| { | |
| "best_metric": 0.377719340186469, | |
| "best_model_checkpoint": "w2v-bert-2.0-CV_Fleurs-lg-10hrs-v4/checkpoint-11286", | |
| "epoch": 25.0, | |
| "eval_steps": 500, | |
| "global_step": 12825, | |
| "is_hyper_param_search": false, | |
| "is_local_process_zero": true, | |
| "is_world_process_zero": true, | |
| "log_history": [ | |
| { | |
| "epoch": 1.0, | |
| "grad_norm": 30.792112350463867, | |
| "learning_rate": 9.998031496062993e-05, | |
| "loss": 1.4106, | |
| "step": 513 | |
| }, | |
| { | |
| "epoch": 1.0, | |
| "eval_cer": 0.1264185884550246, | |
| "eval_loss": 0.5293365716934204, | |
| "eval_runtime": 31.0205, | |
| "eval_samples_per_second": 26.628, | |
| "eval_steps_per_second": 6.673, | |
| "eval_wer": 0.5665790102797036, | |
| "step": 513 | |
| }, | |
| { | |
| "epoch": 2.0, | |
| "grad_norm": 10.16055679321289, | |
| "learning_rate": 9.897047244094488e-05, | |
| "loss": 0.4775, | |
| "step": 1026 | |
| }, | |
| { | |
| "epoch": 2.0, | |
| "eval_cer": 0.11182513176846737, | |
| "eval_loss": 0.4882170259952545, | |
| "eval_runtime": 30.4398, | |
| "eval_samples_per_second": 27.135, | |
| "eval_steps_per_second": 6.8, | |
| "eval_wer": 0.5251016017212526, | |
| "step": 1026 | |
| }, | |
| { | |
| "epoch": 3.0, | |
| "grad_norm": 24.60751724243164, | |
| "learning_rate": 9.796062992125985e-05, | |
| "loss": 0.3696, | |
| "step": 1539 | |
| }, | |
| { | |
| "epoch": 3.0, | |
| "eval_cer": 0.09981164627486981, | |
| "eval_loss": 0.397329181432724, | |
| "eval_runtime": 30.5019, | |
| "eval_samples_per_second": 27.08, | |
| "eval_steps_per_second": 6.786, | |
| "eval_wer": 0.4564905570164953, | |
| "step": 1539 | |
| }, | |
| { | |
| "epoch": 4.0, | |
| "grad_norm": 8.973501205444336, | |
| "learning_rate": 9.695078740157481e-05, | |
| "loss": 0.3007, | |
| "step": 2052 | |
| }, | |
| { | |
| "epoch": 4.0, | |
| "eval_cer": 0.09322717991737761, | |
| "eval_loss": 0.3895888030529022, | |
| "eval_runtime": 29.8209, | |
| "eval_samples_per_second": 27.699, | |
| "eval_steps_per_second": 6.941, | |
| "eval_wer": 0.4311498924217069, | |
| "step": 2052 | |
| }, | |
| { | |
| "epoch": 5.0, | |
| "grad_norm": 3.461097240447998, | |
| "learning_rate": 9.594094488188978e-05, | |
| "loss": 0.2564, | |
| "step": 2565 | |
| }, | |
| { | |
| "epoch": 5.0, | |
| "eval_cer": 0.09659855331676664, | |
| "eval_loss": 0.3743041455745697, | |
| "eval_runtime": 29.9644, | |
| "eval_samples_per_second": 27.566, | |
| "eval_steps_per_second": 6.908, | |
| "eval_wer": 0.4372459956968683, | |
| "step": 2565 | |
| }, | |
| { | |
| "epoch": 6.0, | |
| "grad_norm": 9.60593032836914, | |
| "learning_rate": 9.493110236220472e-05, | |
| "loss": 0.2273, | |
| "step": 3078 | |
| }, | |
| { | |
| "epoch": 6.0, | |
| "eval_cer": 0.0879880973108153, | |
| "eval_loss": 0.3549309968948364, | |
| "eval_runtime": 30.2119, | |
| "eval_samples_per_second": 27.34, | |
| "eval_steps_per_second": 6.852, | |
| "eval_wer": 0.41836002868754485, | |
| "step": 3078 | |
| }, | |
| { | |
| "epoch": 7.0, | |
| "grad_norm": 2.7023348808288574, | |
