| { | |
| "best_global_step": 800, | |
| "best_metric": 0.6969724297523499, | |
| "best_model_checkpoint": "/scratch/skscla001/experiments/datasets/results/whisper-medium-bigcgen-baseline-42/checkpoint-800", | |
| "epoch": 4.878718535469107, | |
| "eval_steps": 200, | |
| "global_step": 1600, | |
| "is_hyper_param_search": false, | |
| "is_local_process_zero": true, | |
| "is_world_process_zero": true, | |
| "log_history": [ | |
| { | |
| "epoch": 0.07627765064836003, | |
| "grad_norm": 32.69679260253906, | |
| "learning_rate": 4.2000000000000006e-07, | |
| "loss": 3.3916, | |
| "step": 25 | |
| }, | |
| { | |
| "epoch": 0.15255530129672007, | |
| "grad_norm": 27.17233657836914, | |
| "learning_rate": 9.200000000000001e-07, | |
| "loss": 2.7163, | |
| "step": 50 | |
| }, | |
| { | |
| "epoch": 0.2288329519450801, | |
| "grad_norm": 22.44352912902832, | |
| "learning_rate": 1.42e-06, | |
| "loss": 2.1063, | |
| "step": 75 | |
| }, | |
| { | |
| "epoch": 0.30511060259344014, | |
| "grad_norm": 17.82234764099121, | |
| "learning_rate": 1.9200000000000003e-06, | |
| "loss": 1.5258, | |
| "step": 100 | |
| }, | |
| { | |
| "epoch": 0.38138825324180015, | |
| "grad_norm": 16.569429397583008, | |
| "learning_rate": 2.42e-06, | |
| "loss": 1.2767, | |
| "step": 125 | |
| }, | |
| { | |
| "epoch": 0.4576659038901602, | |
| "grad_norm": 18.720935821533203, | |
| "learning_rate": 2.9e-06, | |
| "loss": 1.1612, | |
| "step": 150 | |
| }, | |
| { | |
| "epoch": 0.5339435545385202, | |
| "grad_norm": 18.467731475830078, | |
| "learning_rate": 3.4000000000000005e-06, | |
| "loss": 1.0317, | |
| "step": 175 | |
| }, | |
| { | |
| "epoch": 0.6102212051868803, | |
| "grad_norm": 9.183775901794434, | |
| "learning_rate": 3.900000000000001e-06, | |
| "loss": 1.0745, | |
| "step": 200 | |
| }, | |
| { | |
| "epoch": 0.6102212051868803, | |
| "eval_loss": 0.9366264939308167, | |
| "eval_runtime": 278.3426, | |
| "eval_samples_per_second": 1.638, | |
| "eval_steps_per_second": 0.819, | |
| "eval_wer": 0.6486166942539607, | |
| "step": 200 | |
| }, | |
| { | |
| "epoch": 0.6864988558352403, | |
| "grad_norm": 20.13412857055664, | |
| "learning_rate": 4.4e-06, | |
| "loss": 1.0986, | |
| "step": 225 | |
| }, | |
| { | |
| "epoch": 0.7627765064836003, | |
| "grad_norm": 17.606948852539062, | |
| "learning_rate": 4.9000000000000005e-06, | |
| "loss": 0.8614, | |
| "step": 250 | |
| }, | |
| { | |
| "epoch": 0.8390541571319603, | |
| "grad_norm": 13.448685646057129, | |
| "learning_rate": 5.400000000000001e-06, | |
| "loss": 0.9057, | |
| "step": 275 | |
| }, | |
| { | |
| "epoch": 0.9153318077803204, | |
| "grad_norm": 18.768592834472656, | |
| "learning_rate": 5.9e-06, | |
| "loss": 0.9076, | |
| "step": 300 | |
| }, | |
| { | |
| "epoch": 0.9916094584286804, | |
| "grad_norm": 12.053692817687988, | |
| "learning_rate": 6.4000000000000006e-06, | |
| "loss": 0.7214, | |
| "step": 325 | |
| }, | |
