| { | |
| "best_metric": 0.38489122639254125, | |
| "best_model_checkpoint": "w2v-bert-2.0-CV_Fleurs-lg-5hrs-v4/checkpoint-17028", | |
| "epoch": 66.0, | |
| "eval_steps": 500, | |
| "global_step": 17028, | |
| "is_hyper_param_search": false, | |
| "is_local_process_zero": true, | |
| "is_world_process_zero": true, | |
| "log_history": [ | |
| { | |
| "epoch": 1.0, | |
| "grad_norm": 11.246596336364746, | |
| "learning_rate": 5.1200000000000004e-05, | |
| "loss": 1.9959, | |
| "step": 258 | |
| }, | |
| { | |
| "epoch": 1.0, | |
| "eval_cer": 0.13379445701894616, | |
| "eval_loss": 0.5475470423698425, | |
| "eval_runtime": 29.6601, | |
| "eval_samples_per_second": 27.849, | |
| "eval_steps_per_second": 6.979, | |
| "eval_wer": 0.6078173559646187, | |
| "step": 258 | |
| }, | |
| { | |
| "epoch": 2.0, | |
| "grad_norm": 1.1338785886764526, | |
| "learning_rate": 9.994466403162056e-05, | |
| "loss": 0.5447, | |
| "step": 516 | |
| }, | |
| { | |
| "epoch": 2.0, | |
| "eval_cer": 0.1309612371199291, | |
| "eval_loss": 0.5649892687797546, | |
| "eval_runtime": 29.1014, | |
| "eval_samples_per_second": 28.384, | |
| "eval_steps_per_second": 7.113, | |
| "eval_wer": 0.5978962467128854, | |
| "step": 516 | |
| }, | |
| { | |
| "epoch": 3.0, | |
| "grad_norm": 5.585188388824463, | |
| "learning_rate": 9.892490118577076e-05, | |
| "loss": 0.4574, | |
| "step": 774 | |
| }, | |
| { | |
| "epoch": 3.0, | |
| "eval_cer": 0.11218917678342487, | |
| "eval_loss": 0.45441317558288574, | |
| "eval_runtime": 29.5398, | |
| "eval_samples_per_second": 27.962, | |
| "eval_steps_per_second": 7.007, | |
| "eval_wer": 0.5168539325842697, | |
| "step": 774 | |
| }, | |
| { | |
| "epoch": 4.0, | |
| "grad_norm": 3.5034706592559814, | |
| "learning_rate": 9.790513833992096e-05, | |
| "loss": 0.3712, | |
| "step": 1032 | |
| }, | |
| { | |
| "epoch": 4.0, | |
| "eval_cer": 0.10859621076623562, | |
| "eval_loss": 0.4465671181678772, | |
| "eval_runtime": 29.2068, | |
| "eval_samples_per_second": 28.281, | |
| "eval_steps_per_second": 7.087, | |
| "eval_wer": 0.5049007889074827, | |
| "step": 1032 | |
| }, | |
| { | |
| "epoch": 5.0, | |
| "grad_norm": 17.74321174621582, | |
| "learning_rate": 9.688932806324111e-05, | |
| "loss": 0.3187, | |
| "step": 1290 | |
| }, | |
| { | |
| "epoch": 5.0, | |
| "eval_cer": 0.10036562781937036, | |
| "eval_loss": 0.44379866123199463, | |
| "eval_runtime": 29.231, | |
| "eval_samples_per_second": 28.258, | |
| "eval_steps_per_second": 7.082, | |
| "eval_wer": 0.4769304327038011, | |
| "step": 1290 | |
| }, | |
| { | |
| "epoch": 6.0, | |
| "grad_norm": 10.85033893585205, | |
| "learning_rate": 9.586956521739131e-05, | |
| "loss": 0.276, | |
| "step": 1548 | |
| }, | |
| { | |
| "epoch": 6.0, | |
| "eval_cer": 0.1056996786907042, | |
| "eval_loss": 0.4188626706600189, | |
| "eval_runtime": 29.2483, | |
| "eval_samples_per_second": 28.241, | |
| "eval_steps_per_second": 7.077, | |
| "eval_wer": 0.4794405928759264, | |
