| { | |
| "best_metric": 0.4265601933002472, | |
| "best_model_checkpoint": "/scratch/skscla001/speech/results/mms-1b-toigen-male-model/checkpoint-2600", | |
| "epoch": 15.151515151515152, | |
| "eval_steps": 100, | |
| "global_step": 3000, | |
| "is_hyper_param_search": false, | |
| "is_local_process_zero": true, | |
| "is_world_process_zero": true, | |
| "log_history": [ | |
| { | |
| "epoch": 0.5050505050505051, | |
| "grad_norm": 7.066163539886475, | |
| "learning_rate": 0.00028799999999999995, | |
| "loss": 8.0084, | |
| "step": 100 | |
| }, | |
| { | |
| "epoch": 0.5050505050505051, | |
| "eval_loss": 3.745828628540039, | |
| "eval_runtime": 18.1656, | |
| "eval_samples_per_second": 10.68, | |
| "eval_steps_per_second": 2.697, | |
| "eval_wer": 0.9983853606027987, | |
| "step": 100 | |
| }, | |
| { | |
| "epoch": 1.0101010101010102, | |
| "grad_norm": 3.717050552368164, | |
| "learning_rate": 0.0002950684931506849, | |
| "loss": 2.8136, | |
| "step": 200 | |
| }, | |
| { | |
| "epoch": 1.0101010101010102, | |
| "eval_loss": 1.0028560161590576, | |
| "eval_runtime": 18.1412, | |
| "eval_samples_per_second": 10.694, | |
| "eval_steps_per_second": 2.701, | |
| "eval_wer": 0.7421959095801938, | |
| "step": 200 | |
| }, | |
| { | |
| "epoch": 1.5151515151515151, | |
| "grad_norm": 2.0190610885620117, | |
| "learning_rate": 0.00028993150684931507, | |
| "loss": 0.9647, | |
| "step": 300 | |
| }, | |
| { | |
| "epoch": 1.5151515151515151, | |
| "eval_loss": 0.5869513154029846, | |
| "eval_runtime": 18.1111, | |
| "eval_samples_per_second": 10.712, | |
| "eval_steps_per_second": 2.706, | |
| "eval_wer": 0.527448869752422, | |
| "step": 300 | |
| }, | |
| { | |
| "epoch": 2.0202020202020203, | |
| "grad_norm": 1.3322362899780273, | |
| "learning_rate": 0.0002847945205479452, | |
| "loss": 0.8539, | |
| "step": 400 | |
| }, | |
| { | |
| "epoch": 2.0202020202020203, | |
| "eval_loss": 0.5461000800132751, | |
| "eval_runtime": 18.1018, | |
| "eval_samples_per_second": 10.717, | |
| "eval_steps_per_second": 2.707, | |
| "eval_wer": 0.5048439181916039, | |
| "step": 400 | |
| }, | |
| { | |
| "epoch": 2.525252525252525, | |
| "grad_norm": 2.1069583892822266, | |
| "learning_rate": 0.0002796575342465753, | |
| "loss": 0.7525, | |
| "step": 500 | |
| }, | |
| { | |
| "epoch": 2.525252525252525, | |
| "eval_loss": 0.5256428122520447, | |
| "eval_runtime": 18.1961, | |
| "eval_samples_per_second": 10.662, | |
| "eval_steps_per_second": 2.693, | |
| "eval_wer": 0.49892357373519913, | |
| "step": 500 | |
| }, | |
| { | |
| "epoch": 3.0303030303030303, | |
| "grad_norm": 2.1016767024993896, | |
| "learning_rate": 0.0002745205479452055, | |
| "loss": 0.7307, | |
| "step": 600 | |
| }, | |
| { | |
| "epoch": 3.0303030303030303, | |
| "eval_loss": 0.5101361274719238, | |
| "eval_runtime": 18.1622, | |
| "eval_samples_per_second": 10.682, | |