| { | |
| "epoch": 1.0671243325705568, | |
| "grad_norm": 13.019783020019531, | |
| "learning_rate": 6.9e-06, | |
| "loss": 0.7012, | |
| "step": 350 | |
| }, | |
| { | |
| "epoch": 1.1434019832189168, | |
| "grad_norm": 16.344667434692383, | |
| "learning_rate": 7.4e-06, | |
| "loss": 0.7199, | |
| "step": 375 | |
| }, | |
| { | |
| "epoch": 1.2196796338672768, | |
| "grad_norm": 13.561891555786133, | |
| "learning_rate": 7.9e-06, | |
| "loss": 0.6532, | |
| "step": 400 | |
| }, | |
| { | |
| "epoch": 1.2196796338672768, | |
| "eval_loss": 0.7690292000770569, | |
| "eval_runtime": 283.1267, | |
| "eval_samples_per_second": 1.611, | |
| "eval_steps_per_second": 0.805, | |
| "eval_wer": 0.5467013478363679, | |
| "step": 400 | |
| }, | |
| { | |
| "epoch": 1.2959572845156369, | |
| "grad_norm": 15.391003608703613, | |
| "learning_rate": 8.400000000000001e-06, | |
| "loss": 0.616, | |
| "step": 425 | |
| }, | |
| { | |
| "epoch": 1.372234935163997, | |
| "grad_norm": 14.098180770874023, | |
| "learning_rate": 8.900000000000001e-06, | |
| "loss": 0.6977, | |
| "step": 450 | |
| }, | |
| { | |
| "epoch": 1.4485125858123569, | |
| "grad_norm": 12.179559707641602, | |
| "learning_rate": 9.4e-06, | |
| "loss": 0.6716, | |
| "step": 475 | |
| }, | |
| { | |
| "epoch": 1.524790236460717, | |
| "grad_norm": 18.046316146850586, | |
| "learning_rate": 9.9e-06, | |
| "loss": 0.6829, | |
| "step": 500 | |
| }, | |
| { | |
| "epoch": 1.6010678871090769, | |
| "grad_norm": 12.71342658996582, | |
| "learning_rate": 9.955555555555556e-06, | |
| "loss": 0.7193, | |
| "step": 525 | |
| }, | |
| { | |
| "epoch": 1.677345537757437, | |
| "grad_norm": 11.328106880187988, | |
| "learning_rate": 9.9e-06, | |
| "loss": 0.6129, | |
| "step": 550 | |
| }, | |
| { | |
| "epoch": 1.7536231884057971, | |
| "grad_norm": 11.869524002075195, | |
| "learning_rate": 9.844444444444446e-06, | |
| "loss": 0.6352, | |
| "step": 575 | |
| }, | |
| { | |
| "epoch": 1.8299008390541571, | |
| "grad_norm": 12.608686447143555, | |
| "learning_rate": 9.78888888888889e-06, | |
| "loss": 0.6347, | |
| "step": 600 | |
| }, | |
| { | |
| "epoch": 1.8299008390541571, | |
| "eval_loss": 0.7060141563415527, | |
| "eval_runtime": 285.6618, | |
| "eval_samples_per_second": 1.596, | |
| "eval_steps_per_second": 0.798, | |
| "eval_wer": 0.5128872073776306, | |
| "step": 600 | |
| }, | |
| { | |
| "epoch": 1.9061784897025171, | |
| "grad_norm": 14.89091968536377, | |
| "learning_rate": 9.733333333333334e-06, | |
| "loss": 0.6794, | |
| "step": 625 | |
| }, | |
| { | |
| "epoch": 1.9824561403508771, | |
| "grad_norm": 17.0143985748291, | |
| "learning_rate": 9.677777777777778e-06, | |
| "loss": 0.6517, | |
| "step": 650 | |
| }, | |
| { | |
| "epoch": 2.0579710144927534, | |
| "grad_norm": 11.61669921875, | |
| "learning_rate": 9.622222222222222e-06, | |
| "loss": 0.4539, | |
| "step": 675 | |
| }, | |
| { | |
| "epoch": 2.1342486651411137, | |
