| { | |
| "best_metric": 0.6288811564445496, | |
| "best_model_checkpoint": "/scratch/skscla001/results/w2v-bert-bem-natbed-combined-model/checkpoint-2400", | |
| "epoch": 7.509386733416771, | |
| "eval_steps": 200, | |
| "global_step": 3000, | |
| "is_hyper_param_search": false, | |
| "is_local_process_zero": true, | |
| "is_world_process_zero": true, | |
| "log_history": [ | |
| { | |
| "epoch": 0.2503128911138924, | |
| "grad_norm": 3.927278518676758, | |
| "learning_rate": 0.00029699999999999996, | |
| "loss": 2.3529, | |
| "step": 100 | |
| }, | |
| { | |
| "epoch": 0.5006257822277848, | |
| "grad_norm": 2.2314515113830566, | |
| "learning_rate": 0.00029749789385004213, | |
| "loss": 1.0078, | |
| "step": 200 | |
| }, | |
| { | |
| "epoch": 0.5006257822277848, | |
| "eval_loss": 0.9815022945404053, | |
| "eval_runtime": 61.0946, | |
| "eval_samples_per_second": 22.228, | |
| "eval_steps_per_second": 2.783, | |
| "eval_wer": 0.8200802555073294, | |
| "step": 200 | |
| }, | |
| { | |
| "epoch": 0.7509386733416771, | |
| "grad_norm": 2.754998207092285, | |
| "learning_rate": 0.0002949705139005897, | |
| "loss": 0.9989, | |
| "step": 300 | |
| }, | |
| { | |
| "epoch": 1.0012515644555695, | |
| "grad_norm": 3.1453473567962646, | |
| "learning_rate": 0.0002924431339511373, | |
| "loss": 0.8769, | |
| "step": 400 | |
| }, | |
| { | |
| "epoch": 1.0012515644555695, | |
| "eval_loss": 0.9823216199874878, | |
| "eval_runtime": 61.3287, | |
| "eval_samples_per_second": 22.143, | |
| "eval_steps_per_second": 2.772, | |
| "eval_wer": 1.0433215952829415, | |
| "step": 400 | |
| }, | |
| { | |
| "epoch": 1.2515644555694618, | |
| "grad_norm": 11.187702178955078, | |
| "learning_rate": 0.0002899157540016849, | |
| "loss": 0.8422, | |
| "step": 500 | |
| }, | |
| { | |
| "epoch": 1.5018773466833542, | |
| "grad_norm": 2.2132036685943604, | |
| "learning_rate": 0.0002873883740522325, | |
| "loss": 0.805, | |
| "step": 600 | |
| }, | |
| { | |
| "epoch": 1.5018773466833542, | |
| "eval_loss": 0.8306216597557068, | |
| "eval_runtime": 61.3405, | |
| "eval_samples_per_second": 22.139, | |
| "eval_steps_per_second": 2.771, | |
| "eval_wer": 0.8606174760461879, | |
| "step": 600 | |
| }, | |
| { | |
| "epoch": 1.7521902377972465, | |
| "grad_norm": 8.310944557189941, | |
| "learning_rate": 0.0002848609941027801, | |
| "loss": 0.7299, | |
| "step": 700 | |
| }, | |
| { | |
| "epoch": 2.002503128911139, | |
| "grad_norm": 0.9676926136016846, | |
| "learning_rate": 0.0002823336141533277, | |
| "loss": 0.8141, | |
| "step": 800 | |
| }, | |
| { | |
| "epoch": 2.002503128911139, | |
| "eval_loss": 0.7547941207885742, | |
| "eval_runtime": 61.8648, | |
| "eval_samples_per_second": 21.951, | |
| "eval_steps_per_second": 2.748, | |
| "eval_wer": 0.7195970845958561, | |
| "step": 800 | |
| }, | |
| { | |
| "epoch": 2.252816020025031, | |
| "grad_norm": Infinity, | |
| "learning_rate": 0.0002798315080033698, | |
| "loss": 0.7184, | |
| "step": 900 | |
| }, | |
| { | |
| "epoch": 2.5031289111389237, | |
| "grad_norm": 2.475212335586548, | |
| "learning_rate": 0.00027730412805391743, | |
| "loss": 0.7132, | |
| "step": 1000 | |
| }, | |
| { | |
| "epoch": 2.5031289111389237, | |
| "eval_loss": 0.7485368847846985, | |
| "eval_runtime": 65.846, | |
| "eval_samples_per_second": 20.624, | |
| "eval_steps_per_second": 2.582, | |
