| { | |
| "best_metric": 0.2750256657600403, | |
| "best_model_checkpoint": "/scratch/skscla001/speech/results/mms-1b-bemgen-balanced-model/checkpoint-1700", | |
| "epoch": 2.0618556701030926, | |
| "eval_steps": 100, | |
| "global_step": 2000, | |
| "is_hyper_param_search": false, | |
| "is_local_process_zero": true, | |
| "is_world_process_zero": true, | |
| "log_history": [ | |
| { | |
| "epoch": 0.10309278350515463, | |
| "grad_norm": 2.7115845680236816, | |
| "learning_rate": 0.00028799999999999995, | |
| "loss": 6.8327, | |
| "step": 100 | |
| }, | |
| { | |
| "epoch": 0.10309278350515463, | |
| "eval_loss": 0.8479615449905396, | |
| "eval_runtime": 30.8075, | |
| "eval_samples_per_second": 16.003, | |
| "eval_steps_per_second": 4.025, | |
| "eval_wer": 0.788869640537029, | |
| "step": 100 | |
| }, | |
| { | |
| "epoch": 0.20618556701030927, | |
| "grad_norm": 2.6976242065429688, | |
| "learning_rate": 0.0002990068965517241, | |
| "loss": 0.5906, | |
| "step": 200 | |
| }, | |
| { | |
| "epoch": 0.20618556701030927, | |
| "eval_loss": 0.3704410493373871, | |
| "eval_runtime": 30.4245, | |
| "eval_samples_per_second": 16.204, | |
| "eval_steps_per_second": 4.076, | |
| "eval_wer": 0.5818536162841057, | |
| "step": 200 | |
| }, | |
| { | |
| "epoch": 0.30927835051546393, | |
| "grad_norm": 3.159865379333496, | |
| "learning_rate": 0.00029797241379310343, | |
| "loss": 0.4809, | |
| "step": 300 | |
| }, | |
| { | |
| "epoch": 0.30927835051546393, | |
| "eval_loss": 0.33272889256477356, | |
| "eval_runtime": 30.3459, | |
| "eval_samples_per_second": 16.246, | |
| "eval_steps_per_second": 4.086, | |
| "eval_wer": 0.503897791251624, | |
| "step": 300 | |
| }, | |
| { | |
| "epoch": 0.41237113402061853, | |
| "grad_norm": 2.7093732357025146, | |
| "learning_rate": 0.00029693793103448274, | |
| "loss": 0.4495, | |
| "step": 400 | |
| }, | |
| { | |
| "epoch": 0.41237113402061853, | |
| "eval_loss": 0.3171828091144562, | |
| "eval_runtime": 30.503, | |
| "eval_samples_per_second": 16.162, | |
| "eval_steps_per_second": 4.065, | |
| "eval_wer": 0.48938934603724554, | |
| "step": 400 | |
| }, | |
| { | |
| "epoch": 0.5154639175257731, | |
| "grad_norm": 2.6349027156829834, | |
| "learning_rate": 0.00029590344827586204, | |
| "loss": 0.4266, | |
| "step": 500 | |
| }, | |
| { | |
| "epoch": 0.5154639175257731, | |
| "eval_loss": 0.3101615607738495, | |
| "eval_runtime": 30.7842, | |
| "eval_samples_per_second": 16.015, | |
| "eval_steps_per_second": 4.028, | |
| "eval_wer": 0.4631875270679948, | |
| "step": 500 | |
| }, | |
| { | |
| "epoch": 0.6185567010309279, | |
| "grad_norm": 19.215137481689453, | |
| "learning_rate": 0.00029486896551724135, | |
| "loss": 0.4167, | |
| "step": 600 | |
| }, | |
| { | |
| "epoch": 0.6185567010309279, | |
| "eval_loss": 0.30752459168434143, | |
| "eval_runtime": 30.6141, | |
| "eval_samples_per_second": 16.104, | |
| "eval_steps_per_second": 4.05, | |
| "eval_wer": 0.4716327414465136, | |
| "step": 600 | |
| }, | |
| { | |
| "epoch": 0.7216494845360825, | |
| "grad_norm": 1.4912749528884888, | |
| "learning_rate": 0.00029383448275862066, | |
| "loss": 0.4151, | |
| "step": 700 | |
| }, | |
| { | |
