| { | |
| "best_metric": 0.702553391456604, | |
| "best_model_checkpoint": "/scratch/skscla001/results/xls-r-1b-bem-natbed-combined-model/checkpoint-1800", | |
| "epoch": 5.256570713391739, | |
| "eval_steps": 100, | |
| "global_step": 2100, | |
| "is_hyper_param_search": false, | |
| "is_local_process_zero": true, | |
| "is_world_process_zero": true, | |
| "log_history": [ | |
| { | |
| "epoch": 0.2503128911138924, | |
| "grad_norm": 2.982633352279663, | |
| "learning_rate": 0.00029099999999999997, | |
| "loss": 5.0296, | |
| "step": 100 | |
| }, | |
| { | |
| "epoch": 0.2503128911138924, | |
| "eval_loss": 2.807065010070801, | |
| "eval_runtime": 99.4276, | |
| "eval_samples_per_second": 13.658, | |
| "eval_steps_per_second": 1.71, | |
| "eval_wer": 1.0, | |
| "step": 100 | |
| }, | |
| { | |
| "epoch": 0.5006257822277848, | |
| "grad_norm": 2.2083914279937744, | |
| "learning_rate": 0.00029754844144903116, | |
| "loss": 1.5792, | |
| "step": 200 | |
| }, | |
| { | |
| "epoch": 0.5006257822277848, | |
| "eval_loss": 1.1184355020523071, | |
| "eval_runtime": 98.2757, | |
| "eval_samples_per_second": 13.818, | |
| "eval_steps_per_second": 1.73, | |
| "eval_wer": 0.9645076872751063, | |
| "step": 200 | |
| }, | |
| { | |
| "epoch": 0.7509386733416771, | |
| "grad_norm": 3.159029483795166, | |
| "learning_rate": 0.0002950210614995788, | |
| "loss": 1.1422, | |
| "step": 300 | |
| }, | |
| { | |
| "epoch": 0.7509386733416771, | |
| "eval_loss": 1.0389240980148315, | |
| "eval_runtime": 98.789, | |
| "eval_samples_per_second": 13.746, | |
| "eval_steps_per_second": 1.721, | |
| "eval_wer": 0.960582270199542, | |
| "step": 300 | |
| }, | |
| { | |
| "epoch": 1.0012515644555695, | |
| "grad_norm": 6.720183372497559, | |
| "learning_rate": 0.00029249368155012635, | |
| "loss": 0.9883, | |
| "step": 400 | |
| }, | |
| { | |
| "epoch": 1.0012515644555695, | |
| "eval_loss": 1.0493754148483276, | |
| "eval_runtime": 97.5764, | |
| "eval_samples_per_second": 13.917, | |
| "eval_steps_per_second": 1.742, | |
| "eval_wer": 0.9989368662087014, | |
| "step": 400 | |
| }, | |
| { | |
| "epoch": 1.2515644555694618, | |
| "grad_norm": 3.1551387310028076, | |
| "learning_rate": 0.00028996630160067397, | |
| "loss": 0.8999, | |
| "step": 500 | |
| }, | |
| { | |
| "epoch": 1.2515644555694618, | |
| "eval_loss": 0.86915123462677, | |
| "eval_runtime": 98.5884, | |
| "eval_samples_per_second": 13.774, | |
| "eval_steps_per_second": 1.724, | |
| "eval_wer": 0.8682531894013739, | |
| "step": 500 | |
| }, | |
| { | |
| "epoch": 1.5018773466833542, | |
| "grad_norm": 2.8559296131134033, | |
| "learning_rate": 0.00028743892165122154, | |
| "loss": 0.9135, | |
| "step": 600 | |
| }, | |
| { | |
| "epoch": 1.5018773466833542, | |
| "eval_loss": 0.8564479351043701, | |
| "eval_runtime": 98.2671, | |
| "eval_samples_per_second": 13.819, | |
| "eval_steps_per_second": 1.73, | |
| "eval_wer": 0.8429833169774289, | |
| "step": 600 | |
| }, | |
| { | |
| "epoch": 1.7521902377972465, | |
| "grad_norm": 2.4900286197662354, | |
| "learning_rate": 0.00028491154170176916, | |
| "loss": 0.8898, | |
| "step": 700 | |
| }, | |
| { | |
| "epoch": 1.7521902377972465, | |
| "eval_loss": 0.8450906872749329, | |
| "eval_runtime": 98.3583, | |
| "eval_samples_per_second": 13.807, | |
| "eval_steps_per_second": 1.728, | |
| "eval_wer": 0.8522244030094864, | |
| "step": 700 | |
| }, | |
