| { | |
| "best_metric": 0.5884432792663574, | |
| "best_model_checkpoint": "/scratch/skscla001/results/mms-1b-all-bem-natbed-nn-model/checkpoint-2200", | |
| "epoch": 6.320541760722348, | |
| "eval_steps": 100, | |
| "global_step": 2800, | |
| "is_hyper_param_search": false, | |
| "is_local_process_zero": true, | |
| "is_world_process_zero": true, | |
| "log_history": [ | |
| { | |
| "epoch": 0.22573363431151242, | |
| "grad_norm": 1.8729794025421143, | |
| "learning_rate": 0.00028799999999999995, | |
| "loss": 7.9244, | |
| "step": 100 | |
| }, | |
| { | |
| "epoch": 0.22573363431151242, | |
| "eval_loss": 1.35137140750885, | |
| "eval_runtime": 45.1782, | |
| "eval_samples_per_second": 15.671, | |
| "eval_steps_per_second": 1.97, | |
| "eval_wer": 1.0237771739130435, | |
| "step": 100 | |
| }, | |
| { | |
| "epoch": 0.45146726862302483, | |
| "grad_norm": 2.6380980014801025, | |
| "learning_rate": 0.0002978165276724791, | |
| "loss": 1.0236, | |
| "step": 200 | |
| }, | |
| { | |
| "epoch": 0.45146726862302483, | |
| "eval_loss": 0.8354936838150024, | |
| "eval_runtime": 44.7017, | |
| "eval_samples_per_second": 15.838, | |
| "eval_steps_per_second": 1.991, | |
| "eval_wer": 0.6594769021739131, | |
| "step": 200 | |
| }, | |
| { | |
| "epoch": 0.6772009029345373, | |
| "grad_norm": 1.9870244264602661, | |
| "learning_rate": 0.0002955420773313116, | |
| "loss": 0.8005, | |
| "step": 300 | |
| }, | |
| { | |
| "epoch": 0.6772009029345373, | |
| "eval_loss": 0.7836518287658691, | |
| "eval_runtime": 45.0039, | |
| "eval_samples_per_second": 15.732, | |
| "eval_steps_per_second": 1.978, | |
| "eval_wer": 0.6141304347826086, | |
| "step": 300 | |
| }, | |
| { | |
| "epoch": 0.9029345372460497, | |
| "grad_norm": 2.012474536895752, | |
| "learning_rate": 0.00029326762699014404, | |
| "loss": 0.8968, | |
| "step": 400 | |
| }, | |
| { | |
| "epoch": 0.9029345372460497, | |
| "eval_loss": 0.7808529138565063, | |
| "eval_runtime": 44.7704, | |
| "eval_samples_per_second": 15.814, | |
| "eval_steps_per_second": 1.988, | |
| "eval_wer": 0.6042798913043478, | |
| "step": 400 | |
| }, | |
| { | |
| "epoch": 1.1286681715575622, | |
| "grad_norm": 8.587899208068848, | |
| "learning_rate": 0.00029099317664897645, | |
| "loss": 0.8909, | |
| "step": 500 | |
| }, | |
| { | |
| "epoch": 1.1286681715575622, | |
| "eval_loss": 0.7146816849708557, | |
| "eval_runtime": 44.5851, | |
| "eval_samples_per_second": 15.88, | |
| "eval_steps_per_second": 1.996, | |
| "eval_wer": 0.5952785326086957, | |
| "step": 500 | |
| }, | |
| { | |
| "epoch": 1.3544018058690745, | |
| "grad_norm": 8.078141212463379, | |
| "learning_rate": 0.0002887187263078089, | |
| "loss": 0.7983, | |
| "step": 600 | |
| }, | |
| { | |
| "epoch": 1.3544018058690745, | |
| "eval_loss": 0.6989510655403137, | |
| "eval_runtime": 44.4701, | |
| "eval_samples_per_second": 15.921, | |
| "eval_steps_per_second": 2.001, | |
| "eval_wer": 0.5930706521739131, | |
| "step": 600 | |
| }, | |
| { | |
| "epoch": 1.580135440180587, | |
| "grad_norm": 2.5096583366394043, | |
| "learning_rate": 0.00028644427596664137, | |
| "loss": 0.8563, | |
| "step": 700 | |
| }, | |
| { | |
| "epoch": 1.580135440180587, | |
| "eval_loss": 0.6804757118225098, | |
