| { | |
| "best_metric": 0.5324379205703735, | |
| "best_model_checkpoint": "/scratch/skscla001/speech/results/mms-1b-lozgen-balanced-model/checkpoint-2300", | |
| "epoch": 21.13821138211382, | |
| "eval_steps": 100, | |
| "global_step": 2600, | |
| "is_hyper_param_search": false, | |
| "is_local_process_zero": true, | |
| "is_world_process_zero": true, | |
| "log_history": [ | |
| { | |
| "epoch": 0.8130081300813008, | |
| "grad_norm": 2.741861343383789, | |
| "learning_rate": 0.00028799999999999995, | |
| "loss": 6.6518, | |
| "step": 100 | |
| }, | |
| { | |
| "epoch": 0.8130081300813008, | |
| "eval_loss": 3.1670827865600586, | |
| "eval_runtime": 20.9246, | |
| "eval_samples_per_second": 13.142, | |
| "eval_steps_per_second": 3.298, | |
| "eval_wer": 0.9942789034564958, | |
| "step": 100 | |
| }, | |
| { | |
| "epoch": 1.6260162601626016, | |
| "grad_norm": 1.7816331386566162, | |
| "learning_rate": 0.0002919777158774373, | |
| "loss": 2.6718, | |
| "step": 200 | |
| }, | |
| { | |
| "epoch": 1.6260162601626016, | |
| "eval_loss": 2.2608625888824463, | |
| "eval_runtime": 20.9525, | |
| "eval_samples_per_second": 13.125, | |
| "eval_steps_per_second": 3.293, | |
| "eval_wer": 0.931585220500596, | |
| "step": 200 | |
| }, | |
| { | |
| "epoch": 2.4390243902439024, | |
| "grad_norm": 1.4348615407943726, | |
| "learning_rate": 0.0002836211699164345, | |
| "loss": 1.4567, | |
| "step": 300 | |
| }, | |
| { | |
| "epoch": 2.4390243902439024, | |
| "eval_loss": 0.740405261516571, | |
| "eval_runtime": 20.9256, | |
| "eval_samples_per_second": 13.142, | |
| "eval_steps_per_second": 3.297, | |
| "eval_wer": 0.7191895113230036, | |
| "step": 300 | |
| }, | |
| { | |
| "epoch": 3.252032520325203, | |
| "grad_norm": 2.2145164012908936, | |
| "learning_rate": 0.0002752646239554317, | |
| "loss": 0.7044, | |
| "step": 400 | |
| }, | |
| { | |
| "epoch": 3.252032520325203, | |
| "eval_loss": 0.6401560306549072, | |
| "eval_runtime": 20.8861, | |
| "eval_samples_per_second": 13.167, | |
| "eval_steps_per_second": 3.304, | |
| "eval_wer": 0.532061978545888, | |
| "step": 400 | |
| }, | |
| { | |
| "epoch": 4.065040650406504, | |
| "grad_norm": 1.0280542373657227, | |
| "learning_rate": 0.00026690807799442895, | |
| "loss": 0.6221, | |
| "step": 500 | |
| }, | |
| { | |
| "epoch": 4.065040650406504, | |
| "eval_loss": 0.6065016984939575, | |
| "eval_runtime": 20.9642, | |
| "eval_samples_per_second": 13.118, | |
| "eval_steps_per_second": 3.291, | |
| "eval_wer": 0.5165673420738975, | |
| "step": 500 | |
| }, | |
| { | |
| "epoch": 4.878048780487805, | |
| "grad_norm": 1.9754825830459595, | |
| "learning_rate": 0.00025855153203342614, | |
| "loss": 0.6016, | |
| "step": 600 | |
| }, | |
| { | |
| "epoch": 4.878048780487805, | |
| "eval_loss": 0.5947937965393066, | |
| "eval_runtime": 20.9641, | |
| "eval_samples_per_second": 13.118, | |
| "eval_steps_per_second": 3.291, | |
| "eval_wer": 0.4786650774731824, | |
| "step": 600 | |
| }, | |
| { | |
| "epoch": 5.691056910569106, | |
| "grad_norm": 14.851791381835938, | |
| "learning_rate": 0.0002502785515320334, | |
