| { | |
| "best_metric": 0.29688435792922974, | |
| "best_model_checkpoint": "/scratch/skscla001/results/w2v-bert-bem-genbed-combined-model/checkpoint-2400", | |
| "epoch": 8.241758241758241, | |
| "eval_steps": 200, | |
| "global_step": 3000, | |
| "is_hyper_param_search": false, | |
| "is_local_process_zero": true, | |
| "is_world_process_zero": true, | |
| "log_history": [ | |
| { | |
| "epoch": 0.27472527472527475, | |
| "grad_norm": 3.8470664024353027, | |
| "learning_rate": 0.00029699999999999996, | |
| "loss": 1.5247, | |
| "step": 100 | |
| }, | |
| { | |
| "epoch": 0.5494505494505495, | |
| "grad_norm": 4.275773048400879, | |
| "learning_rate": 0.00029725508317929756, | |
| "loss": 0.6407, | |
| "step": 200 | |
| }, | |
| { | |
| "epoch": 0.5494505494505495, | |
| "eval_loss": 0.6847326159477234, | |
| "eval_runtime": 81.9887, | |
| "eval_samples_per_second": 23.65, | |
| "eval_steps_per_second": 2.964, | |
| "eval_wer": 0.8380540540540541, | |
| "step": 200 | |
| }, | |
| { | |
| "epoch": 0.8241758241758241, | |
| "grad_norm": 4.533022403717041, | |
| "learning_rate": 0.0002944824399260628, | |
| "loss": 0.5465, | |
| "step": 300 | |
| }, | |
| { | |
| "epoch": 1.098901098901099, | |
| "grad_norm": 1.5235211849212646, | |
| "learning_rate": 0.00029170979667282804, | |
| "loss": 0.458, | |
| "step": 400 | |
| }, | |
| { | |
| "epoch": 1.098901098901099, | |
| "eval_loss": 0.4855729043483734, | |
| "eval_runtime": 81.2075, | |
| "eval_samples_per_second": 23.877, | |
| "eval_steps_per_second": 2.992, | |
| "eval_wer": 0.6787027027027027, | |
| "step": 400 | |
| }, | |
| { | |
| "epoch": 1.3736263736263736, | |
| "grad_norm": 1.3080778121948242, | |
| "learning_rate": 0.00028893715341959334, | |
| "loss": 0.4539, | |
| "step": 500 | |
| }, | |
| { | |
| "epoch": 1.6483516483516483, | |
| "grad_norm": 1.1141129732131958, | |
| "learning_rate": 0.0002861645101663586, | |
| "loss": 0.4014, | |
| "step": 600 | |
| }, | |
| { | |
| "epoch": 1.6483516483516483, | |
| "eval_loss": 0.4310147762298584, | |
| "eval_runtime": 81.4044, | |
| "eval_samples_per_second": 23.819, | |
| "eval_steps_per_second": 2.985, | |
| "eval_wer": 0.6258378378378379, | |
| "step": 600 | |
| }, | |
| { | |
| "epoch": 1.9230769230769231, | |
| "grad_norm": 1.3589787483215332, | |
| "learning_rate": 0.00028339186691312383, | |
| "loss": 0.4043, | |
| "step": 700 | |
| }, | |
| { | |
| "epoch": 2.197802197802198, | |
| "grad_norm": 1.0690594911575317, | |
| "learning_rate": 0.0002806192236598891, | |
| "loss": 0.3523, | |
| "step": 800 | |
| }, | |
| { | |
| "epoch": 2.197802197802198, | |
| "eval_loss": 0.36540287733078003, | |
| "eval_runtime": 81.4173, | |
| "eval_samples_per_second": 23.816, | |
| "eval_steps_per_second": 2.985, | |
| "eval_wer": 0.5421621621621622, | |
| "step": 800 | |
| }, | |
| { | |
| "epoch": 2.4725274725274726, | |
| "grad_norm": 1.0713083744049072, | |
| "learning_rate": 0.0002778465804066543, | |
| "loss": 0.3256, | |
| "step": 900 | |
| }, | |
| { | |
| "epoch": 2.7472527472527473, | |
| "grad_norm": 1.2681094408035278, | |
| "learning_rate": 0.00027507393715341956, | |
| "loss": 0.3298, | |
| "step": 1000 | |
| }, | |
| { | |
| "epoch": 2.7472527472527473, | |
| "eval_loss": 0.3534020483493805, | |
| "eval_runtime": 81.532, | |
| "eval_samples_per_second": 23.782, | |
| "eval_steps_per_second": 2.98, | |