| "eval_runtime": 44.8305, | |
| "eval_samples_per_second": 15.793, | |
| "eval_steps_per_second": 1.985, | |
| "eval_wer": 0.5964673913043478, | |
| "step": 700 | |
| }, | |
| { | |
| "epoch": 1.8058690744920993, | |
| "grad_norm": 2.3668673038482666, | |
| "learning_rate": 0.00028416982562547383, | |
| "loss": 0.7094, | |
| "step": 800 | |
| }, | |
| { | |
| "epoch": 1.8058690744920993, | |
| "eval_loss": 0.6849333047866821, | |
| "eval_runtime": 44.6427, | |
| "eval_samples_per_second": 15.859, | |
| "eval_steps_per_second": 1.994, | |
| "eval_wer": 0.5808423913043478, | |
| "step": 800 | |
| }, | |
| { | |
| "epoch": 2.0316027088036117, | |
| "grad_norm": 1.9050077199935913, | |
| "learning_rate": 0.0002819181197877179, | |
| "loss": 0.7499, | |
| "step": 900 | |
| }, | |
| { | |
| "epoch": 2.0316027088036117, | |
| "eval_loss": 0.6456880569458008, | |
| "eval_runtime": 44.6644, | |
| "eval_samples_per_second": 15.852, | |
| "eval_steps_per_second": 1.993, | |
| "eval_wer": 0.5934103260869565, | |
| "step": 900 | |
| }, | |
| { | |
| "epoch": 2.2573363431151243, | |
| "grad_norm": 2.829195737838745, | |
| "learning_rate": 0.0002796436694465504, | |
| "loss": 0.7722, | |
| "step": 1000 | |
| }, | |
| { | |
| "epoch": 2.2573363431151243, | |
| "eval_loss": 0.6565266251564026, | |
| "eval_runtime": 44.7397, | |
| "eval_samples_per_second": 15.825, | |
| "eval_steps_per_second": 1.989, | |
| "eval_wer": 0.5874660326086957, | |
| "step": 1000 | |
| }, | |
| { | |
| "epoch": 2.4830699774266365, | |
| "grad_norm": 0.8484971523284912, | |
| "learning_rate": 0.00027736921910538284, | |
| "loss": 0.7099, | |
| "step": 1100 | |
| }, | |
| { | |
| "epoch": 2.4830699774266365, | |
| "eval_loss": 0.6419216394424438, | |
| "eval_runtime": 44.7023, | |
| "eval_samples_per_second": 15.838, | |
| "eval_steps_per_second": 1.991, | |
| "eval_wer": 0.5596127717391305, | |
| "step": 1100 | |
| }, | |
| { | |
| "epoch": 2.708803611738149, | |
| "grad_norm": 2.4867944717407227, | |
| "learning_rate": 0.0002750947687642153, | |
| "loss": 0.7416, | |
| "step": 1200 | |
| }, | |
| { | |
| "epoch": 2.708803611738149, | |
| "eval_loss": 0.6195096373558044, | |
| "eval_runtime": 44.7966, | |
| "eval_samples_per_second": 15.805, | |
| "eval_steps_per_second": 1.987, | |
| "eval_wer": 0.561141304347826, | |
| "step": 1200 | |
| }, | |
| { | |
| "epoch": 2.9345372460496613, | |
| "grad_norm": 4.84053373336792, | |
| "learning_rate": 0.00027282031842304776, | |
| "loss": 0.6385, | |
| "step": 1300 | |
| }, | |
| { | |
| "epoch": 2.9345372460496613, | |
| "eval_loss": 0.6227733492851257, | |
| "eval_runtime": 44.9228, | |
| "eval_samples_per_second": 15.76, | |
| "eval_steps_per_second": 1.981, | |
| "eval_wer": 0.5647078804347826, | |
| "step": 1300 | |
| }, | |
| { | |
| "epoch": 3.160270880361174, | |
| "grad_norm": 0.9543440341949463, | |
| "learning_rate": 0.0002705458680818802, | |
| "loss": 0.6436, | |
| "step": 1400 | |
| }, | |
| { | |
| "epoch": 3.160270880361174, | |
| "eval_loss": 0.6184154152870178, | |
| "eval_runtime": 44.7251, | |
| "eval_samples_per_second": 15.83, | |
| "eval_steps_per_second": 1.99, | |
| "eval_wer": 0.5509510869565217, | |
| "step": 1400 | |
| }, | |
| { | |
| "epoch": 3.386004514672686, | |