| "learning_rate": 9.392125984251969e-05, | |
| "loss": 0.203, | |
| "step": 3591 | |
| }, | |
| { | |
| "epoch": 7.0, | |
| "eval_cer": 0.08468003608794061, | |
| "eval_loss": 0.3835783302783966, | |
| "eval_runtime": 29.6315, | |
| "eval_samples_per_second": 27.876, | |
| "eval_steps_per_second": 6.986, | |
| "eval_wer": 0.4000717188620607, | |
| "step": 3591 | |
| }, | |
| { | |
| "epoch": 8.0, | |
| "grad_norm": 2.196669340133667, | |
| "learning_rate": 9.291141732283465e-05, | |
| "loss": 0.1775, | |
| "step": 4104 | |
| }, | |
| { | |
| "epoch": 8.0, | |
| "eval_cer": 0.08550309438262714, | |
| "eval_loss": 0.3693934977054596, | |
| "eval_runtime": 30.2095, | |
| "eval_samples_per_second": 27.342, | |
| "eval_steps_per_second": 6.852, | |
| "eval_wer": 0.4132201769065264, | |
| "step": 4104 | |
| }, | |
| { | |
| "epoch": 9.0, | |
| "grad_norm": 3.045036792755127, | |
| "learning_rate": 9.190354330708662e-05, | |
| "loss": 0.159, | |
| "step": 4617 | |
| }, | |
| { | |
| "epoch": 9.0, | |
| "eval_cer": 0.0889694360467877, | |
| "eval_loss": 0.3487217128276825, | |
| "eval_runtime": 29.9136, | |
| "eval_samples_per_second": 27.613, | |
| "eval_steps_per_second": 6.92, | |
| "eval_wer": 0.4136983026535979, | |
| "step": 4617 | |
| }, | |
| { | |
| "epoch": 10.0, | |
| "grad_norm": 3.3439998626708984, | |
| "learning_rate": 9.089370078740158e-05, | |
| "loss": 0.1444, | |
| "step": 5130 | |
| }, | |
| { | |
| "epoch": 10.0, | |
| "eval_cer": 0.08915937257633075, | |
| "eval_loss": 0.3705226480960846, | |
| "eval_runtime": 30.2455, | |
| "eval_samples_per_second": 27.31, | |
| "eval_steps_per_second": 6.844, | |
| "eval_wer": 0.42768348075543866, | |
| "step": 5130 | |
| }, | |
| { | |
| "epoch": 11.0, | |
| "grad_norm": 3.339082956314087, | |
| "learning_rate": 8.988385826771653e-05, | |
| "loss": 0.1262, | |
| "step": 5643 | |
| }, | |
| { | |
| "epoch": 11.0, | |
| "eval_cer": 0.08284398296902452, | |
| "eval_loss": 0.3619905114173889, | |
| "eval_runtime": 30.7416, | |
| "eval_samples_per_second": 26.869, | |
| "eval_steps_per_second": 6.734, | |
| "eval_wer": 0.3981592158737748, | |
| "step": 5643 | |
| }, | |
| { | |
| "epoch": 12.0, | |
| "grad_norm": 0.1660577654838562, | |
| "learning_rate": 8.88740157480315e-05, | |
| "loss": 0.1139, | |
| "step": 6156 | |
| }, | |
| { | |
| "epoch": 12.0, | |
| "eval_cer": 0.08422102280821159, | |
| "eval_loss": 0.3665634095668793, | |
| "eval_runtime": 30.3996, | |
| "eval_samples_per_second": 27.171, | |
| "eval_steps_per_second": 6.809, | |
| "eval_wer": 0.40545063351661487, | |
| "step": 6156 | |
| }, | |
| { | |
| "epoch": 13.0, | |
| "grad_norm": 0.19767990708351135, | |
| "learning_rate": 8.786417322834646e-05, | |
| "loss": 0.1015, | |
| "step": 6669 | |
| }, | |
| { | |
| "epoch": 13.0, | |
| "eval_cer": 0.08355624495481094, | |
| "eval_loss": 0.3656393885612488, | |
| "eval_runtime": 30.0647, | |
| "eval_samples_per_second": 27.474, | |
| "eval_steps_per_second": 6.885, | |
| "eval_wer": 0.40114750179297154, | |
| "step": 6669 | |
| }, | |
| { | |