| "eval_wer": 0.5373513513513514, | |
| "step": 1000 | |
| }, | |
| { | |
| "epoch": 3.021978021978022, | |
| "grad_norm": 1.4569008350372314, | |
| "learning_rate": 0.0002723012939001848, | |
| "loss": 0.3323, | |
| "step": 1100 | |
| }, | |
| { | |
| "epoch": 3.2967032967032965, | |
| "grad_norm": 0.8082828521728516, | |
| "learning_rate": 0.00026952865064695005, | |
| "loss": 0.2749, | |
| "step": 1200 | |
| }, | |
| { | |
| "epoch": 3.2967032967032965, | |
| "eval_loss": 0.34022244811058044, | |
| "eval_runtime": 81.3749, | |
| "eval_samples_per_second": 23.828, | |
| "eval_steps_per_second": 2.986, | |
| "eval_wer": 0.5196216216216216, | |
| "step": 1200 | |
| }, | |
| { | |
| "epoch": 3.571428571428571, | |
| "grad_norm": 0.5922394394874573, | |
| "learning_rate": 0.0002667560073937153, | |
| "loss": 0.2681, | |
| "step": 1300 | |
| }, | |
| { | |
| "epoch": 3.8461538461538463, | |
| "grad_norm": 0.7247134447097778, | |
| "learning_rate": 0.00026398336414048054, | |
| "loss": 0.2705, | |
| "step": 1400 | |
| }, | |
| { | |
| "epoch": 3.8461538461538463, | |
| "eval_loss": 0.3283730447292328, | |
| "eval_runtime": 86.9096, | |
| "eval_samples_per_second": 22.311, | |
| "eval_steps_per_second": 2.796, | |
| "eval_wer": 0.5250270270270271, | |
| "step": 1400 | |
| }, | |
| { | |
| "epoch": 4.1208791208791204, | |
| "grad_norm": 0.9202386140823364, | |
| "learning_rate": 0.00026121072088724584, | |
| "loss": 0.2516, | |
| "step": 1500 | |
| }, | |
| { | |
| "epoch": 4.395604395604396, | |
| "grad_norm": 1.350644588470459, | |
| "learning_rate": 0.0002584380776340111, | |
| "loss": 0.249, | |
| "step": 1600 | |
| }, | |
| { | |
| "epoch": 4.395604395604396, | |
| "eval_loss": 0.3498856723308563, | |
| "eval_runtime": 81.8027, | |
| "eval_samples_per_second": 23.703, | |
| "eval_steps_per_second": 2.971, | |
| "eval_wer": 0.5299459459459459, | |
| "step": 1600 | |
| }, | |
| { | |
| "epoch": 4.670329670329671, | |
| "grad_norm": 1.1849825382232666, | |
| "learning_rate": 0.00025566543438077633, | |
| "loss": 0.2415, | |
| "step": 1700 | |
| }, | |
| { | |
| "epoch": 4.945054945054945, | |
| "grad_norm": 1.1748567819595337, | |
| "learning_rate": 0.0002528927911275416, | |
| "loss": 0.2508, | |
| "step": 1800 | |
| }, | |
| { | |
| "epoch": 4.945054945054945, | |
| "eval_loss": 0.35116615891456604, | |
| "eval_runtime": 81.4977, | |
| "eval_samples_per_second": 23.792, | |
| "eval_steps_per_second": 2.982, | |
| "eval_wer": 0.5582162162162162, | |
| "step": 1800 | |
| }, | |
| { | |
| "epoch": 5.21978021978022, | |
| "grad_norm": 0.8483781814575195, | |
| "learning_rate": 0.0002501201478743068, | |
| "loss": 0.2172, | |
| "step": 1900 | |
| }, | |
| { | |
| "epoch": 5.4945054945054945, | |
| "grad_norm": 0.7349032759666443, | |
| "learning_rate": 0.00024734750462107206, | |
| "loss": 0.2081, | |
| "step": 2000 | |
| }, | |
| { | |
| "epoch": 5.4945054945054945, | |
| "eval_loss": 0.3216859996318817, | |
| "eval_runtime": 81.8476, | |
| "eval_samples_per_second": 23.69, | |
| "eval_steps_per_second": 2.969, | |
| "eval_wer": 0.4808108108108108, | |
| "step": 2000 | |
| }, | |
| { | |
| "epoch": 5.769230769230769, | |
| "grad_norm": 0.5661817193031311, | |
| "learning_rate": 0.0002445748613678373, | |
| "loss": 0.2039, | |
| "step": 2100 | |
| }, | |
| { | |
| "epoch": 6.043956043956044, | |