| "grad_norm": 0.8133373856544495, | |
| "learning_rate": 0.0002682714177407126, | |
| "loss": 0.6795, | |
| "step": 1500 | |
| }, | |
| { | |
| "epoch": 3.386004514672686, | |
| "eval_loss": 0.6156527996063232, | |
| "eval_runtime": 45.3372, | |
| "eval_samples_per_second": 15.616, | |
| "eval_steps_per_second": 1.963, | |
| "eval_wer": 0.553328804347826, | |
| "step": 1500 | |
| }, | |
| { | |
| "epoch": 3.6117381489841986, | |
| "grad_norm": 25.58840560913086, | |
| "learning_rate": 0.0002659969673995451, | |
| "loss": 0.7027, | |
| "step": 1600 | |
| }, | |
| { | |
| "epoch": 3.6117381489841986, | |
| "eval_loss": 0.6343082785606384, | |
| "eval_runtime": 45.529, | |
| "eval_samples_per_second": 15.551, | |
| "eval_steps_per_second": 1.955, | |
| "eval_wer": 0.5426290760869565, | |
| "step": 1600 | |
| }, | |
| { | |
| "epoch": 3.837471783295711, | |
| "grad_norm": 0.6009318828582764, | |
| "learning_rate": 0.00026372251705837754, | |
| "loss": 0.6585, | |
| "step": 1700 | |
| }, | |
| { | |
| "epoch": 3.837471783295711, | |
| "eval_loss": 0.6057115793228149, | |
| "eval_runtime": 44.8336, | |
| "eval_samples_per_second": 15.792, | |
| "eval_steps_per_second": 1.985, | |
| "eval_wer": 0.5427989130434783, | |
| "step": 1700 | |
| }, | |
| { | |
| "epoch": 4.063205417607223, | |
| "grad_norm": 0.929165244102478, | |
| "learning_rate": 0.00026144806671720994, | |
| "loss": 0.6351, | |
| "step": 1800 | |
| }, | |
| { | |
| "epoch": 4.063205417607223, | |
| "eval_loss": 0.6017059683799744, | |
| "eval_runtime": 44.5067, | |
| "eval_samples_per_second": 15.908, | |
| "eval_steps_per_second": 2.0, | |
| "eval_wer": 0.54296875, | |
| "step": 1800 | |
| }, | |
| { | |
| "epoch": 4.288939051918736, | |
| "grad_norm": 1.4761062860488892, | |
| "learning_rate": 0.00025917361637604246, | |
| "loss": 0.6528, | |
| "step": 1900 | |
| }, | |
| { | |
| "epoch": 4.288939051918736, | |
| "eval_loss": 0.6098975539207458, | |
| "eval_runtime": 44.8754, | |
| "eval_samples_per_second": 15.777, | |
| "eval_steps_per_second": 1.983, | |
| "eval_wer": 0.5339673913043478, | |
| "step": 1900 | |
| }, | |
| { | |
| "epoch": 4.514672686230249, | |
| "grad_norm": 1.2957922220230103, | |
| "learning_rate": 0.00025689916603487486, | |
| "loss": 0.6603, | |
| "step": 2000 | |
| }, | |
| { | |
| "epoch": 4.514672686230249, | |
| "eval_loss": 0.621790885925293, | |
| "eval_runtime": 45.2703, | |
| "eval_samples_per_second": 15.639, | |
| "eval_steps_per_second": 1.966, | |
| "eval_wer": 0.5334578804347826, | |
| "step": 2000 | |
| }, | |
| { | |
| "epoch": 4.74040632054176, | |
| "grad_norm": 3.225343942642212, | |
| "learning_rate": 0.0002546247156937073, | |
| "loss": 0.6676, | |
| "step": 2100 | |
| }, | |
| { | |
| "epoch": 4.74040632054176, | |
| "eval_loss": 0.5977216958999634, | |
| "eval_runtime": 44.9573, | |
| "eval_samples_per_second": 15.748, | |
| "eval_steps_per_second": 1.98, | |
| "eval_wer": 0.5322690217391305, | |
| "step": 2100 | |
| }, | |
| { | |
| "epoch": 4.966139954853273, | |
| "grad_norm": 1.4750922918319702, | |
| "learning_rate": 0.0002523502653525398, | |
| "loss": 0.6304, | |
| "step": 2200 | |
| }, | |
| { | |
| "epoch": 4.966139954853273, | |
| "eval_loss": 0.5884432792663574, | |
| "eval_runtime": 45.456, | |