| "step": 1548 | |
| }, | |
| { | |
| "epoch": 7.0, | |
| "grad_norm": 1.633737564086914, | |
| "learning_rate": 9.484980237154151e-05, | |
| "loss": 0.235, | |
| "step": 1806 | |
| }, | |
| { | |
| "epoch": 7.0, | |
| "eval_cer": 0.1004447680400133, | |
| "eval_loss": 0.4695667326450348, | |
| "eval_runtime": 29.6156, | |
| "eval_samples_per_second": 27.891, | |
| "eval_steps_per_second": 6.99, | |
| "eval_wer": 0.4796796557494621, | |
| "step": 1806 | |
| }, | |
| { | |
| "epoch": 8.0, | |
| "grad_norm": 7.645702838897705, | |
| "learning_rate": 9.383003952569171e-05, | |
| "loss": 0.2086, | |
| "step": 2064 | |
| }, | |
| { | |
| "epoch": 8.0, | |
| "eval_cer": 0.09632947656658067, | |
| "eval_loss": 0.4218059182167053, | |
| "eval_runtime": 29.7338, | |
| "eval_samples_per_second": 27.78, | |
| "eval_steps_per_second": 6.962, | |
| "eval_wer": 0.4603155629930672, | |
| "step": 2064 | |
| }, | |
| { | |
| "epoch": 9.0, | |
| "grad_norm": 5.886650562286377, | |
| "learning_rate": 9.281027667984191e-05, | |
| "loss": 0.1883, | |
| "step": 2322 | |
| }, | |
| { | |
| "epoch": 9.0, | |
| "eval_cer": 0.09757989205273904, | |
| "eval_loss": 0.4250072240829468, | |
| "eval_runtime": 29.3849, | |
| "eval_samples_per_second": 28.11, | |
| "eval_steps_per_second": 7.044, | |
| "eval_wer": 0.45206789385608415, | |
| "step": 2322 | |
| }, | |
| { | |
| "epoch": 10.0, | |
| "grad_norm": 1.1357783079147339, | |
| "learning_rate": 9.179051383399209e-05, | |
| "loss": 0.1671, | |
| "step": 2580 | |
| }, | |
| { | |
| "epoch": 10.0, | |
| "eval_cer": 0.0972633311701673, | |
| "eval_loss": 0.45072105526924133, | |
| "eval_runtime": 29.0789, | |
| "eval_samples_per_second": 28.405, | |
| "eval_steps_per_second": 7.119, | |
| "eval_wer": 0.458642122878317, | |
| "step": 2580 | |
| }, | |
| { | |
| "epoch": 11.0, | |
| "grad_norm": 0.13285361230373383, | |
| "learning_rate": 9.077075098814229e-05, | |
| "loss": 0.1466, | |
| "step": 2838 | |
| }, | |
| { | |
| "epoch": 11.0, | |
| "eval_cer": 0.09696259833172415, | |
| "eval_loss": 0.4737579822540283, | |
| "eval_runtime": 29.2438, | |
| "eval_samples_per_second": 28.245, | |
| "eval_steps_per_second": 7.078, | |
| "eval_wer": 0.4634233803490318, | |
| "step": 2838 | |
| }, | |
| { | |
| "epoch": 12.0, | |
| "grad_norm": 18.049510955810547, | |
| "learning_rate": 8.975098814229249e-05, | |
| "loss": 0.1313, | |
| "step": 3096 | |
| }, | |
| { | |
| "epoch": 12.0, | |
| "eval_cer": 0.09357539688820653, | |
| "eval_loss": 0.46005183458328247, | |
| "eval_runtime": 29.8062, | |
| "eval_samples_per_second": 27.712, | |
| "eval_steps_per_second": 6.945, | |
| "eval_wer": 0.4410710016734401, | |
| "step": 3096 | |
| }, | |
| { | |
| "epoch": 13.0, | |
| "grad_norm": 1.1974396705627441, | |
| "learning_rate": 8.873517786561266e-05, | |
| "loss": 0.1181, | |
| "step": 3354 | |
| }, | |
| { | |
| "epoch": 13.0, | |
| "eval_cer": 0.08908023235568781, | |
| "eval_loss": 0.47361627221107483, | |
| "eval_runtime": 29.1477, | |