| "eval_steps_per_second": 2.698, | |
| "eval_wer": 0.48708288482238965, | |
| "step": 600 | |
| }, | |
| { | |
| "epoch": 3.5353535353535355, | |
| "grad_norm": 1.1222645044326782, | |
| "learning_rate": 0.0002693835616438356, | |
| "loss": 0.6997, | |
| "step": 700 | |
| }, | |
| { | |
| "epoch": 3.5353535353535355, | |
| "eval_loss": 0.5032119750976562, | |
| "eval_runtime": 18.0324, | |
| "eval_samples_per_second": 10.758, | |
| "eval_steps_per_second": 2.717, | |
| "eval_wer": 0.468783638320775, | |
| "step": 700 | |
| }, | |
| { | |
| "epoch": 4.040404040404041, | |
| "grad_norm": 1.4940961599349976, | |
| "learning_rate": 0.0002642465753424657, | |
| "loss": 0.6882, | |
| "step": 800 | |
| }, | |
| { | |
| "epoch": 4.040404040404041, | |
| "eval_loss": 0.4878801703453064, | |
| "eval_runtime": 18.1093, | |
| "eval_samples_per_second": 10.713, | |
| "eval_steps_per_second": 2.706, | |
| "eval_wer": 0.4736275565123789, | |
| "step": 800 | |
| }, | |
| { | |
| "epoch": 4.545454545454545, | |
| "grad_norm": 0.9596569538116455, | |
| "learning_rate": 0.0002591095890410959, | |
| "loss": 0.651, | |
| "step": 900 | |
| }, | |
| { | |
| "epoch": 4.545454545454545, | |
| "eval_loss": 0.4788007438182831, | |
| "eval_runtime": 18.2582, | |
| "eval_samples_per_second": 10.625, | |
| "eval_steps_per_second": 2.684, | |
| "eval_wer": 0.45586652314316467, | |
| "step": 900 | |
| }, | |
| { | |
| "epoch": 5.05050505050505, | |
| "grad_norm": 6.235934734344482, | |
| "learning_rate": 0.000253972602739726, | |
| "loss": 0.6623, | |
| "step": 1000 | |
| }, | |
| { | |
| "epoch": 5.05050505050505, | |
| "eval_loss": 0.4798834025859833, | |
| "eval_runtime": 18.2341, | |
| "eval_samples_per_second": 10.639, | |
| "eval_steps_per_second": 2.687, | |
| "eval_wer": 0.4526372443487621, | |
| "step": 1000 | |
| }, | |
| { | |
| "epoch": 5.555555555555555, | |
| "grad_norm": 1.2491483688354492, | |
| "learning_rate": 0.00024883561643835613, | |
| "loss": 0.6339, | |
| "step": 1100 | |
| }, | |
| { | |
| "epoch": 5.555555555555555, | |
| "eval_loss": 0.467692494392395, | |
| "eval_runtime": 18.2104, | |
| "eval_samples_per_second": 10.653, | |
| "eval_steps_per_second": 2.691, | |
| "eval_wer": 0.4418729817007535, | |
| "step": 1100 | |
| }, | |
| { | |
| "epoch": 6.0606060606060606, | |
| "grad_norm": 1.8275915384292603, | |
| "learning_rate": 0.0002436986301369863, | |
| "loss": 0.6424, | |
| "step": 1200 | |
| }, | |
| { | |
| "epoch": 6.0606060606060606, | |
| "eval_loss": 0.46495306491851807, | |
| "eval_runtime": 18.1846, | |
| "eval_samples_per_second": 10.668, | |
| "eval_steps_per_second": 2.695, | |
| "eval_wer": 0.4429494079655544, | |
| "step": 1200 | |
| }, | |
| { | |
| "epoch": 6.565656565656566, | |
| "grad_norm": 12.825818061828613, | |
| "learning_rate": 0.00023856164383561642, | |
| "loss": 0.6365, | |
| "step": 1300 | |
| }, | |
| { | |
| "epoch": 6.565656565656566, | |