| { | |
| "epoch": 2.002503128911139, | |
| "grad_norm": 2.2157928943634033, | |
| "learning_rate": 0.0002823841617523167, | |
| "loss": 0.9089, | |
| "step": 800 | |
| }, | |
| { | |
| "epoch": 2.002503128911139, | |
| "eval_loss": 0.8857161402702332, | |
| "eval_runtime": 98.023, | |
| "eval_samples_per_second": 13.854, | |
| "eval_steps_per_second": 1.734, | |
| "eval_wer": 0.8484625449787373, | |
| "step": 800 | |
| }, | |
| { | |
| "epoch": 2.252816020025031, | |
| "grad_norm": 10.16332721710205, | |
| "learning_rate": 0.00027985678180286435, | |
| "loss": 0.8292, | |
| "step": 900 | |
| }, | |
| { | |
| "epoch": 2.252816020025031, | |
| "eval_loss": 0.8661824464797974, | |
| "eval_runtime": 98.5047, | |
| "eval_samples_per_second": 13.786, | |
| "eval_steps_per_second": 1.726, | |
| "eval_wer": 0.8580307491004252, | |
| "step": 900 | |
| }, | |
| { | |
| "epoch": 2.5031289111389237, | |
| "grad_norm": 1.2531850337982178, | |
| "learning_rate": 0.00027732940185341197, | |
| "loss": 0.7921, | |
| "step": 1000 | |
| }, | |
| { | |
| "epoch": 2.5031289111389237, | |
| "eval_loss": 0.796419084072113, | |
| "eval_runtime": 98.0484, | |
| "eval_samples_per_second": 13.85, | |
| "eval_steps_per_second": 1.734, | |
| "eval_wer": 0.7968596663395485, | |
| "step": 1000 | |
| }, | |
| { | |
| "epoch": 2.7534418022528158, | |
| "grad_norm": 6.90547513961792, | |
| "learning_rate": 0.00027480202190395954, | |
| "loss": 0.7983, | |
| "step": 1100 | |
| }, | |
| { | |
| "epoch": 2.7534418022528158, | |
| "eval_loss": 0.7895970940589905, | |
| "eval_runtime": 98.3739, | |
| "eval_samples_per_second": 13.804, | |
| "eval_steps_per_second": 1.728, | |
| "eval_wer": 0.7951422963689893, | |
| "step": 1100 | |
| }, | |
| { | |
| "epoch": 3.0037546933667083, | |
| "grad_norm": 1.1230757236480713, | |
| "learning_rate": 0.00027227464195450716, | |
| "loss": 0.7946, | |
| "step": 1200 | |
| }, | |
| { | |
| "epoch": 3.0037546933667083, | |
| "eval_loss": 0.7666952013969421, | |
| "eval_runtime": 97.1256, | |
| "eval_samples_per_second": 13.982, | |
| "eval_steps_per_second": 1.75, | |
| "eval_wer": 0.7946516192345436, | |
| "step": 1200 | |
| }, | |
| { | |
| "epoch": 3.254067584480601, | |
| "grad_norm": 2.657978057861328, | |
| "learning_rate": 0.00026974726200505473, | |
| "loss": 0.7488, | |
| "step": 1300 | |
| }, | |
| { | |
| "epoch": 3.254067584480601, | |
| "eval_loss": 0.8179810047149658, | |
| "eval_runtime": 98.614, | |
| "eval_samples_per_second": 13.771, | |
| "eval_steps_per_second": 1.724, | |
| "eval_wer": 0.8495256787700359, | |
| "step": 1300 | |
| }, | |
| { | |
| "epoch": 3.504380475594493, | |
| "grad_norm": 1.1082779169082642, | |
| "learning_rate": 0.00026721988205560235, | |
| "loss": 0.7428, | |
| "step": 1400 | |
| }, | |
| { | |
| "epoch": 3.504380475594493, | |
| "eval_loss": 0.7548468708992004, | |
| "eval_runtime": 98.2436, | |
| "eval_samples_per_second": 13.823, | |
| "eval_steps_per_second": 1.73, | |
| "eval_wer": 0.7688092901537456, | |
| "step": 1400 | |
| }, | |
| { | |
| "epoch": 3.7546933667083855, | |
| "grad_norm": 0.6124172210693359, | |
| "learning_rate": 0.0002646925021061499, | |
| "loss": 0.7256, | |
| "step": 1500 | |
| }, | |
| { | |
| "epoch": 3.7546933667083855, | |
| "eval_loss": 0.7258334755897522, | |
| "eval_runtime": 98.5216, | |
| "eval_samples_per_second": 13.784, | |
| "eval_steps_per_second": 1.726, | |