| "eval_samples_per_second": 28.338, | |
| "eval_steps_per_second": 7.102, | |
| "eval_wer": 0.4217069089170452, | |
| "step": 3354 | |
| }, | |
| { | |
| "epoch": 14.0, | |
| "grad_norm": 1.3642749786376953, | |
| "learning_rate": 8.771541501976284e-05, | |
| "loss": 0.1026, | |
| "step": 3612 | |
| }, | |
| { | |
| "epoch": 14.0, | |
| "eval_cer": 0.09121701831304706, | |
| "eval_loss": 0.4470472037792206, | |
| "eval_runtime": 29.2097, | |
| "eval_samples_per_second": 28.278, | |
| "eval_steps_per_second": 7.087, | |
| "eval_wer": 0.42780301219220657, | |
| "step": 3612 | |
| }, | |
| { | |
| "epoch": 15.0, | |
| "grad_norm": 50.79623031616211, | |
| "learning_rate": 8.669565217391304e-05, | |
| "loss": 0.092, | |
| "step": 3870 | |
| }, | |
| { | |
| "epoch": 15.0, | |
| "eval_cer": 0.09560138653666567, | |
| "eval_loss": 0.4730050265789032, | |
| "eval_runtime": 29.5306, | |
| "eval_samples_per_second": 27.971, | |
| "eval_steps_per_second": 7.01, | |
| "eval_wer": 0.4619890031078174, | |
| "step": 3870 | |
| }, | |
| { | |
| "epoch": 16.0, | |
| "grad_norm": 0.18782581388950348, | |
| "learning_rate": 8.567588932806324e-05, | |
| "loss": 0.083, | |
| "step": 4128 | |
| }, | |
| { | |
| "epoch": 16.0, | |
| "eval_cer": 0.09370202124123522, | |
| "eval_loss": 0.533858597278595, | |
| "eval_runtime": 30.1467, | |
| "eval_samples_per_second": 27.399, | |
| "eval_steps_per_second": 6.866, | |
| "eval_wer": 0.44824288787951233, | |
| "step": 4128 | |
| }, | |
| { | |
| "epoch": 17.0, | |
| "grad_norm": 0.19636400043964386, | |
| "learning_rate": 8.465612648221344e-05, | |
| "loss": 0.0744, | |
| "step": 4386 | |
| }, | |
| { | |
| "epoch": 17.0, | |
| "eval_cer": 0.09436679909463587, | |
| "eval_loss": 0.485524982213974, | |
| "eval_runtime": 30.0117, | |
| "eval_samples_per_second": 27.523, | |
| "eval_steps_per_second": 6.897, | |
| "eval_wer": 0.4508725794884055, | |
| "step": 4386 | |
| }, | |
| { | |
| "epoch": 18.0, | |
| "grad_norm": 3.5025405883789062, | |
| "learning_rate": 8.363636363636364e-05, | |
| "loss": 0.0697, | |
| "step": 4644 | |
| }, | |
| { | |
| "epoch": 18.0, | |
| "eval_cer": 0.09033064784184619, | |
| "eval_loss": 0.5221356153488159, | |
| "eval_runtime": 29.9898, | |
| "eval_samples_per_second": 27.543, | |
| "eval_steps_per_second": 6.902, | |
| "eval_wer": 0.43748505857040404, | |
| "step": 4644 | |
| }, | |
| { | |
| "epoch": 19.0, | |
| "grad_norm": 1.851936936378479, | |
| "learning_rate": 8.261660079051384e-05, | |
| "loss": 0.0602, | |
| "step": 4902 | |
| }, | |
| { | |
| "epoch": 19.0, | |
| "eval_cer": 0.08944427737064531, | |
| "eval_loss": 0.51482093334198, | |
| "eval_runtime": 29.3263, | |
| "eval_samples_per_second": 28.166, | |
| "eval_steps_per_second": 7.059, | |
| "eval_wer": 0.42708582357159935, | |
| "step": 4902 | |
| }, | |
| { | |
| "epoch": 20.0, | |
| "grad_norm": 0.04486146196722984, | |
| "learning_rate": 8.159683794466404e-05, | |
| "loss": 0.056, | |
| "step": 5160 | |