| "epoch": 0.7216494845360825, | |
| "eval_loss": 0.2996384799480438, | |
| "eval_runtime": 30.5121, | |
| "eval_samples_per_second": 16.158, | |
| "eval_steps_per_second": 4.064, | |
| "eval_wer": 0.48289302728453876, | |
| "step": 700 | |
| }, | |
| { | |
| "epoch": 0.8247422680412371, | |
| "grad_norm": 2.437255859375, | |
| "learning_rate": 0.00029279999999999996, | |
| "loss": 0.3955, | |
| "step": 800 | |
| }, | |
| { | |
| "epoch": 0.8247422680412371, | |
| "eval_loss": 0.29851067066192627, | |
| "eval_runtime": 30.5988, | |
| "eval_samples_per_second": 16.112, | |
| "eval_steps_per_second": 4.052, | |
| "eval_wer": 0.47119965352966653, | |
| "step": 800 | |
| }, | |
| { | |
| "epoch": 0.9278350515463918, | |
| "grad_norm": 1.967859148979187, | |
| "learning_rate": 0.00029176551724137927, | |
| "loss": 0.3802, | |
| "step": 900 | |
| }, | |
| { | |
| "epoch": 0.9278350515463918, | |
| "eval_loss": 0.2959710359573364, | |
| "eval_runtime": 30.8901, | |
| "eval_samples_per_second": 15.96, | |
| "eval_steps_per_second": 4.014, | |
| "eval_wer": 0.49263750541359896, | |
| "step": 900 | |
| }, | |
| { | |
| "epoch": 1.0309278350515463, | |
| "grad_norm": 0.8750647902488708, | |
| "learning_rate": 0.00029073103448275857, | |
| "loss": 0.392, | |
| "step": 1000 | |
| }, | |
| { | |
| "epoch": 1.0309278350515463, | |
| "eval_loss": 0.28385233879089355, | |
| "eval_runtime": 30.857, | |
| "eval_samples_per_second": 15.977, | |
| "eval_steps_per_second": 4.019, | |
| "eval_wer": 0.43741879601559114, | |
| "step": 1000 | |
| }, | |
| { | |
| "epoch": 1.134020618556701, | |
| "grad_norm": 0.8763826489448547, | |
| "learning_rate": 0.0002896965517241379, | |
| "loss": 0.375, | |
| "step": 1100 | |
| }, | |
| { | |
| "epoch": 1.134020618556701, | |
| "eval_loss": 0.28370916843414307, | |
| "eval_runtime": 30.6261, | |
| "eval_samples_per_second": 16.097, | |
| "eval_steps_per_second": 4.049, | |
| "eval_wer": 0.4317886530965786, | |
| "step": 1100 | |
| }, | |
| { | |
| "epoch": 1.2371134020618557, | |
| "grad_norm": 0.6871535181999207, | |
| "learning_rate": 0.0002886620689655172, | |
| "loss": 0.3885, | |
| "step": 1200 | |
| }, | |
| { | |
| "epoch": 1.2371134020618557, | |
| "eval_loss": 0.281158447265625, | |
| "eval_runtime": 30.5181, | |
| "eval_samples_per_second": 16.154, | |
| "eval_steps_per_second": 4.063, | |
| "eval_wer": 0.4257254222607189, | |
| "step": 1200 | |
| }, | |
| { | |
| "epoch": 1.3402061855670104, | |
| "grad_norm": 1.2791551351547241, | |
| "learning_rate": 0.00028762758620689654, | |
| "loss": 0.3824, | |
| "step": 1300 | |
| }, | |
| { | |
| "epoch": 1.3402061855670104, | |
| "eval_loss": 0.282485693693161, | |
| "eval_runtime": 30.9775, | |
| "eval_samples_per_second": 15.915, | |
| "eval_steps_per_second": 4.003, | |
| "eval_wer": 0.4255088783022954, | |
| "step": 1300 | |
| }, | |
| { | |
| "epoch": 1.443298969072165, | |
| "grad_norm": 2.2985143661499023, | |
| "learning_rate": 0.00028659310344827585, | |
| "loss": 0.3906, | |
| "step": 1400 | |
| }, | |
| { | |
| "epoch": 1.443298969072165, | |
| "eval_loss": 0.2794145941734314, | |
| "eval_runtime": 30.8505, | |
| "eval_samples_per_second": 15.98, | |
| "eval_steps_per_second": 4.019, | |