| "eval_loss": 0.47455188632011414, | |
| "eval_runtime": 18.2175, | |
| "eval_samples_per_second": 10.649, | |
| "eval_steps_per_second": 2.69, | |
| "eval_wer": 0.4461786867599569, | |
| "step": 1300 | |
| }, | |
| { | |
| "epoch": 7.070707070707071, | |
| "grad_norm": 0.8040870428085327, | |
| "learning_rate": 0.00023342465753424657, | |
| "loss": 0.556, | |
| "step": 1400 | |
| }, | |
| { | |
| "epoch": 7.070707070707071, | |
| "eval_loss": 0.45120927691459656, | |
| "eval_runtime": 18.2585, | |
| "eval_samples_per_second": 10.625, | |
| "eval_steps_per_second": 2.684, | |
| "eval_wer": 0.4381054897739505, | |
| "step": 1400 | |
| }, | |
| { | |
| "epoch": 7.575757575757576, | |
| "grad_norm": 1.3401854038238525, | |
| "learning_rate": 0.00022828767123287669, | |
| "loss": 0.5969, | |
| "step": 1500 | |
| }, | |
| { | |
| "epoch": 7.575757575757576, | |
| "eval_loss": 0.4596523940563202, | |
| "eval_runtime": 18.1217, | |
| "eval_samples_per_second": 10.705, | |
| "eval_steps_per_second": 2.704, | |
| "eval_wer": 0.4413347685683531, | |
| "step": 1500 | |
| }, | |
| { | |
| "epoch": 8.080808080808081, | |
| "grad_norm": 1.6330450773239136, | |
| "learning_rate": 0.00022315068493150683, | |
| "loss": 0.5772, | |
| "step": 1600 | |
| }, | |
| { | |
| "epoch": 8.080808080808081, | |
| "eval_loss": 0.4455384612083435, | |
| "eval_runtime": 18.0777, | |
| "eval_samples_per_second": 10.731, | |
| "eval_steps_per_second": 2.711, | |
| "eval_wer": 0.4284176533907427, | |
| "step": 1600 | |
| }, | |
| { | |
| "epoch": 8.585858585858587, | |
| "grad_norm": 2.6647837162017822, | |
| "learning_rate": 0.00021801369863013698, | |
| "loss": 0.5695, | |
| "step": 1700 | |
| }, | |
| { | |
| "epoch": 8.585858585858587, | |
| "eval_loss": 0.45646098256111145, | |
| "eval_runtime": 18.2232, | |
| "eval_samples_per_second": 10.646, | |
| "eval_steps_per_second": 2.689, | |
| "eval_wer": 0.4268030139935414, | |
| "step": 1700 | |
| }, | |
| { | |
| "epoch": 9.090909090909092, | |
| "grad_norm": 0.6165424585342407, | |
| "learning_rate": 0.0002129280821917808, | |
| "loss": 0.5752, | |
| "step": 1800 | |
| }, | |
| { | |
| "epoch": 9.090909090909092, | |
| "eval_loss": 0.44141411781311035, | |
| "eval_runtime": 18.1955, | |
| "eval_samples_per_second": 10.662, | |
| "eval_steps_per_second": 2.693, | |
| "eval_wer": 0.418729817007535, | |
| "step": 1800 | |
| }, | |
| { | |
| "epoch": 9.595959595959595, | |
| "grad_norm": 1.7799068689346313, | |
| "learning_rate": 0.00020779109589041093, | |
| "loss": 0.5734, | |
| "step": 1900 | |
| }, | |
| { | |
| "epoch": 9.595959595959595, | |
| "eval_loss": 0.4450058937072754, | |
| "eval_runtime": 18.1771, | |
| "eval_samples_per_second": 10.673, | |
| "eval_steps_per_second": 2.696, | |
| "eval_wer": 0.4085037674919268, | |
| "step": 1900 | |
| }, | |
| { | |
| "epoch": 10.1010101010101, | |