| }, | |
| { | |
| "epoch": 20.0, | |
| "eval_cer": 0.08983997847385998, | |
| "eval_loss": 0.5517579317092896, | |
| "eval_runtime": 29.8422, | |
| "eval_samples_per_second": 27.679, | |
| "eval_steps_per_second": 6.936, | |
| "eval_wer": 0.4312694238584748, | |
| "step": 5160 | |
| }, | |
| { | |
| "epoch": 21.0, | |
| "grad_norm": 0.09858173131942749, | |
| "learning_rate": 8.057707509881424e-05, | |
| "loss": 0.05, | |
| "step": 5418 | |
| }, | |
| { | |
| "epoch": 21.0, | |
| "eval_cer": 0.09120119026891847, | |
| "eval_loss": 0.5373868346214294, | |
| "eval_runtime": 29.6954, | |
| "eval_samples_per_second": 27.816, | |
| "eval_steps_per_second": 6.971, | |
| "eval_wer": 0.431030360984939, | |
| "step": 5418 | |
| }, | |
| { | |
| "epoch": 22.0, | |
| "grad_norm": 0.7042592167854309, | |
| "learning_rate": 7.955731225296444e-05, | |
| "loss": 0.0464, | |
| "step": 5676 | |
| }, | |
| { | |
| "epoch": 22.0, | |
| "eval_cer": 0.08990329065037433, | |
| "eval_loss": 0.5166798233985901, | |
| "eval_runtime": 29.6186, | |
| "eval_samples_per_second": 27.888, | |
| "eval_steps_per_second": 6.989, | |
| "eval_wer": 0.42648816638776, | |
| "step": 5676 | |
| }, | |
| { | |
| "epoch": 23.0, | |
| "grad_norm": 8.724380493164062, | |
| "learning_rate": 7.853754940711463e-05, | |
| "loss": 0.0453, | |
| "step": 5934 | |
| }, | |
| { | |
| "epoch": 23.0, | |
| "eval_cer": 0.08813054970797259, | |
| "eval_loss": 0.5781718492507935, | |
| "eval_runtime": 29.9397, | |
| "eval_samples_per_second": 27.589, | |
| "eval_steps_per_second": 6.914, | |
| "eval_wer": 0.4226631604111881, | |
| "step": 5934 | |
| }, | |
| { | |
| "epoch": 24.0, | |
| "grad_norm": 0.4805084764957428, | |
| "learning_rate": 7.751778656126482e-05, | |
| "loss": 0.0412, | |
| "step": 6192 | |
| }, | |
| { | |
| "epoch": 24.0, | |
| "eval_cer": 0.09294227512306304, | |
| "eval_loss": 0.5274640321731567, | |
| "eval_runtime": 29.0914, | |
| "eval_samples_per_second": 28.393, | |
| "eval_steps_per_second": 7.116, | |
| "eval_wer": 0.43533349270858235, | |
| "step": 6192 | |
| }, | |
| { | |
| "epoch": 25.0, | |
| "grad_norm": 0.1090458482503891, | |
| "learning_rate": 7.649802371541502e-05, | |
| "loss": 0.0369, | |
| "step": 6450 | |
| }, | |
| { | |
| "epoch": 25.0, | |
| "eval_cer": 0.09186596812231912, | |
| "eval_loss": 0.6112269759178162, | |
| "eval_runtime": 29.3026, | |
| "eval_samples_per_second": 28.189, | |
| "eval_steps_per_second": 7.064, | |
| "eval_wer": 0.4233803490317954, | |
| "step": 6450 | |
| }, | |
| { | |
| "epoch": 26.0, | |
| "grad_norm": 0.0822739452123642, | |
| "learning_rate": 7.547826086956522e-05, | |
| "loss": 0.0339, | |
| "step": 6708 | |
| }, | |
| { | |
| "epoch": 26.0, | |
| "eval_cer": 0.09094794156286108, | |
| "eval_loss": 0.6159156560897827, | |
| "eval_runtime": 29.9499, | |
| "eval_samples_per_second": 27.579, | |
| "eval_steps_per_second": 6.912, | |
| "eval_wer": 0.4164475256992589, | |
| "step": 6708 | |
| }, | |
| { | |
