| { | |
| "best_metric": 0.2620340585708618, | |
| "best_model_checkpoint": "/scratch/skscla001/results/w2v-bert-bem-bembaspeech-model/checkpoint-2600", | |
| "epoch": 4.497540407589599, | |
| "eval_steps": 200, | |
| "global_step": 3200, | |
| "is_hyper_param_search": false, | |
| "is_local_process_zero": true, | |
| "is_world_process_zero": true, | |
| "log_history": [ | |
| { | |
| "epoch": 0.14054813773717498, | |
| "grad_norm": 3.1719706058502197, | |
| "learning_rate": 0.00029699999999999996, | |
| "loss": 1.5404, | |
| "step": 100 | |
| }, | |
| { | |
| "epoch": 0.28109627547434995, | |
| "grad_norm": 2.3361048698425293, | |
| "learning_rate": 0.00029860103626943, | |
| "loss": 0.7066, | |
| "step": 200 | |
| }, | |
| { | |
| "epoch": 0.28109627547434995, | |
| "eval_loss": 0.5066322684288025, | |
| "eval_runtime": 58.7073, | |
| "eval_samples_per_second": 25.414, | |
| "eval_steps_per_second": 3.185, | |
| "eval_wer": 0.7647720391037287, | |
| "step": 200 | |
| }, | |
| { | |
| "epoch": 0.42164441321152496, | |
| "grad_norm": 2.280653476715088, | |
| "learning_rate": 0.0002971879415920866, | |
| "loss": 0.5841, | |
| "step": 300 | |
| }, | |
| { | |
| "epoch": 0.5621925509486999, | |
| "grad_norm": 3.3147284984588623, | |
| "learning_rate": 0.00029577484691474326, | |
| "loss": 0.5538, | |
| "step": 400 | |
| }, | |
| { | |
| "epoch": 0.5621925509486999, | |
| "eval_loss": 0.4313148260116577, | |
| "eval_runtime": 57.0018, | |
| "eval_samples_per_second": 26.175, | |
| "eval_steps_per_second": 3.281, | |
| "eval_wer": 0.7232459555324855, | |
| "step": 400 | |
| }, | |
| { | |
| "epoch": 0.7027406886858749, | |
| "grad_norm": 2.074723243713379, | |
| "learning_rate": 0.00029436175223739985, | |
| "loss": 0.5065, | |
| "step": 500 | |
| }, | |
| { | |
| "epoch": 0.8432888264230499, | |
| "grad_norm": 1.6516566276550293, | |
| "learning_rate": 0.0002929486575600565, | |
| "loss": 0.4574, | |
| "step": 600 | |
| }, | |
| { | |
| "epoch": 0.8432888264230499, | |
| "eval_loss": 0.41018426418304443, | |
| "eval_runtime": 57.3352, | |
| "eval_samples_per_second": 26.022, | |
| "eval_steps_per_second": 3.262, | |
| "eval_wer": 0.6956484124924301, | |
| "step": 600 | |
| }, | |
| { | |
| "epoch": 0.9838369641602249, | |
| "grad_norm": 2.077942132949829, | |
| "learning_rate": 0.0002915355628827131, | |
| "loss": 0.4548, | |
| "step": 700 | |
| }, | |
| { | |
| "epoch": 1.1243851018973998, | |
| "grad_norm": 3.208420753479004, | |
| "learning_rate": 0.00029012246820536974, | |
| "loss": 0.4084, | |
| "step": 800 | |
| }, | |
| { | |
| "epoch": 1.1243851018973998, | |
| "eval_loss": 0.3528524935245514, | |
| "eval_runtime": 57.2723, | |
| "eval_samples_per_second": 26.051, | |
| "eval_steps_per_second": 3.265, | |
| "eval_wer": 0.6276494506445194, | |
| "step": 800 | |
| }, | |
| { | |
| "epoch": 1.264933239634575, | |
| "grad_norm": 2.2176339626312256, | |
| "learning_rate": 0.00028870937352802633, | |
| "loss": 0.4086, | |
| "step": 900 | |
| }, | |
| { | |
| "epoch": 1.4054813773717498, | |
| "grad_norm": 1.1724199056625366, | |
| "learning_rate": 0.000287296278850683, | |
| "loss": 0.388, | |
| "step": 1000 | |
| }, | |
| { | |
| "epoch": 1.4054813773717498, | |
| "eval_loss": 0.3004082143306732, | |
| "eval_runtime": 56.7843, | |
| "eval_samples_per_second": 26.275, | |
| "eval_steps_per_second": 3.293, | |
| "eval_wer": 0.5723678518903019, | |
| "step": 1000 | |
| }, | |
| { | |
| "epoch": 1.5460295151089247, | |
| "grad_norm": 1.3748365640640259, | |
| "learning_rate": 0.00028588318417333957, | |