| "eval_wer": 0.6932274179018917, | |
| "step": 1000 | |
| }, | |
| { | |
| "epoch": 2.7534418022528158, | |
| "grad_norm": 12.120068550109863, | |
| "learning_rate": 0.000274776748104465, | |
| "loss": 0.706, | |
| "step": 1100 | |
| }, | |
| { | |
| "epoch": 3.0037546933667083, | |
| "grad_norm": 1.1044033765792847, | |
| "learning_rate": 0.0002722493681550126, | |
| "loss": 0.7058, | |
| "step": 1200 | |
| }, | |
| { | |
| "epoch": 3.0037546933667083, | |
| "eval_loss": 0.7280192375183105, | |
| "eval_runtime": 61.369, | |
| "eval_samples_per_second": 22.128, | |
| "eval_steps_per_second": 2.77, | |
| "eval_wer": 0.6916714437801982, | |
| "step": 1200 | |
| }, | |
| { | |
| "epoch": 3.254067584480601, | |
| "grad_norm": 1.5886726379394531, | |
| "learning_rate": 0.0002697219882055602, | |
| "loss": 0.6273, | |
| "step": 1300 | |
| }, | |
| { | |
| "epoch": 3.504380475594493, | |
| "grad_norm": 5.7870659828186035, | |
| "learning_rate": 0.0002671946082561078, | |
| "loss": 0.6563, | |
| "step": 1400 | |
| }, | |
| { | |
| "epoch": 3.504380475594493, | |
| "eval_loss": 0.7046008706092834, | |
| "eval_runtime": 61.6596, | |
| "eval_samples_per_second": 22.024, | |
| "eval_steps_per_second": 2.757, | |
| "eval_wer": 0.7045287036278765, | |
| "step": 1400 | |
| }, | |
| { | |
| "epoch": 3.7546933667083855, | |
| "grad_norm": 1.0691437721252441, | |
| "learning_rate": 0.0002646672283066554, | |
| "loss": 0.6538, | |
| "step": 1500 | |
| }, | |
| { | |
| "epoch": 4.005006257822278, | |
| "grad_norm": 3.255420684814453, | |
| "learning_rate": 0.000262139848357203, | |
| "loss": 0.6232, | |
| "step": 1600 | |
| }, | |
| { | |
| "epoch": 4.005006257822278, | |
| "eval_loss": 0.7186261415481567, | |
| "eval_runtime": 61.5315, | |
| "eval_samples_per_second": 22.07, | |
| "eval_steps_per_second": 2.763, | |
| "eval_wer": 0.7408893620506101, | |
| "step": 1600 | |
| }, | |
| { | |
| "epoch": 4.25531914893617, | |
| "grad_norm": 0.7094090580940247, | |
| "learning_rate": 0.00025961246840775057, | |
| "loss": 0.585, | |
| "step": 1700 | |
| }, | |
| { | |
| "epoch": 4.505632040050062, | |
| "grad_norm": 3.2348897457122803, | |
| "learning_rate": 0.0002570850884582982, | |
| "loss": 0.6093, | |
| "step": 1800 | |
| }, | |
| { | |
| "epoch": 4.505632040050062, | |
| "eval_loss": 0.7048168182373047, | |
| "eval_runtime": 62.5823, | |
| "eval_samples_per_second": 21.699, | |
| "eval_steps_per_second": 2.716, | |
| "eval_wer": 0.643436246007698, | |
| "step": 1800 | |
| }, | |
| { | |
| "epoch": 4.755944931163955, | |
| "grad_norm": 0.6349893808364868, | |
| "learning_rate": 0.0002545577085088458, | |
| "loss": 0.6006, | |
| "step": 1900 | |
| }, | |
| { | |
| "epoch": 5.006257822277847, | |
| "grad_norm": 1.0852922201156616, | |
| "learning_rate": 0.0002520303285593934, | |
| "loss": 0.5767, | |
| "step": 2000 | |
| }, | |
| { | |
| "epoch": 5.006257822277847, | |
| "eval_loss": 0.6520542502403259, | |
| "eval_runtime": 61.2377, | |
| "eval_samples_per_second": 22.176, | |
| "eval_steps_per_second": 2.776, | |
| "eval_wer": 0.6474490213741708, | |
| "step": 2000 | |
| }, | |
| { | |
| "epoch": 5.256570713391739, | |
| "grad_norm": 1.3806486129760742, | |
| "learning_rate": 0.000249502948609941, | |
| "loss": 0.5647, | |
| "step": 2100 | |
| }, | |
| { | |
| "epoch": 5.506883604505632, | |
| "grad_norm": 1.1985052824020386, | |