| "grad_norm": 0.5281690359115601, | |
| "learning_rate": 0.00024180221811460255, | |
| "loss": 0.2176, | |
| "step": 2200 | |
| }, | |
| { | |
| "epoch": 6.043956043956044, | |
| "eval_loss": 0.3141148090362549, | |
| "eval_runtime": 81.6241, | |
| "eval_samples_per_second": 23.755, | |
| "eval_steps_per_second": 2.977, | |
| "eval_wer": 0.472, | |
| "step": 2200 | |
| }, | |
| { | |
| "epoch": 6.318681318681318, | |
| "grad_norm": 0.5380920171737671, | |
| "learning_rate": 0.0002390295748613678, | |
| "loss": 0.1719, | |
| "step": 2300 | |
| }, | |
| { | |
| "epoch": 6.593406593406593, | |
| "grad_norm": 0.6829392910003662, | |
| "learning_rate": 0.0002362569316081331, | |
| "loss": 0.1784, | |
| "step": 2400 | |
| }, | |
| { | |
| "epoch": 6.593406593406593, | |
| "eval_loss": 0.29688435792922974, | |
| "eval_runtime": 81.2312, | |
| "eval_samples_per_second": 23.87, | |
| "eval_steps_per_second": 2.991, | |
| "eval_wer": 0.4669189189189189, | |
| "step": 2400 | |
| }, | |
| { | |
| "epoch": 6.868131868131869, | |
| "grad_norm": 0.5835310220718384, | |
| "learning_rate": 0.00023348428835489834, | |
| "loss": 0.1842, | |
| "step": 2500 | |
| }, | |
| { | |
| "epoch": 7.142857142857143, | |
| "grad_norm": 0.5060920715332031, | |
| "learning_rate": 0.00023071164510166358, | |
| "loss": 0.166, | |
| "step": 2600 | |
| }, | |
| { | |
| "epoch": 7.142857142857143, | |
| "eval_loss": 0.3367214798927307, | |
| "eval_runtime": 81.6166, | |
| "eval_samples_per_second": 23.757, | |
| "eval_steps_per_second": 2.977, | |
| "eval_wer": 0.49135135135135133, | |
| "step": 2600 | |
| }, | |
| { | |
| "epoch": 7.417582417582418, | |
| "grad_norm": 0.5145979523658752, | |
| "learning_rate": 0.00022793900184842883, | |
| "loss": 0.1562, | |
| "step": 2700 | |
| }, | |
| { | |
| "epoch": 7.6923076923076925, | |
| "grad_norm": 0.8669461011886597, | |
| "learning_rate": 0.00022516635859519407, | |
| "loss": 0.157, | |
| "step": 2800 | |
| }, | |
| { | |
| "epoch": 7.6923076923076925, | |
| "eval_loss": 0.32062390446662903, | |
| "eval_runtime": 81.6437, | |
| "eval_samples_per_second": 23.75, | |
| "eval_steps_per_second": 2.976, | |
| "eval_wer": 0.49027027027027026, | |
| "step": 2800 | |
| }, | |
| { | |
| "epoch": 7.967032967032967, | |
| "grad_norm": 0.7915688753128052, | |
| "learning_rate": 0.00022239371534195932, | |
| "loss": 0.1602, | |
| "step": 2900 | |
| }, | |
| { | |
| "epoch": 8.241758241758241, | |
| "grad_norm": 0.7078336477279663, | |
| "learning_rate": 0.00021962107208872456, | |
| "loss": 0.1398, | |
| "step": 3000 | |
| }, | |
| { | |
| "epoch": 8.241758241758241, | |
| "eval_loss": 0.32599326968193054, | |
| "eval_runtime": 81.2462, | |
| "eval_samples_per_second": 23.866, | |
| "eval_steps_per_second": 2.991, | |
| "eval_wer": 0.4616756756756757, | |
| "step": 3000 | |
| }, | |
| { | |
| "epoch": 8.241758241758241, | |
| "step": 3000, | |
| "total_flos": 8.892175962497035e+18, | |
| "train_loss": 0.32454829279581704, | |
| "train_runtime": 5336.5113, | |
| "train_samples_per_second": 32.701, | |
| "train_steps_per_second": 2.046 | |
| } | |
| ], | |
| "logging_steps": 100, | |
| "max_steps": 10920, | |
| "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": 8.892175962497035e+18, | |
| "train_batch_size": 8, | |
| "trial_name": null, | |
| "trial_params": null | |
| } | |