| "grad_norm": 6.573554515838623, | |
| "learning_rate": 9.566666666666668e-06, | |
| "loss": 0.3674, | |
| "step": 700 | |
| }, | |
| { | |
| "epoch": 2.2105263157894735, | |
| "grad_norm": 11.21182918548584, | |
| "learning_rate": 9.511111111111112e-06, | |
| "loss": 0.3828, | |
| "step": 725 | |
| }, | |
| { | |
| "epoch": 2.2868039664378337, | |
| "grad_norm": 14.237340927124023, | |
| "learning_rate": 9.455555555555557e-06, | |
| "loss": 0.3999, | |
| "step": 750 | |
| }, | |
| { | |
| "epoch": 2.363081617086194, | |
| "grad_norm": 9.172303199768066, | |
| "learning_rate": 9.4e-06, | |
| "loss": 0.4579, | |
| "step": 775 | |
| }, | |
| { | |
| "epoch": 2.4393592677345537, | |
| "grad_norm": 10.187737464904785, | |
| "learning_rate": 9.344444444444446e-06, | |
| "loss": 0.4066, | |
| "step": 800 | |
| }, | |
| { | |
| "epoch": 2.4393592677345537, | |
| "eval_loss": 0.6969724297523499, | |
| "eval_runtime": 289.906, | |
| "eval_samples_per_second": 1.573, | |
| "eval_steps_per_second": 0.786, | |
| "eval_wer": 0.526129108536297, | |
| "step": 800 | |
| }, | |
| { | |
| "epoch": 2.515636918382914, | |
| "grad_norm": 11.924391746520996, | |
| "learning_rate": 9.28888888888889e-06, | |
| "loss": 0.382, | |
| "step": 825 | |
| }, | |
| { | |
| "epoch": 2.5919145690312737, | |
| "grad_norm": 10.492657661437988, | |
| "learning_rate": 9.233333333333334e-06, | |
| "loss": 0.3796, | |
| "step": 850 | |
| }, | |
| { | |
| "epoch": 2.668192219679634, | |
| "grad_norm": 9.099035263061523, | |
| "learning_rate": 9.17777777777778e-06, | |
| "loss": 0.5154, | |
| "step": 875 | |
| }, | |
| { | |
| "epoch": 2.744469870327994, | |
| "grad_norm": 11.572857856750488, | |
| "learning_rate": 9.122222222222223e-06, | |
| "loss": 0.4487, | |
| "step": 900 | |
| }, | |
| { | |
| "epoch": 2.820747520976354, | |
| "grad_norm": 13.59435749053955, | |
| "learning_rate": 9.066666666666667e-06, | |
| "loss": 0.4478, | |
| "step": 925 | |
| }, | |
| { | |
| "epoch": 2.8970251716247137, | |
| "grad_norm": 11.392141342163086, | |
| "learning_rate": 9.011111111111111e-06, | |
| "loss": 0.352, | |
| "step": 950 | |
| }, | |
| { | |
| "epoch": 2.973302822273074, | |
| "grad_norm": 8.065723419189453, | |
| "learning_rate": 8.955555555555555e-06, | |
| "loss": 0.3986, | |
| "step": 975 | |
| }, | |
| { | |
| "epoch": 3.0488176964149503, | |
| "grad_norm": 7.743978977203369, | |
| "learning_rate": 8.900000000000001e-06, | |
| "loss": 0.2542, | |
| "step": 1000 | |
| }, | |
| { | |
| "epoch": 3.0488176964149503, | |
| "eval_loss": 0.7139758467674255, | |
| "eval_runtime": 280.6617, | |
| "eval_samples_per_second": 1.625, | |
| "eval_steps_per_second": 0.812, | |
| "eval_wer": 0.5034287065500118, | |
| "step": 1000 | |
| }, | |
| { | |
| "epoch": 3.1250953470633105, | |
| "grad_norm": 6.686219215393066, | |
| "learning_rate": 8.844444444444445e-06, | |
| "loss": 0.1859, | |
| "step": 1025 | |