| "eval_wer": 0.7595682041216879, | |
| "step": 1500 | |
| }, | |
| { | |
| "epoch": 4.005006257822278, | |
| "grad_norm": 1.6707249879837036, | |
| "learning_rate": 0.00026216512215669754, | |
| "loss": 0.741, | |
| "step": 1600 | |
| }, | |
| { | |
| "epoch": 4.005006257822278, | |
| "eval_loss": 0.7665364146232605, | |
| "eval_runtime": 98.0591, | |
| "eval_samples_per_second": 13.849, | |
| "eval_steps_per_second": 1.734, | |
| "eval_wer": 0.7717533529604187, | |
| "step": 1600 | |
| }, | |
| { | |
| "epoch": 4.25531914893617, | |
| "grad_norm": 1.7184687852859497, | |
| "learning_rate": 0.00025963774220724516, | |
| "loss": 0.6775, | |
| "step": 1700 | |
| }, | |
| { | |
| "epoch": 4.25531914893617, | |
| "eval_loss": 0.7922295928001404, | |
| "eval_runtime": 97.9024, | |
| "eval_samples_per_second": 13.871, | |
| "eval_steps_per_second": 1.736, | |
| "eval_wer": 0.77747791952895, | |
| "step": 1700 | |
| }, | |
| { | |
| "epoch": 4.505632040050062, | |
| "grad_norm": 0.8240795135498047, | |
| "learning_rate": 0.00025711036225779273, | |
| "loss": 0.6795, | |
| "step": 1800 | |
| }, | |
| { | |
| "epoch": 4.505632040050062, | |
| "eval_loss": 0.702553391456604, | |
| "eval_runtime": 96.6836, | |
| "eval_samples_per_second": 14.046, | |
| "eval_steps_per_second": 1.758, | |
| "eval_wer": 0.7512266928361139, | |
| "step": 1800 | |
| }, | |
| { | |
| "epoch": 4.755944931163955, | |
| "grad_norm": 0.8530161380767822, | |
| "learning_rate": 0.00025458298230834035, | |
| "loss": 0.683, | |
| "step": 1900 | |
| }, | |
| { | |
| "epoch": 4.755944931163955, | |
| "eval_loss": 0.7051005363464355, | |
| "eval_runtime": 98.456, | |
| "eval_samples_per_second": 13.793, | |
| "eval_steps_per_second": 1.727, | |
| "eval_wer": 0.72252208047105, | |
| "step": 1900 | |
| }, | |
| { | |
| "epoch": 5.006257822277847, | |
| "grad_norm": 1.205073595046997, | |
| "learning_rate": 0.0002520556023588879, | |
| "loss": 0.6838, | |
| "step": 2000 | |
| }, | |
| { | |
| "epoch": 5.006257822277847, | |
| "eval_loss": 0.7196346521377563, | |
| "eval_runtime": 98.5975, | |
| "eval_samples_per_second": 13.773, | |
| "eval_steps_per_second": 1.724, | |
| "eval_wer": 0.7503271180896304, | |
| "step": 2000 | |
| }, | |
| { | |
| "epoch": 5.256570713391739, | |
| "grad_norm": 0.9071811437606812, | |
| "learning_rate": 0.00024952822240943554, | |
| "loss": 0.6005, | |
| "step": 2100 | |
| }, | |
| { | |
| "epoch": 5.256570713391739, | |
| "eval_loss": 0.7032491564750671, | |
| "eval_runtime": 98.4687, | |
| "eval_samples_per_second": 13.791, | |
| "eval_steps_per_second": 1.726, | |
| "eval_wer": 0.7423945044160942, | |
| "step": 2100 | |
| }, | |
| { | |
| "epoch": 5.256570713391739, | |
| "step": 2100, | |
| "total_flos": 2.148232424796796e+19, | |
| "train_loss": 1.0403752663021997, | |
| "train_runtime": 6352.2325, | |
| "train_samples_per_second": 30.164, | |
| "train_steps_per_second": 1.884 | |
| } | |
| ], | |
| "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": 2 | |
| } | |
| }, | |
| "TrainerControl": { | |
| "args": { | |
| "should_epoch_stop": false, | |
| "should_evaluate": false, | |
| "should_log": false, | |
| "should_save": true, | |
| "should_training_stop": false | |
| }, | |
| "attributes": {} | |
| } | |
| }, | |
| "total_flos": 2.148232424796796e+19, | |
| "train_batch_size": 8, | |
| "trial_name": null, | |
| "trial_params": null | |
| } | |