| "epoch": 27.0, | |
| "grad_norm": 0.11275064945220947, | |
| "learning_rate": 7.445849802371542e-05, | |
| "loss": 0.0316, | |
| "step": 6966 | |
| }, | |
| { | |
| "epoch": 27.0, | |
| "eval_cer": 0.08453758369078333, | |
| "eval_loss": 0.5938315987586975, | |
| "eval_runtime": 29.2749, | |
| "eval_samples_per_second": 28.215, | |
| "eval_steps_per_second": 7.071, | |
| "eval_wer": 0.4031795362180253, | |
| "step": 6966 | |
| }, | |
| { | |
| "epoch": 28.0, | |
| "grad_norm": 0.01683381013572216, | |
| "learning_rate": 7.343873517786562e-05, | |
| "loss": 0.0263, | |
| "step": 7224 | |
| }, | |
| { | |
| "epoch": 28.0, | |
| "eval_cer": 0.08710172683961442, | |
| "eval_loss": 0.5883095860481262, | |
| "eval_runtime": 29.7516, | |
| "eval_samples_per_second": 27.763, | |
| "eval_steps_per_second": 6.958, | |
| "eval_wer": 0.4093951709299546, | |
| "step": 7224 | |
| }, | |
| { | |
| "epoch": 29.0, | |
| "grad_norm": 0.0025164771359413862, | |
| "learning_rate": 7.242292490118577e-05, | |
| "loss": 0.0268, | |
| "step": 7482 | |
| }, | |
| { | |
| "epoch": 29.0, | |
| "eval_cer": 0.0871492109720002, | |
| "eval_loss": 0.6013171076774597, | |
| "eval_runtime": 29.0617, | |
| "eval_samples_per_second": 28.422, | |
| "eval_steps_per_second": 7.123, | |
| "eval_wer": 0.41477408558450873, | |
| "step": 7482 | |
| }, | |
| { | |
| "epoch": 30.0, | |
| "grad_norm": 0.012442449107766151, | |
| "learning_rate": 7.140316205533597e-05, | |
| "loss": 0.0269, | |
| "step": 7740 | |
| }, | |
| { | |
| "epoch": 30.0, | |
| "eval_cer": 0.09121701831304706, | |
| "eval_loss": 0.6137306690216064, | |
| "eval_runtime": 29.4462, | |
| "eval_samples_per_second": 28.051, | |
| "eval_steps_per_second": 7.03, | |
| "eval_wer": 0.42182644035381306, | |
| "step": 7740 | |
| }, | |
| { | |
| "epoch": 31.0, | |
| "grad_norm": 2.4109809398651123, | |
| "learning_rate": 7.038339920948617e-05, | |
| "loss": 0.0234, | |
| "step": 7998 | |
| }, | |
| { | |
| "epoch": 31.0, | |
| "eval_cer": 0.08730749141328606, | |
| "eval_loss": 0.5840467214584351, | |
| "eval_runtime": 29.7315, | |
| "eval_samples_per_second": 27.782, | |
| "eval_steps_per_second": 6.962, | |
| "eval_wer": 0.40987329667702604, | |
| "step": 7998 | |
| }, | |
| { | |
| "epoch": 32.0, | |
| "grad_norm": 1.8709559440612793, | |
| "learning_rate": 6.936363636363637e-05, | |
| "loss": 0.0229, | |
| "step": 8256 | |
| }, | |
| { | |
| "epoch": 32.0, | |
| "eval_cer": 0.08613621614777062, | |
| "eval_loss": 0.628635823726654, | |
| "eval_runtime": 29.901, | |
| "eval_samples_per_second": 27.624, | |
| "eval_steps_per_second": 6.923, | |
| "eval_wer": 0.4041357877121683, | |
| "step": 8256 | |
| }, | |
| { | |
| "epoch": 33.0, | |
| "grad_norm": 21.66620445251465, | |
| "learning_rate": 6.834387351778657e-05, | |
| "loss": 0.0205, | |
| "step": 8514 | |
| }, | |
| { | |
| "epoch": 33.0, | |
| "eval_cer": 0.0859146235299704, | |
| "eval_loss": 0.5922638773918152, | |
| "eval_runtime": 29.394, | |