| "epoch": 14.0, | |
| "grad_norm": 2.6208112239837646, | |
| "learning_rate": 8.685433070866143e-05, | |
| "loss": 0.0914, | |
| "step": 7182 | |
| }, | |
| { | |
| "epoch": 14.0, | |
| "eval_cer": 0.08339796451352506, | |
| "eval_loss": 0.3572923541069031, | |
| "eval_runtime": 30.3516, | |
| "eval_samples_per_second": 27.214, | |
| "eval_steps_per_second": 6.82, | |
| "eval_wer": 0.39278030121922064, | |
| "step": 7182 | |
| }, | |
| { | |
| "epoch": 15.0, | |
| "grad_norm": 3.129692316055298, | |
| "learning_rate": 8.584448818897639e-05, | |
| "loss": 0.0836, | |
| "step": 7695 | |
| }, | |
| { | |
| "epoch": 15.0, | |
| "eval_cer": 0.07895028411339211, | |
| "eval_loss": 0.379753053188324, | |
| "eval_runtime": 30.1224, | |
| "eval_samples_per_second": 27.421, | |
| "eval_steps_per_second": 6.872, | |
| "eval_wer": 0.3795123117379871, | |
| "step": 7695 | |
| }, | |
| { | |
| "epoch": 16.0, | |
| "grad_norm": 4.029957294464111, | |
| "learning_rate": 8.483464566929134e-05, | |
| "loss": 0.0729, | |
| "step": 8208 | |
| }, | |
| { | |
| "epoch": 16.0, | |
| "eval_cer": 0.08219503315975245, | |
| "eval_loss": 0.4288513660430908, | |
| "eval_runtime": 30.1728, | |
| "eval_samples_per_second": 27.376, | |
| "eval_steps_per_second": 6.86, | |
| "eval_wer": 0.3895529524264882, | |
| "step": 8208 | |
| }, | |
| { | |
| "epoch": 17.0, | |
| "grad_norm": 0.5638870596885681, | |
| "learning_rate": 8.38248031496063e-05, | |
| "loss": 0.0652, | |
| "step": 8721 | |
| }, | |
| { | |
| "epoch": 17.0, | |
| "eval_cer": 0.087386631633929, | |
| "eval_loss": 0.4778367280960083, | |
| "eval_runtime": 30.6189, | |
| "eval_samples_per_second": 26.977, | |
| "eval_steps_per_second": 6.761, | |
| "eval_wer": 0.4157303370786517, | |
| "step": 8721 | |
| }, | |
| { | |
| "epoch": 18.0, | |
| "grad_norm": 0.010239621624350548, | |
| "learning_rate": 8.281496062992127e-05, | |
| "loss": 0.0599, | |
| "step": 9234 | |
| }, | |
| { | |
| "epoch": 18.0, | |
| "eval_cer": 0.07918770477532092, | |
| "eval_loss": 0.4570063352584839, | |
| "eval_runtime": 30.2821, | |
| "eval_samples_per_second": 27.277, | |
| "eval_steps_per_second": 6.836, | |
| "eval_wer": 0.37963184317475496, | |
| "step": 9234 | |
| }, | |
| { | |
| "epoch": 19.0, | |
| "grad_norm": 3.2985517978668213, | |
| "learning_rate": 8.180511811023623e-05, | |
| "loss": 0.0504, | |
| "step": 9747 | |
| }, | |
| { | |
| "epoch": 19.0, | |
| "eval_cer": 0.08564554677978442, | |
| "eval_loss": 0.44442683458328247, | |
| "eval_runtime": 30.5951, | |
| "eval_samples_per_second": 26.998, | |
| "eval_steps_per_second": 6.766, | |
| "eval_wer": 0.4125029882859192, | |
| "step": 9747 | |
| }, | |
| { | |
| "epoch": 20.0, | |
| "grad_norm": 0.014229847118258476, | |
| "learning_rate": 8.079724409448819e-05, | |
| "loss": 0.0498, | |
| "step": 10260 | |
| }, | |
| { | |
| "epoch": 20.0, | |
| "eval_cer": 0.08422102280821159, | |
| "eval_loss": 0.4611632823944092, | |
| "eval_runtime": 30.4634, | |
| "eval_samples_per_second": 27.115, | |
| "eval_steps_per_second": 6.795, | |
| "eval_wer": 0.3944537413339708, | |