| "learning_rate": 0.0002469755686604886, | |
| "loss": 0.5628, | |
| "step": 2200 | |
| }, | |
| { | |
| "epoch": 5.506883604505632, | |
| "eval_loss": 0.6322329640388489, | |
| "eval_runtime": 61.1301, | |
| "eval_samples_per_second": 22.215, | |
| "eval_steps_per_second": 2.781, | |
| "eval_wer": 0.6017525182212758, | |
| "step": 2200 | |
| }, | |
| { | |
| "epoch": 5.7571964956195245, | |
| "grad_norm": 1.5096967220306396, | |
| "learning_rate": 0.0002444481887110362, | |
| "loss": 0.5281, | |
| "step": 2300 | |
| }, | |
| { | |
| "epoch": 6.007509386733417, | |
| "grad_norm": 1.2591967582702637, | |
| "learning_rate": 0.00024192080876158381, | |
| "loss": 0.5569, | |
| "step": 2400 | |
| }, | |
| { | |
| "epoch": 6.007509386733417, | |
| "eval_loss": 0.6288811564445496, | |
| "eval_runtime": 61.1464, | |
| "eval_samples_per_second": 22.209, | |
| "eval_steps_per_second": 2.78, | |
| "eval_wer": 0.6078126279583982, | |
| "step": 2400 | |
| }, | |
| { | |
| "epoch": 6.257822277847309, | |
| "grad_norm": 1.0905272960662842, | |
| "learning_rate": 0.0002393934288121314, | |
| "loss": 0.538, | |
| "step": 2500 | |
| }, | |
| { | |
| "epoch": 6.508135168961202, | |
| "grad_norm": 0.9892663359642029, | |
| "learning_rate": 0.000236866048862679, | |
| "loss": 0.5156, | |
| "step": 2600 | |
| }, | |
| { | |
| "epoch": 6.508135168961202, | |
| "eval_loss": 0.6504250764846802, | |
| "eval_runtime": 60.9759, | |
| "eval_samples_per_second": 22.271, | |
| "eval_steps_per_second": 2.788, | |
| "eval_wer": 0.6373761362705757, | |
| "step": 2600 | |
| }, | |
| { | |
| "epoch": 6.758448060075094, | |
| "grad_norm": 4.6128106117248535, | |
| "learning_rate": 0.0002343386689132266, | |
| "loss": 0.5022, | |
| "step": 2700 | |
| }, | |
| { | |
| "epoch": 7.008760951188986, | |
| "grad_norm": 1.0073637962341309, | |
| "learning_rate": 0.00023181128896377422, | |
| "loss": 0.5074, | |
| "step": 2800 | |
| }, | |
| { | |
| "epoch": 7.008760951188986, | |
| "eval_loss": 0.663825273513794, | |
| "eval_runtime": 60.8911, | |
| "eval_samples_per_second": 22.302, | |
| "eval_steps_per_second": 2.792, | |
| "eval_wer": 0.6222258619277701, | |
| "step": 2800 | |
| }, | |
| { | |
| "epoch": 7.259073842302879, | |
| "grad_norm": 1.6625531911849976, | |
| "learning_rate": 0.00022928390901432182, | |
| "loss": 0.478, | |
| "step": 2900 | |
| }, | |
| { | |
| "epoch": 7.509386733416771, | |
| "grad_norm": 4.11372184753418, | |
| "learning_rate": 0.0002267565290648694, | |
| "loss": 0.4906, | |
| "step": 3000 | |
| }, | |
| { | |
| "epoch": 7.509386733416771, | |
| "eval_loss": 0.6743665933609009, | |
| "eval_runtime": 61.0318, | |
| "eval_samples_per_second": 22.251, | |
| "eval_steps_per_second": 2.785, | |
| "eval_wer": 0.5884038981246417, | |
| "step": 3000 | |
| }, | |
| { | |
| "epoch": 7.509386733416771, | |
| "step": 3000, | |
| "total_flos": 9.634997827277466e+18, | |
| "train_loss": 0.7149289563496908, | |
| "train_runtime": 5175.4065, | |
| "train_samples_per_second": 37.023, | |
| "train_steps_per_second": 2.313 | |
| } | |
| ], | |
| "logging_steps": 100, | |
| "max_steps": 11970, | |
| "num_input_tokens_seen": 0, | |
| "num_train_epochs": 30, | |
| "save_steps": 200, | |
| "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": 9.634997827277466e+18, | |
| "train_batch_size": 8, | |
| "trial_name": null, | |
| "trial_params": null | |
| } | |