| "grad_norm": 4.881310939788818, | |
| "learning_rate": 0.00020265410958904108, | |
| "loss": 0.5465, | |
| "step": 2000 | |
| }, | |
| { | |
| "epoch": 10.1010101010101, | |
| "eval_loss": 0.43734970688819885, | |
| "eval_runtime": 18.1427, | |
| "eval_samples_per_second": 10.693, | |
| "eval_steps_per_second": 2.701, | |
| "eval_wer": 0.4155005382131324, | |
| "step": 2000 | |
| }, | |
| { | |
| "epoch": 10.606060606060606, | |
| "grad_norm": 4.940012454986572, | |
| "learning_rate": 0.00019751712328767122, | |
| "loss": 0.5553, | |
| "step": 2100 | |
| }, | |
| { | |
| "epoch": 10.606060606060606, | |
| "eval_loss": 0.45197948813438416, | |
| "eval_runtime": 18.2849, | |
| "eval_samples_per_second": 10.61, | |
| "eval_steps_per_second": 2.68, | |
| "eval_wer": 0.4241119483315393, | |
| "step": 2100 | |
| }, | |
| { | |
| "epoch": 11.11111111111111, | |
| "grad_norm": 1.0321776866912842, | |
| "learning_rate": 0.00019238013698630134, | |
| "loss": 0.5289, | |
| "step": 2200 | |
| }, | |
| { | |
| "epoch": 11.11111111111111, | |
| "eval_loss": 0.430617094039917, | |
| "eval_runtime": 18.3569, | |
| "eval_samples_per_second": 10.568, | |
| "eval_steps_per_second": 2.669, | |
| "eval_wer": 0.4085037674919268, | |
| "step": 2200 | |
| }, | |
| { | |
| "epoch": 11.616161616161616, | |
| "grad_norm": 1.051830768585205, | |
| "learning_rate": 0.0001872431506849315, | |
| "loss": 0.5122, | |
| "step": 2300 | |
| }, | |
| { | |
| "epoch": 11.616161616161616, | |
| "eval_loss": 0.4372076392173767, | |
| "eval_runtime": 18.1559, | |
| "eval_samples_per_second": 10.685, | |
| "eval_steps_per_second": 2.699, | |
| "eval_wer": 0.4015069967707212, | |
| "step": 2300 | |
| }, | |
| { | |
| "epoch": 12.121212121212121, | |
| "grad_norm": 1.2487713098526, | |
| "learning_rate": 0.0001821061643835616, | |
| "loss": 0.5659, | |
| "step": 2400 | |
| }, | |
| { | |
| "epoch": 12.121212121212121, | |
| "eval_loss": 0.4407643973827362, | |
| "eval_runtime": 18.201, | |
| "eval_samples_per_second": 10.659, | |
| "eval_steps_per_second": 2.692, | |
| "eval_wer": 0.4009687836383208, | |
| "step": 2400 | |
| }, | |
| { | |
| "epoch": 12.626262626262626, | |
| "grad_norm": 0.9535735845565796, | |
| "learning_rate": 0.00017696917808219175, | |
| "loss": 0.5007, | |
| "step": 2500 | |
| }, | |
| { | |
| "epoch": 12.626262626262626, | |
| "eval_loss": 0.4274228811264038, | |
| "eval_runtime": 18.2686, | |
| "eval_samples_per_second": 10.619, | |
| "eval_steps_per_second": 2.682, | |
| "eval_wer": 0.3982777179763186, | |
| "step": 2500 | |
| }, | |
| { | |
| "epoch": 13.131313131313131, | |
| "grad_norm": 0.9828107357025146, | |
| "learning_rate": 0.00017188356164383559, | |
| "loss": 0.5366, | |
| "step": 2600 | |
| }, | |
| { | |
| "epoch": 13.131313131313131, | |
| "eval_loss": 0.4265601933002472, | |
| "eval_runtime": 18.3638, | |
| "eval_samples_per_second": 10.564, | |