| "eval_wer": 0.42897358163707233, | |
| "step": 1400 | |
| }, | |
| { | |
| "epoch": 1.5463917525773194, | |
| "grad_norm": 0.9774999618530273, | |
| "learning_rate": 0.00028555862068965516, | |
| "loss": 0.3465, | |
| "step": 1500 | |
| }, | |
| { | |
| "epoch": 1.5463917525773194, | |
| "eval_loss": 0.28065648674964905, | |
| "eval_runtime": 30.5852, | |
| "eval_samples_per_second": 16.119, | |
| "eval_steps_per_second": 4.054, | |
| "eval_wer": 0.42832394976180166, | |
| "step": 1500 | |
| }, | |
| { | |
| "epoch": 1.6494845360824741, | |
| "grad_norm": 0.6370195746421814, | |
| "learning_rate": 0.00028452413793103446, | |
| "loss": 0.3564, | |
| "step": 1600 | |
| }, | |
| { | |
| "epoch": 1.6494845360824741, | |
| "eval_loss": 0.2772924304008484, | |
| "eval_runtime": 30.7481, | |
| "eval_samples_per_second": 16.034, | |
| "eval_steps_per_second": 4.033, | |
| "eval_wer": 0.4237765266349069, | |
| "step": 1600 | |
| }, | |
| { | |
| "epoch": 1.7525773195876289, | |
| "grad_norm": 0.919018566608429, | |
| "learning_rate": 0.00028348965517241377, | |
| "loss": 0.3617, | |
| "step": 1700 | |
| }, | |
| { | |
| "epoch": 1.7525773195876289, | |
| "eval_loss": 0.2750256657600403, | |
| "eval_runtime": 31.0143, | |
| "eval_samples_per_second": 15.896, | |
| "eval_steps_per_second": 3.998, | |
| "eval_wer": 0.4452143785188393, | |
| "step": 1700 | |
| }, | |
| { | |
| "epoch": 1.8556701030927836, | |
| "grad_norm": 1.0795314311981201, | |
| "learning_rate": 0.0002824551724137931, | |
| "loss": 0.3808, | |
| "step": 1800 | |
| }, | |
| { | |
| "epoch": 1.8556701030927836, | |
| "eval_loss": 0.27826789021492004, | |
| "eval_runtime": 30.9277, | |
| "eval_samples_per_second": 15.94, | |
| "eval_steps_per_second": 4.009, | |
| "eval_wer": 0.42291035080121264, | |
| "step": 1800 | |
| }, | |
| { | |
| "epoch": 1.9587628865979383, | |
| "grad_norm": 1.087866187095642, | |
| "learning_rate": 0.0002814206896551724, | |
| "loss": 0.3661, | |
| "step": 1900 | |
| }, | |
| { | |
| "epoch": 1.9587628865979383, | |
| "eval_loss": 0.2761477828025818, | |
| "eval_runtime": 30.7566, | |
| "eval_samples_per_second": 16.029, | |
| "eval_steps_per_second": 4.032, | |
| "eval_wer": 0.4517106972715461, | |
| "step": 1900 | |
| }, | |
| { | |
| "epoch": 2.0618556701030926, | |
| "grad_norm": 0.7742441892623901, | |
| "learning_rate": 0.0002803862068965517, | |
| "loss": 0.3952, | |
| "step": 2000 | |
| }, | |
| { | |
| "epoch": 2.0618556701030926, | |
| "eval_loss": 0.27530166506767273, | |
| "eval_runtime": 30.8874, | |
| "eval_samples_per_second": 15.961, | |
| "eval_steps_per_second": 4.015, | |
| "eval_wer": 0.4200952793417064, | |
| "step": 2000 | |
| }, | |
| { | |
| "epoch": 2.0618556701030926, | |
| "step": 2000, | |
| "total_flos": 4.819386981327346e+18, | |
| "train_loss": 0.7261434955596924, | |
| "train_runtime": 1981.0289, | |
| "train_samples_per_second": 58.727, | |
| "train_steps_per_second": 14.689 | |
| } | |
| ], | |
| "logging_steps": 100, | |
| "max_steps": 29100, | |
| "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": 4.819386981327346e+18, | |
| "train_batch_size": 4, | |
| "trial_name": null, | |
| "trial_params": null | |
| } | |