| "step": 10260 | |
| }, | |
| { | |
| "epoch": 21.0, | |
| "grad_norm": 0.029411843046545982, | |
| "learning_rate": 7.978740157480316e-05, | |
| "loss": 0.0434, | |
| "step": 10773 | |
| }, | |
| { | |
| "epoch": 21.0, | |
| "eval_cer": 0.08431599107298311, | |
| "eval_loss": 0.48810598254203796, | |
| "eval_runtime": 30.3298, | |
| "eval_samples_per_second": 27.234, | |
| "eval_steps_per_second": 6.825, | |
| "eval_wer": 0.40019125029882857, | |
| "step": 10773 | |
| }, | |
| { | |
| "epoch": 22.0, | |
| "grad_norm": 0.19160379469394684, | |
| "learning_rate": 7.877755905511811e-05, | |
| "loss": 0.0386, | |
| "step": 11286 | |
| }, | |
| { | |
| "epoch": 22.0, | |
| "eval_cer": 0.08035898004083636, | |
| "eval_loss": 0.509917676448822, | |
| "eval_runtime": 31.249, | |
| "eval_samples_per_second": 26.433, | |
| "eval_steps_per_second": 6.624, | |
| "eval_wer": 0.377719340186469, | |
| "step": 11286 | |
| }, | |
| { | |
| "epoch": 23.0, | |
| "grad_norm": 0.18706804513931274, | |
| "learning_rate": 7.776771653543308e-05, | |
| "loss": 0.0381, | |
| "step": 11799 | |
| }, | |
| { | |
| "epoch": 23.0, | |
| "eval_cer": 0.08241662577755267, | |
| "eval_loss": 0.49042966961860657, | |
| "eval_runtime": 30.3145, | |
| "eval_samples_per_second": 27.248, | |
| "eval_steps_per_second": 6.828, | |
| "eval_wer": 0.38656466650729143, | |
| "step": 11799 | |
| }, | |
| { | |
| "epoch": 24.0, | |
| "grad_norm": 0.048221975564956665, | |
| "learning_rate": 7.675984251968504e-05, | |
| "loss": 0.0344, | |
| "step": 12312 | |
| }, | |
| { | |
| "epoch": 24.0, | |
| "eval_cer": 0.08344544864591083, | |
| "eval_loss": 0.4621882736682892, | |
| "eval_runtime": 30.4774, | |
| "eval_samples_per_second": 27.102, | |
| "eval_steps_per_second": 6.792, | |
| "eval_wer": 0.4028209419077217, | |
| "step": 12312 | |
| }, | |
| { | |
| "epoch": 25.0, | |
| "grad_norm": 1.851014494895935, | |
| "learning_rate": 7.575e-05, | |
| "loss": 0.0302, | |
| "step": 12825 | |
| }, | |
| { | |
| "epoch": 25.0, | |
| "eval_cer": 0.08198926858608081, | |
| "eval_loss": 0.4985555112361908, | |
| "eval_runtime": 30.5557, | |
| "eval_samples_per_second": 27.033, | |
| "eval_steps_per_second": 6.775, | |
| "eval_wer": 0.39182404972507767, | |
| "step": 12825 | |
| } | |
| ], | |
| "logging_steps": 500, | |
| "max_steps": 51300, | |
| "num_input_tokens_seen": 0, | |
| "num_train_epochs": 100, | |
| "save_steps": 500, | |
| "stateful_callbacks": { | |
| "EarlyStoppingCallback": { | |
| "args": { | |
| "early_stopping_patience": 20, | |
| "early_stopping_threshold": 0.001 | |
| }, | |
| "attributes": { | |
| "early_stopping_patience_counter": 3 | |
| } | |
| }, | |
| "TrainerControl": { | |
| "args": { | |
| "should_epoch_stop": false, | |
| "should_evaluate": false, | |
| "should_log": false, | |
| "should_save": true, | |
| "should_training_stop": false | |
| }, | |
| "attributes": {} | |
| } | |
| }, | |
| "total_flos": 2.7282880107684676e+19, | |
| "train_batch_size": 8, | |
| "trial_name": null, | |
| "trial_params": null | |
| } | |