.3326, | |
| "step": 500 | |
| }, | |
| { | |
| "epoch": 1.3812154696132597, | |
| "eval_loss": 0.19188211858272552, | |
| "eval_runtime": 15.5114, | |
| "eval_samples_per_second": 10.895, | |
| "eval_steps_per_second": 2.772, | |
| "eval_wer": 0.29425079221367134, | |
| "step": 500 | |
| }, | |
| { | |
| "epoch": 1.6574585635359116, | |
| "grad_norm": 1.2119916677474976, | |
| "learning_rate": 0.00028619888475836427, | |
| "loss": 0.322, | |
| "step": 600 | |
| }, | |
| { | |
| "epoch": 1.6574585635359116, | |
| "eval_loss": 0.18680231273174286, | |
| "eval_runtime": 15.4576, | |
| "eval_samples_per_second": 10.933, | |
| "eval_steps_per_second": 2.782, | |
| "eval_wer": 0.28610230873698506, | |
| "step": 600 | |
| }, | |
| { | |
| "epoch": 1.9337016574585635, | |
| "grad_norm": 1.408340573310852, | |
| "learning_rate": 0.00028341078066914494, | |
| "loss": 0.3025, | |
| "step": 700 | |
| }, | |
| { | |
| "epoch": 1.9337016574585635, | |
| "eval_loss": 0.1849866360425949, | |
| "eval_runtime": 15.3489, | |
| "eval_samples_per_second": 11.011, | |
| "eval_steps_per_second": 2.801, | |
| "eval_wer": 0.2901765504753282, | |
| "step": 700 | |
| }, | |
| { | |
| "epoch": 2.2099447513812156, | |
| "grad_norm": 0.6426145434379578, | |
| "learning_rate": 0.0002806226765799256, | |
| "loss": 0.2939, | |
| "step": 800 | |
| }, | |
| { | |
| "epoch": 2.2099447513812156, | |
| "eval_loss": 0.17766940593719482, | |
| "eval_runtime": 15.4593, | |
| "eval_samples_per_second": 10.932, | |
| "eval_steps_per_second": 2.781, | |
| "eval_wer": 0.2698053417836125, | |
| "step": 800 | |
| }, | |
| { | |
| "epoch": 2.4861878453038675, | |
| "grad_norm": 0.5668926239013672, | |
| "learning_rate": 0.0002778345724907063, | |
| "loss": 0.2971, | |
| "step": 900 | |
| }, | |
| { | |
| "epoch": 2.4861878453038675, | |
| "eval_loss": 0.17163243889808655, | |
| "eval_runtime": 15.5806, | |
| "eval_samples_per_second": 10.847, | |
| "eval_steps_per_second": 2.76, | |
| "eval_wer": 0.2675418741511996, | |
| "step": 900 | |
| }, | |
| { | |
| "epoch": 2.7624309392265194, | |
| "grad_norm": 0.42915207147598267, | |
| "learning_rate": 0.00027504646840148696, | |
| "loss": 0.2787, | |
| "step": 1000 | |
| }, | |
| { | |
| "epoch": 2.7624309392265194, | |
| "eval_loss": 0.17503149807453156, | |
| "eval_runtime": 15.5493, | |
| "eval_samples_per_second": 10.869, | |
| "eval_steps_per_second": 2.765, | |
| "eval_wer": 0.27161611588954276, | |
| "step": 1000 | |
| }, | |
| { | |
| "epoch": 3.0386740331491713, | |
| "grad_norm": 0.994978129863739, | |
| "learning_rate": 0.00027225836431226763, | |
| "loss": 0.32, | |
| "step": 1100 | |
| }, | |
| { | |
| "epoch": 3.0386740331491713, | |
| "eval_loss": 0.1725001335144043, | |
| "eval_runtime": 15.57, | |
| "eval_samples_per_second": 10.854, | |
| "eval_steps_per_second": 2.762, | |
| "eval_wer": 0.27342688999547304, | |
| "step": 1100 | |
| }, | |
| { | |
| "epoch": 3.314917127071823, | |
| "grad_norm": 0.6422222852706909, | |
| "learning_rate": 0.0002694702602230483, | |
| "loss": 0.2738, | |
| "step": 1200 | |
| }, | |
| { | |
| "epoch": 3.314917127071823, | |
| "eval_loss": 0.18027880787849426, | |
| "eval_runtime": 15.2933, | |
| "eval_samples_per_second": 11.051, | |
| "eval_steps_per_second": 2.812, | |
| "eval_wer": 0.2544137618832051, | |
| "step": 1200 | |
| }, | |
| { | |
| "epoch": 3.314917127071823, | |
| "step": 1200, | |
| "total_flos": 4.522297158190472e+18, | |
| "train_loss": 0.9028558111190796, | |
| "train_runtime": 1060.5768, | |
| "train_samples_per_second": 40.874, | |
| "train_steps_per_second": 10.24 | |
| } | |
| ], | |
| "logging_steps": 100, | |
| "max_steps": 10860, | |
| "num_input_tokens_seen": 0, | |
| "num_train_epochs": 30, | |
| "save_steps": 400, | |
| "stateful_callbacks": { | |
| "EarlyStoppingCallback": { | |
| "args": { | |
| "early_stopping_patience": 3, | |
| "early_stopping_threshold": 0.0 | |
| }, | |
| "attributes": { | |
| "early_stopping_patience_counter": 3 | |
| } | |
| }, | |
| "TrainerControl": { | |
| "args": { | |
| "should_epoch_stop": false, | |
| "should_evaluate": false, | |
| "should_log": false, | |
| "should_save": true, | |
| "should_training_stop": true | |
| }, | |
| "attributes": {} | |
| } | |
| }, | |
| "total_flos": 4.522297158190472e+18, | |
| "train_batch_size": 4, | |
| "trial_name": null, | |
| "trial_params": null | |
| } | |