| "loss": 0.5686, | |
| "step": 700 | |
| }, | |
| { | |
| "epoch": 5.691056910569106, | |
| "eval_loss": 0.580583393573761, | |
| "eval_runtime": 20.7413, | |
| "eval_samples_per_second": 13.259, | |
| "eval_steps_per_second": 3.327, | |
| "eval_wer": 0.4641239570917759, | |
| "step": 700 | |
| }, | |
| { | |
| "epoch": 6.504065040650406, | |
| "grad_norm": 0.918157696723938, | |
| "learning_rate": 0.00024192200557103064, | |
| "loss": 0.6054, | |
| "step": 800 | |
| }, | |
| { | |
| "epoch": 6.504065040650406, | |
| "eval_loss": 0.5715970396995544, | |
| "eval_runtime": 20.8551, | |
| "eval_samples_per_second": 13.186, | |
| "eval_steps_per_second": 3.309, | |
| "eval_wer": 0.44862932061978544, | |
| "step": 800 | |
| }, | |
| { | |
| "epoch": 7.317073170731708, | |
| "grad_norm": 1.2295010089874268, | |
| "learning_rate": 0.00023356545961002783, | |
| "loss": 0.4871, | |
| "step": 900 | |
| }, | |
| { | |
| "epoch": 7.317073170731708, | |
| "eval_loss": 0.5732107162475586, | |
| "eval_runtime": 20.9032, | |
| "eval_samples_per_second": 13.156, | |
| "eval_steps_per_second": 3.301, | |
| "eval_wer": 0.4445768772348033, | |
| "step": 900 | |
| }, | |
| { | |
| "epoch": 8.130081300813009, | |
| "grad_norm": 1.2786221504211426, | |
| "learning_rate": 0.00022520891364902505, | |
| "loss": 0.5275, | |
| "step": 1000 | |
| }, | |
| { | |
| "epoch": 8.130081300813009, | |
| "eval_loss": 0.5667155385017395, | |
| "eval_runtime": 20.9254, | |
| "eval_samples_per_second": 13.142, | |
| "eval_steps_per_second": 3.297, | |
| "eval_wer": 0.43504171632896305, | |
| "step": 1000 | |
| }, | |
| { | |
| "epoch": 8.94308943089431, | |
| "grad_norm": 2.170154571533203, | |
| "learning_rate": 0.00021685236768802224, | |
| "loss": 0.5199, | |
| "step": 1100 | |
| }, | |
| { | |
| "epoch": 8.94308943089431, | |
| "eval_loss": 0.5688361525535583, | |
| "eval_runtime": 20.7636, | |
| "eval_samples_per_second": 13.244, | |
| "eval_steps_per_second": 3.323, | |
| "eval_wer": 0.4300357568533969, | |
| "step": 1100 | |
| }, | |
| { | |
| "epoch": 9.75609756097561, | |
| "grad_norm": 0.6021554470062256, | |
| "learning_rate": 0.0002084958217270195, | |
| "loss": 0.5031, | |
| "step": 1200 | |
| }, | |
| { | |
| "epoch": 9.75609756097561, | |
| "eval_loss": 0.5516279935836792, | |
| "eval_runtime": 20.7896, | |
| "eval_samples_per_second": 13.228, | |
| "eval_steps_per_second": 3.319, | |
| "eval_wer": 0.4443384982121573, | |
| "step": 1200 | |
| }, | |
| { | |
| "epoch": 10.56910569105691, | |
| "grad_norm": 0.8330698609352112, | |
| "learning_rate": 0.00020013927576601671, | |
| "loss": 0.4533, | |
| "step": 1300 | |
| }, | |
| { | |
| "epoch": 10.56910569105691, | |
| "eval_loss": 0.557715117931366, | |
| "eval_runtime": 21.0144, | |
| "eval_samples_per_second": 13.086, | |
| "eval_steps_per_second": 3.283, | |
| "eval_wer": 0.41787842669845054, | |
| "step": 1300 | |
| }, | |
| { | |
| "epoch": 11.382113821138212, | |
| "grad_norm": 1.5642321109771729, | |
| "learning_rate": 0.0001917827298050139, | |
| "loss": 0.4738, | |
| "step": 1400 | |
| }, | |
| { | |
| "epoch": 11.382113821138212, | |