| "loss": 0.3791, | |
| "step": 1100 | |
| }, | |
| { | |
| "epoch": 1.6865776528460998, | |
| "grad_norm": 1.1393488645553589, | |
| "learning_rate": 0.0002844700894959962, | |
| "loss": 0.3803, | |
| "step": 1200 | |
| }, | |
| { | |
| "epoch": 1.6865776528460998, | |
| "eval_loss": 0.3376190662384033, | |
| "eval_runtime": 56.8427, | |
| "eval_samples_per_second": 26.248, | |
| "eval_steps_per_second": 3.29, | |
| "eval_wer": 0.647720391037287, | |
| "step": 1200 | |
| }, | |
| { | |
| "epoch": 1.8271257905832747, | |
| "grad_norm": 2.872753143310547, | |
| "learning_rate": 0.0002830569948186528, | |
| "loss": 0.3644, | |
| "step": 1300 | |
| }, | |
| { | |
| "epoch": 1.9676739283204498, | |
| "grad_norm": 0.8893818855285645, | |
| "learning_rate": 0.00028164390014130946, | |
| "loss": 0.367, | |
| "step": 1400 | |
| }, | |
| { | |
| "epoch": 1.9676739283204498, | |
| "eval_loss": 0.29106405377388, | |
| "eval_runtime": 56.9876, | |
| "eval_samples_per_second": 26.181, | |
| "eval_steps_per_second": 3.281, | |
| "eval_wer": 0.5801539925599101, | |
| "step": 1400 | |
| }, | |
| { | |
| "epoch": 2.1082220660576247, | |
| "grad_norm": 0.9697826504707336, | |
| "learning_rate": 0.00028023080546396605, | |
| "loss": 0.3335, | |
| "step": 1500 | |
| }, | |
| { | |
| "epoch": 2.2487702037947996, | |
| "grad_norm": 1.1089781522750854, | |
| "learning_rate": 0.0002788177107866227, | |
| "loss": 0.3168, | |
| "step": 1600 | |
| }, | |
| { | |
| "epoch": 2.2487702037947996, | |
| "eval_loss": 0.31056511402130127, | |
| "eval_runtime": 57.0385, | |
| "eval_samples_per_second": 26.158, | |
| "eval_steps_per_second": 3.278, | |
| "eval_wer": 0.5724543645644087, | |
| "step": 1600 | |
| }, | |
| { | |
| "epoch": 2.3893183415319745, | |
| "grad_norm": 0.8601323962211609, | |
| "learning_rate": 0.0002774046161092793, | |
| "loss": 0.3268, | |
| "step": 1700 | |
| }, | |
| { | |
| "epoch": 2.52986647926915, | |
| "grad_norm": 0.7583444714546204, | |
| "learning_rate": 0.00027599152143193594, | |
| "loss": 0.3227, | |
| "step": 1800 | |
| }, | |
| { | |
| "epoch": 2.52986647926915, | |
| "eval_loss": 0.26542848348617554, | |
| "eval_runtime": 57.1567, | |
| "eval_samples_per_second": 26.104, | |
| "eval_steps_per_second": 3.272, | |
| "eval_wer": 0.5348213513279696, | |
| "step": 1800 | |
| }, | |
| { | |
| "epoch": 2.6704146170063248, | |
| "grad_norm": 1.2712171077728271, | |
| "learning_rate": 0.00027457842675459253, | |
| "loss": 0.3081, | |
| "step": 1900 | |
| }, | |
| { | |
| "epoch": 2.8109627547434997, | |
| "grad_norm": 0.7406817674636841, | |
| "learning_rate": 0.0002731653320772492, | |
| "loss": 0.3111, | |
| "step": 2000 | |
| }, | |
| { | |
| "epoch": 2.8109627547434997, | |
| "eval_loss": 0.26211297512054443, | |
| "eval_runtime": 56.777, | |
| "eval_samples_per_second": 26.278, | |
| "eval_steps_per_second": 3.294, | |
| "eval_wer": 0.5494419932520114, | |
| "step": 2000 | |
| }, | |
| { | |
| "epoch": 2.9515108924806746, | |
| "grad_norm": 1.053320050239563, | |
| "learning_rate": 0.0002717522373999058, | |
| "loss": 0.3181, | |
| "step": 2100 | |
| }, | |
| { | |
| "epoch": 3.0920590302178494, | |
| "grad_norm": 0.7663154006004333, | |
| "learning_rate": 0.00027033914272256237, | |
| "loss": 0.2823, | |
| "step": 2200 | |
| }, | |
| { | |
| "epoch": 3.0920590302178494, | |
| "eval_loss": 0.2665364444255829, | |
| "eval_runtime": 56.8072, | |
| "eval_samples_per_second": 26.264, | |
| "eval_steps_per_second": 3.292, | |
| "eval_wer": 0.5421749286270439, | |
| "step": 2200 | |
| }, | |
| { | |