| }, | |
| { | |
| "epoch": 3.2013729977116703, | |
| "grad_norm": 8.531457901000977, | |
| "learning_rate": 8.788888888888891e-06, | |
| "loss": 0.2228, | |
| "step": 1050 | |
| }, | |
| { | |
| "epoch": 3.2776506483600305, | |
| "grad_norm": 8.079030990600586, | |
| "learning_rate": 8.733333333333333e-06, | |
| "loss": 0.233, | |
| "step": 1075 | |
| }, | |
| { | |
| "epoch": 3.3539282990083907, | |
| "grad_norm": 7.210234642028809, | |
| "learning_rate": 8.677777777777779e-06, | |
| "loss": 0.278, | |
| "step": 1100 | |
| }, | |
| { | |
| "epoch": 3.4302059496567505, | |
| "grad_norm": 9.321666717529297, | |
| "learning_rate": 8.622222222222223e-06, | |
| "loss": 0.2022, | |
| "step": 1125 | |
| }, | |
| { | |
| "epoch": 3.5064836003051107, | |
| "grad_norm": 8.422845840454102, | |
| "learning_rate": 8.566666666666667e-06, | |
| "loss": 0.2462, | |
| "step": 1150 | |
| }, | |
| { | |
| "epoch": 3.5827612509534705, | |
| "grad_norm": 7.176828384399414, | |
| "learning_rate": 8.511111111111113e-06, | |
| "loss": 0.2568, | |
| "step": 1175 | |
| }, | |
| { | |
| "epoch": 3.6590389016018308, | |
| "grad_norm": 7.077821731567383, | |
| "learning_rate": 8.455555555555555e-06, | |
| "loss": 0.252, | |
| "step": 1200 | |
| }, | |
| { | |
| "epoch": 3.6590389016018308, | |
| "eval_loss": 0.7221043705940247, | |
| "eval_runtime": 282.4738, | |
| "eval_samples_per_second": 1.614, | |
| "eval_steps_per_second": 0.807, | |
| "eval_wer": 0.4833293922913218, | |
| "step": 1200 | |
| }, | |
| { | |
| "epoch": 3.7353165522501905, | |
| "grad_norm": 8.898653984069824, | |
| "learning_rate": 8.400000000000001e-06, | |
| "loss": 0.2351, | |
| "step": 1225 | |
| }, | |
| { | |
| "epoch": 3.8115942028985508, | |
| "grad_norm": 8.900397300720215, | |
| "learning_rate": 8.344444444444445e-06, | |
| "loss": 0.2656, | |
| "step": 1250 | |
| }, | |
| { | |
| "epoch": 3.887871853546911, | |
| "grad_norm": 6.858781814575195, | |
| "learning_rate": 8.288888888888889e-06, | |
| "loss": 0.2253, | |
| "step": 1275 | |
| }, | |
| { | |
| "epoch": 3.964149504195271, | |
| "grad_norm": 5.616495609283447, | |
| "learning_rate": 8.233333333333335e-06, | |
| "loss": 0.2518, | |
| "step": 1300 | |
| }, | |
| { | |
| "epoch": 4.0396643783371475, | |
| "grad_norm": 2.58622407913208, | |
| "learning_rate": 8.177777777777779e-06, | |
| "loss": 0.1329, | |
| "step": 1325 | |
| }, | |
| { | |
| "epoch": 4.115942028985507, | |
| "grad_norm": 4.7402544021606445, | |
| "learning_rate": 8.122222222222223e-06, | |
| "loss": 0.0939, | |
| "step": 1350 | |
| }, | |
| { | |
| "epoch": 4.192219679633867, | |
| "grad_norm": 7.061638355255127, | |
| "learning_rate": 8.066666666666667e-06, | |
| "loss": 0.1122, | |
| "step": 1375 | |
| }, | |
| { | |
| "epoch": 4.268497330282227, | |
| "grad_norm": 3.740872383117676, | |
| "learning_rate": 8.011111111111113e-06, | |
| "loss": 0.137, | |
| "step": 1400 | |
| }, | |
| { | |