| "eval_loss": 0.553620457649231, | |
| "eval_runtime": 20.9609, | |
| "eval_samples_per_second": 13.12, | |
| "eval_steps_per_second": 3.292, | |
| "eval_wer": 0.4057210965435042, | |
| "step": 1400 | |
| }, | |
| { | |
| "epoch": 12.195121951219512, | |
| "grad_norm": 6.336544513702393, | |
| "learning_rate": 0.00018342618384401113, | |
| "loss": 0.4925, | |
| "step": 1500 | |
| }, | |
| { | |
| "epoch": 12.195121951219512, | |
| "eval_loss": 0.5502671003341675, | |
| "eval_runtime": 20.8854, | |
| "eval_samples_per_second": 13.167, | |
| "eval_steps_per_second": 3.304, | |
| "eval_wer": 0.39690107270560193, | |
| "step": 1500 | |
| }, | |
| { | |
| "epoch": 13.008130081300813, | |
| "grad_norm": 1.8177986145019531, | |
| "learning_rate": 0.00017506963788300832, | |
| "loss": 0.441, | |
| "step": 1600 | |
| }, | |
| { | |
| "epoch": 13.008130081300813, | |
| "eval_loss": 0.5402783155441284, | |
| "eval_runtime": 20.8181, | |
| "eval_samples_per_second": 13.21, | |
| "eval_steps_per_second": 3.314, | |
| "eval_wer": 0.40047675804529204, | |
| "step": 1600 | |
| }, | |
| { | |
| "epoch": 13.821138211382113, | |
| "grad_norm": 0.8598074316978455, | |
| "learning_rate": 0.00016671309192200557, | |
| "loss": 0.4177, | |
| "step": 1700 | |
| }, | |
| { | |
| "epoch": 13.821138211382113, | |
| "eval_loss": 0.5563377737998962, | |
| "eval_runtime": 20.9575, | |
| "eval_samples_per_second": 13.122, | |
| "eval_steps_per_second": 3.292, | |
| "eval_wer": 0.3914183551847437, | |
| "step": 1700 | |
| }, | |
| { | |
| "epoch": 14.634146341463415, | |
| "grad_norm": 2.3619320392608643, | |
| "learning_rate": 0.00015835654596100277, | |
| "loss": 0.4589, | |
| "step": 1800 | |
| }, | |
| { | |
| "epoch": 14.634146341463415, | |
| "eval_loss": 0.5394493341445923, | |
| "eval_runtime": 21.0011, | |
| "eval_samples_per_second": 13.095, | |
| "eval_steps_per_second": 3.286, | |
| "eval_wer": 0.3876042908224076, | |
| "step": 1800 | |
| }, | |
| { | |
| "epoch": 15.447154471544716, | |
| "grad_norm": 3.2565107345581055, | |
| "learning_rate": 0.00015, | |
| "loss": 0.4131, | |
| "step": 1900 | |
| }, | |
| { | |
| "epoch": 15.447154471544716, | |
| "eval_loss": 0.5424718856811523, | |
| "eval_runtime": 20.8003, | |
| "eval_samples_per_second": 13.221, | |
| "eval_steps_per_second": 3.317, | |
| "eval_wer": 0.3957091775923719, | |
| "step": 1900 | |
| }, | |
| { | |
| "epoch": 16.260162601626018, | |
| "grad_norm": 1.814285397529602, | |
| "learning_rate": 0.0001416434540389972, | |
| "loss": 0.393, | |
| "step": 2000 | |
| }, | |
| { | |
| "epoch": 16.260162601626018, | |
| "eval_loss": 0.546908974647522, | |
| "eval_runtime": 20.881, | |
| "eval_samples_per_second": 13.17, | |
| "eval_steps_per_second": 3.304, | |
| "eval_wer": 0.3907032181168057, | |
| "step": 2000 | |
| }, | |
| { | |
| "epoch": 17.073170731707318, | |
| "grad_norm": 0.762374997138977, | |
| "learning_rate": 0.00013328690807799443, | |
| "loss": 0.4235, | |
| "step": 2100 | |
| }, | |
| { | |
| "epoch": 17.073170731707318, | |
| "eval_loss": 0.5356802940368652, | |
| "eval_runtime": 20.9384, | |