| "epoch": 3.232607167955025, | |
| "grad_norm": 0.5670559406280518, | |
| "learning_rate": 0.000268926048045219, | |
| "loss": 0.2593, | |
| "step": 2300 | |
| }, | |
| { | |
| "epoch": 3.3731553056921997, | |
| "grad_norm": 0.800305187702179, | |
| "learning_rate": 0.0002675129533678756, | |
| "loss": 0.2603, | |
| "step": 2400 | |
| }, | |
| { | |
| "epoch": 3.3731553056921997, | |
| "eval_loss": 0.26233434677124023, | |
| "eval_runtime": 56.2939, | |
| "eval_samples_per_second": 26.504, | |
| "eval_steps_per_second": 3.322, | |
| "eval_wer": 0.5174323038325115, | |
| "step": 2400 | |
| }, | |
| { | |
| "epoch": 3.5137034434293746, | |
| "grad_norm": 0.581555187702179, | |
| "learning_rate": 0.00026609985869053225, | |
| "loss": 0.2835, | |
| "step": 2500 | |
| }, | |
| { | |
| "epoch": 3.6542515811665495, | |
| "grad_norm": 0.5515788793563843, | |
| "learning_rate": 0.00026468676401318885, | |
| "loss": 0.2735, | |
| "step": 2600 | |
| }, | |
| { | |
| "epoch": 3.6542515811665495, | |
| "eval_loss": 0.2620340585708618, | |
| "eval_runtime": 57.2283, | |
| "eval_samples_per_second": 26.071, | |
| "eval_steps_per_second": 3.268, | |
| "eval_wer": 0.5353404273726101, | |
| "step": 2600 | |
| }, | |
| { | |
| "epoch": 3.7947997189037244, | |
| "grad_norm": 0.6277577877044678, | |
| "learning_rate": 0.0002632736693358455, | |
| "loss": 0.2806, | |
| "step": 2700 | |
| }, | |
| { | |
| "epoch": 3.9353478566408997, | |
| "grad_norm": 0.5215665698051453, | |
| "learning_rate": 0.0002618605746585021, | |
| "loss": 0.2666, | |
| "step": 2800 | |
| }, | |
| { | |
| "epoch": 3.9353478566408997, | |
| "eval_loss": 0.2752685844898224, | |
| "eval_runtime": 57.0665, | |
| "eval_samples_per_second": 26.145, | |
| "eval_steps_per_second": 3.277, | |
| "eval_wer": 0.5450298468725668, | |
| "step": 2800 | |
| }, | |
| { | |
| "epoch": 4.075895994378074, | |
| "grad_norm": 0.745657205581665, | |
| "learning_rate": 0.0002604474799811587, | |
| "loss": 0.2479, | |
| "step": 2900 | |
| }, | |
| { | |
| "epoch": 4.2164441321152495, | |
| "grad_norm": 0.44574543833732605, | |
| "learning_rate": 0.00025903438530381533, | |
| "loss": 0.2248, | |
| "step": 3000 | |
| }, | |
| { | |
| "epoch": 4.2164441321152495, | |
| "eval_loss": 0.28806060552597046, | |
| "eval_runtime": 56.7213, | |
| "eval_samples_per_second": 26.304, | |
| "eval_steps_per_second": 3.297, | |
| "eval_wer": 0.5817977333679384, | |
| "step": 3000 | |
| }, | |
| { | |
| "epoch": 4.356992269852425, | |
| "grad_norm": 0.690984845161438, | |
| "learning_rate": 0.0002576212906264719, | |
| "loss": 0.2502, | |
| "step": 3100 | |
| }, | |
| { | |
| "epoch": 4.497540407589599, | |
| "grad_norm": 0.47614070773124695, | |
| "learning_rate": 0.00025620819594912857, | |
| "loss": 0.2408, | |
| "step": 3200 | |
| }, | |
| { | |
| "epoch": 4.497540407589599, | |
| "eval_loss": 0.2748269736766815, | |
| "eval_runtime": 57.1007, | |
| "eval_samples_per_second": 26.129, | |
| "eval_steps_per_second": 3.275, | |
| "eval_wer": 0.5323989964529804, | |
| "step": 3200 | |
| }, | |
| { | |
| "epoch": 4.497540407589599, | |
| "step": 3200, | |
| "total_flos": 9.439925733716597e+18, | |
| "train_loss": 0.3939369261264801, | |
| "train_runtime": 5287.5609, | |
| "train_samples_per_second": 64.55, | |
| "train_steps_per_second": 4.034 | |
| } | |
| ], | |
| "logging_steps": 100, | |
| "max_steps": 21330, | |
| "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": 9.439925733716597e+18, | |
| "train_batch_size": 8, | |
| "trial_name": null, | |
| "trial_params": null | |
| } | |