| "eval_samples_per_second": 15.575, | |
| "eval_steps_per_second": 1.958, | |
| "eval_wer": 0.5332880434782609, | |
| "step": 2200 | |
| }, | |
| { | |
| "epoch": 5.191873589164786, | |
| "grad_norm": 0.7652086615562439, | |
| "learning_rate": 0.00025007581501137224, | |
| "loss": 0.5976, | |
| "step": 2300 | |
| }, | |
| { | |
| "epoch": 5.191873589164786, | |
| "eval_loss": 0.5955621600151062, | |
| "eval_runtime": 45.3065, | |
| "eval_samples_per_second": 15.627, | |
| "eval_steps_per_second": 1.964, | |
| "eval_wer": 0.5227581521739131, | |
| "step": 2300 | |
| }, | |
| { | |
| "epoch": 5.417607223476298, | |
| "grad_norm": 3.0203540325164795, | |
| "learning_rate": 0.00024780136467020465, | |
| "loss": 0.6564, | |
| "step": 2400 | |
| }, | |
| { | |
| "epoch": 5.417607223476298, | |
| "eval_loss": 0.5956509709358215, | |
| "eval_runtime": 45.0964, | |
| "eval_samples_per_second": 15.7, | |
| "eval_steps_per_second": 1.974, | |
| "eval_wer": 0.5302309782608695, | |
| "step": 2400 | |
| }, | |
| { | |
| "epoch": 5.643340857787811, | |
| "grad_norm": 1.8780221939086914, | |
| "learning_rate": 0.00024552691432903716, | |
| "loss": 0.6717, | |
| "step": 2500 | |
| }, | |
| { | |
| "epoch": 5.643340857787811, | |
| "eval_loss": 0.5767239332199097, | |
| "eval_runtime": 44.9865, | |
| "eval_samples_per_second": 15.738, | |
| "eval_steps_per_second": 1.978, | |
| "eval_wer": 0.5183423913043478, | |
| "step": 2500 | |
| }, | |
| { | |
| "epoch": 5.8690744920993225, | |
| "grad_norm": 0.8542383909225464, | |
| "learning_rate": 0.00024325246398786956, | |
| "loss": 0.6091, | |
| "step": 2600 | |
| }, | |
| { | |
| "epoch": 5.8690744920993225, | |
| "eval_loss": 0.592084527015686, | |
| "eval_runtime": 44.6533, | |
| "eval_samples_per_second": 15.855, | |
| "eval_steps_per_second": 1.993, | |
| "eval_wer": 0.52734375, | |
| "step": 2600 | |
| }, | |
| { | |
| "epoch": 6.094808126410835, | |
| "grad_norm": 1.4033461809158325, | |
| "learning_rate": 0.00024097801364670205, | |
| "loss": 0.6168, | |
| "step": 2700 | |
| }, | |
| { | |
| "epoch": 6.094808126410835, | |
| "eval_loss": 0.5894186496734619, | |
| "eval_runtime": 45.0539, | |
| "eval_samples_per_second": 15.715, | |
| "eval_steps_per_second": 1.975, | |
| "eval_wer": 0.5275135869565217, | |
| "step": 2700 | |
| }, | |
| { | |
| "epoch": 6.320541760722348, | |
| "grad_norm": 10.08028507232666, | |
| "learning_rate": 0.00023870356330553448, | |
| "loss": 0.6495, | |
| "step": 2800 | |
| }, | |
| { | |
| "epoch": 6.320541760722348, | |
| "eval_loss": 0.6036040782928467, | |
| "eval_runtime": 45.0407, | |
| "eval_samples_per_second": 15.719, | |
| "eval_steps_per_second": 1.976, | |
| "eval_wer": 0.5197010869565217, | |
| "step": 2800 | |
| }, | |
| { | |
| "epoch": 6.320541760722348, | |
| "step": 2800, | |
| "total_flos": 1.3313268682658783e+19, | |
| "train_loss": 0.972996187210083, | |
| "train_runtime": 3933.908, | |
| "train_samples_per_second": 27.011, | |
| "train_steps_per_second": 3.378 | |
| } | |
| ], | |
| "logging_steps": 100, | |
| "max_steps": 13290, | |
| "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": 1.3313268682658783e+19, | |
| "train_batch_size": 8, | |
| "trial_name": null, | |
| "trial_params": null | |
| } | |