| "eval_steps_per_second": 2.668, | |
| "eval_wer": 0.4025834230355221, | |
| "step": 2600 | |
| }, | |
| { | |
| "epoch": 13.636363636363637, | |
| "grad_norm": 1.9459073543548584, | |
| "learning_rate": 0.00016674657534246576, | |
| "loss": 0.5068, | |
| "step": 2700 | |
| }, | |
| { | |
| "epoch": 13.636363636363637, | |
| "eval_loss": 0.43655750155448914, | |
| "eval_runtime": 18.2672, | |
| "eval_samples_per_second": 10.62, | |
| "eval_steps_per_second": 2.682, | |
| "eval_wer": 0.3961248654467169, | |
| "step": 2700 | |
| }, | |
| { | |
| "epoch": 14.141414141414142, | |
| "grad_norm": 1.0564210414886475, | |
| "learning_rate": 0.00016160958904109588, | |
| "loss": 0.507, | |
| "step": 2800 | |
| }, | |
| { | |
| "epoch": 14.141414141414142, | |
| "eval_loss": 0.4359418749809265, | |
| "eval_runtime": 18.2049, | |
| "eval_samples_per_second": 10.656, | |
| "eval_steps_per_second": 2.692, | |
| "eval_wer": 0.39720129171151775, | |
| "step": 2800 | |
| }, | |
| { | |
| "epoch": 14.646464646464647, | |
| "grad_norm": 0.9299136400222778, | |
| "learning_rate": 0.00015647260273972602, | |
| "loss": 0.5031, | |
| "step": 2900 | |
| }, | |
| { | |
| "epoch": 14.646464646464647, | |
| "eval_loss": 0.4333823323249817, | |
| "eval_runtime": 18.3722, | |
| "eval_samples_per_second": 10.559, | |
| "eval_steps_per_second": 2.667, | |
| "eval_wer": 0.3966630785791173, | |
| "step": 2900 | |
| }, | |
| { | |
| "epoch": 15.151515151515152, | |
| "grad_norm": 2.234483242034912, | |
| "learning_rate": 0.00015133561643835617, | |
| "loss": 0.4949, | |
| "step": 3000 | |
| }, | |
| { | |
| "epoch": 15.151515151515152, | |
| "eval_loss": 0.43481922149658203, | |
| "eval_runtime": 18.3081, | |
| "eval_samples_per_second": 10.596, | |
| "eval_steps_per_second": 2.676, | |
| "eval_wer": 0.39881593110871905, | |
| "step": 3000 | |
| }, | |
| { | |
| "epoch": 15.151515151515152, | |
| "step": 3000, | |
| "total_flos": 1.0546657971478737e+19, | |
| "train_loss": 0.9314683748881022, | |
| "train_runtime": 2759.24, | |
| "train_samples_per_second": 8.611, | |
| "train_steps_per_second": 2.153 | |
| } | |
| ], | |
| "logging_steps": 100, | |
| "max_steps": 5940, | |
| "num_input_tokens_seen": 0, | |
| "num_train_epochs": 30, | |
| "save_steps": 400, | |
| "stateful_callbacks": { | |
| "EarlyStoppingCallback": { | |
| "args": { | |
| "early_stopping_patience": 4, | |
| "early_stopping_threshold": 0.0 | |
| }, | |
| "attributes": { | |
| "early_stopping_patience_counter": 2 | |
| } | |
| }, | |
| "TrainerControl": { | |
| "args": { | |
| "should_epoch_stop": false, | |
| "should_evaluate": false, | |
| "should_log": false, | |
| "should_save": true, | |
| "should_training_stop": false | |
| }, | |
| "attributes": {} | |
| } | |
| }, | |
| "total_flos": 1.0546657971478737e+19, | |
| "train_batch_size": 4, | |
| "trial_name": null, | |
| "trial_params": null | |
| } | |