| "eval_samples_per_second": 13.134, | |
| "eval_steps_per_second": 3.295, | |
| "eval_wer": 0.3878426698450536, | |
| "step": 2100 | |
| }, | |
| { | |
| "epoch": 17.88617886178862, | |
| "grad_norm": 0.8089348077774048, | |
| "learning_rate": 0.00012493036211699162, | |
| "loss": 0.4113, | |
| "step": 2200 | |
| }, | |
| { | |
| "epoch": 17.88617886178862, | |
| "eval_loss": 0.5390728712081909, | |
| "eval_runtime": 20.9645, | |
| "eval_samples_per_second": 13.117, | |
| "eval_steps_per_second": 3.291, | |
| "eval_wer": 0.3802145411203814, | |
| "step": 2200 | |
| }, | |
| { | |
| "epoch": 18.69918699186992, | |
| "grad_norm": 1.4261620044708252, | |
| "learning_rate": 0.00011657381615598886, | |
| "loss": 0.3781, | |
| "step": 2300 | |
| }, | |
| { | |
| "epoch": 18.69918699186992, | |
| "eval_loss": 0.5324379205703735, | |
| "eval_runtime": 20.8837, | |
| "eval_samples_per_second": 13.168, | |
| "eval_steps_per_second": 3.304, | |
| "eval_wer": 0.3728247914183552, | |
| "step": 2300 | |
| }, | |
| { | |
| "epoch": 19.51219512195122, | |
| "grad_norm": 0.8106199502944946, | |
| "learning_rate": 0.00010821727019498607, | |
| "loss": 0.3706, | |
| "step": 2400 | |
| }, | |
| { | |
| "epoch": 19.51219512195122, | |
| "eval_loss": 0.5462723970413208, | |
| "eval_runtime": 20.8289, | |
| "eval_samples_per_second": 13.203, | |
| "eval_steps_per_second": 3.313, | |
| "eval_wer": 0.38259833134684146, | |
| "step": 2400 | |
| }, | |
| { | |
| "epoch": 20.32520325203252, | |
| "grad_norm": 1.8274589776992798, | |
| "learning_rate": 9.986072423398329e-05, | |
| "loss": 0.3617, | |
| "step": 2500 | |
| }, | |
| { | |
| "epoch": 20.32520325203252, | |
| "eval_loss": 0.5391108989715576, | |
| "eval_runtime": 21.0365, | |
| "eval_samples_per_second": 13.072, | |
| "eval_steps_per_second": 3.28, | |
| "eval_wer": 0.3697258641239571, | |
| "step": 2500 | |
| }, | |
| { | |
| "epoch": 21.13821138211382, | |
| "grad_norm": 2.291131019592285, | |
| "learning_rate": 9.15041782729805e-05, | |
| "loss": 0.401, | |
| "step": 2600 | |
| }, | |
| { | |
| "epoch": 21.13821138211382, | |
| "eval_loss": 0.5417346358299255, | |
| "eval_runtime": 21.0259, | |
| "eval_samples_per_second": 13.079, | |
| "eval_steps_per_second": 3.282, | |
| "eval_wer": 0.366388557806913, | |
| "step": 2600 | |
| }, | |
| { | |
| "epoch": 21.13821138211382, | |
| "step": 2600, | |
| "total_flos": 7.953394513554737e+18, | |
| "train_loss": 0.8388280222966121, | |
| "train_runtime": 2359.9883, | |
| "train_samples_per_second": 6.242, | |
| "train_steps_per_second": 1.564 | |
| } | |
| ], | |
| "logging_steps": 100, | |
| "max_steps": 3690, | |
| "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": 1 | |
| } | |
| }, | |
| "TrainerControl": { | |
| "args": { | |
| "should_epoch_stop": false, | |
| "should_evaluate": false, | |
| "should_log": false, | |
| "should_save": true, | |
| "should_training_stop": false | |
| }, | |
| "attributes": {} | |
| } | |
| }, | |
| "total_flos": 7.953394513554737e+18, | |
| "train_batch_size": 4, | |
| "trial_name": null, | |
| "trial_params": null | |
| } | |