| "epoch": 4.268497330282227, | |
| "eval_loss": 0.757293701171875, | |
| "eval_runtime": 278.9207, | |
| "eval_samples_per_second": 1.635, | |
| "eval_steps_per_second": 0.817, | |
| "eval_wer": 0.4878221801844408, | |
| "step": 1400 | |
| }, | |
| { | |
| "epoch": 4.344774980930588, | |
| "grad_norm": 4.499791622161865, | |
| "learning_rate": 7.955555555555557e-06, | |
| "loss": 0.0954, | |
| "step": 1425 | |
| }, | |
| { | |
| "epoch": 4.421052631578947, | |
| "grad_norm": 4.751846790313721, | |
| "learning_rate": 7.9e-06, | |
| "loss": 0.1071, | |
| "step": 1450 | |
| }, | |
| { | |
| "epoch": 4.497330282227307, | |
| "grad_norm": 9.211798667907715, | |
| "learning_rate": 7.844444444444446e-06, | |
| "loss": 0.102, | |
| "step": 1475 | |
| }, | |
| { | |
| "epoch": 4.573607932875667, | |
| "grad_norm": 6.26869535446167, | |
| "learning_rate": 7.788888888888889e-06, | |
| "loss": 0.0968, | |
| "step": 1500 | |
| }, | |
| { | |
| "epoch": 4.649885583524028, | |
| "grad_norm": 6.295158386230469, | |
| "learning_rate": 7.733333333333334e-06, | |
| "loss": 0.1196, | |
| "step": 1525 | |
| }, | |
| { | |
| "epoch": 4.726163234172388, | |
| "grad_norm": 9.435331344604492, | |
| "learning_rate": 7.677777777777778e-06, | |
| "loss": 0.1279, | |
| "step": 1550 | |
| }, | |
| { | |
| "epoch": 4.802440884820747, | |
| "grad_norm": 3.417738914489746, | |
| "learning_rate": 7.622222222222223e-06, | |
| "loss": 0.1082, | |
| "step": 1575 | |
| }, | |
| { | |
| "epoch": 4.878718535469107, | |
| "grad_norm": 7.170241832733154, | |
| "learning_rate": 7.566666666666667e-06, | |
| "loss": 0.114, | |
| "step": 1600 | |
| }, | |
| { | |
| "epoch": 4.878718535469107, | |
| "eval_loss": 0.793424129486084, | |
| "eval_runtime": 277.9238, | |
| "eval_samples_per_second": 1.641, | |
| "eval_steps_per_second": 0.82, | |
| "eval_wer": 0.4812012296051076, | |
| "step": 1600 | |
| }, | |
| { | |
| "epoch": 4.878718535469107, | |
| "step": 1600, | |
| "total_flos": 1.305149555736576e+19, | |
| "train_loss": 0.5828571186959743, | |
| "train_runtime": 5664.3172, | |
| "train_samples_per_second": 7.062, | |
| "train_steps_per_second": 0.883 | |
| } | |
| ], | |
| "logging_steps": 25, | |
| "max_steps": 5000, | |
| "num_input_tokens_seen": 0, | |
| "num_train_epochs": 16, | |
| "save_steps": 200, | |
| "stateful_callbacks": { | |
| "EarlyStoppingCallback": { | |
| "args": { | |
| "early_stopping_patience": 4, | |
| "early_stopping_threshold": 0.0 | |
| }, | |
| "attributes": { | |
| "early_stopping_patience_counter": 4 | |
| } | |
| }, | |
| "TrainerControl": { | |
| "args": { | |
| "should_epoch_stop": false, | |
| "should_evaluate": false, | |
| "should_log": false, | |
| "should_save": true, | |
| "should_training_stop": true | |
| }, | |
| "attributes": {} | |
| } | |
| }, | |
| "total_flos": 1.305149555736576e+19, | |
| "train_batch_size": 2, | |
| "trial_name": null, | |
| "trial_params": null | |
| } | |