| "eval_samples_per_second": 28.101, | |
| "eval_steps_per_second": 7.042, | |
| "eval_wer": 0.3968443700693282, | |
| "step": 8514 | |
| }, | |
| { | |
| "epoch": 34.0, | |
| "grad_norm": 0.023612063378095627, | |
| "learning_rate": 6.732411067193675e-05, | |
| "loss": 0.0196, | |
| "step": 8772 | |
| }, | |
| { | |
| "epoch": 34.0, | |
| "eval_cer": 0.08950758954715966, | |
| "eval_loss": 0.6187863945960999, | |
| "eval_runtime": 29.6115, | |
| "eval_samples_per_second": 27.895, | |
| "eval_steps_per_second": 6.991, | |
| "eval_wer": 0.4049725077695434, | |
| "step": 8772 | |
| }, | |
| { | |
| "epoch": 35.0, | |
| "grad_norm": 0.011560407467186451, | |
| "learning_rate": 6.630434782608695e-05, | |
| "loss": 0.0191, | |
| "step": 9030 | |
| }, | |
| { | |
| "epoch": 35.0, | |
| "eval_cer": 0.08724417923677172, | |
| "eval_loss": 0.6254743933677673, | |
| "eval_runtime": 30.6227, | |
| "eval_samples_per_second": 26.973, | |
| "eval_steps_per_second": 6.76, | |
| "eval_wer": 0.4148936170212766, | |
| "step": 9030 | |
| }, | |
| { | |
| "epoch": 36.0, | |
| "grad_norm": 0.30145859718322754, | |
| "learning_rate": 6.528458498023716e-05, | |
| "loss": 0.0185, | |
| "step": 9288 | |
| }, | |
| { | |
| "epoch": 36.0, | |
| "eval_cer": 0.08858956298770161, | |
| "eval_loss": 0.5937672853469849, | |
| "eval_runtime": 29.5915, | |
| "eval_samples_per_second": 27.913, | |
| "eval_steps_per_second": 6.995, | |
| "eval_wer": 0.41035142242409756, | |
| "step": 9288 | |
| }, | |
| { | |
| "epoch": 37.0, | |
| "grad_norm": 0.013502486981451511, | |
| "learning_rate": 6.426482213438736e-05, | |
| "loss": 0.0163, | |
| "step": 9546 | |
| }, | |
| { | |
| "epoch": 37.0, | |
| "eval_cer": 0.08642112094208519, | |
| "eval_loss": 0.6004359126091003, | |
| "eval_runtime": 29.3597, | |
| "eval_samples_per_second": 28.134, | |
| "eval_steps_per_second": 7.05, | |
| "eval_wer": 0.4076021993784365, | |
| "step": 9546 | |
| }, | |
| { | |
| "epoch": 38.0, | |
| "grad_norm": 4.638584136962891, | |
| "learning_rate": 6.324505928853756e-05, | |
| "loss": 0.0171, | |
| "step": 9804 | |
| }, | |
| { | |
| "epoch": 38.0, | |
| "eval_cer": 0.08646860507447095, | |
| "eval_loss": 0.6485220193862915, | |
| "eval_runtime": 29.8389, | |
| "eval_samples_per_second": 27.682, | |
| "eval_steps_per_second": 6.937, | |
| "eval_wer": 0.4010279703562037, | |
| "step": 9804 | |
| }, | |
| { | |
| "epoch": 39.0, | |
| "grad_norm": 0.03485848009586334, | |
| "learning_rate": 6.222529644268775e-05, | |
| "loss": 0.0163, | |
| "step": 10062 | |
| }, | |
| { | |
| "epoch": 39.0, | |
| "eval_cer": 0.08507573719115528, | |
| "eval_loss": 0.6360179781913757, | |
| "eval_runtime": 29.5268, | |
| "eval_samples_per_second": 27.975, | |
| "eval_steps_per_second": 7.011, | |
| "eval_wer": 0.40353813052832893, | |
| "step": 10062 | |
| }, | |
| { | |
| "epoch": 40.0, | |
| "grad_norm": 0.004591576289385557, | |
| "learning_rate": 6.120553359683795e-05, | |
| "loss": 0.0144, | |
| "step": 10320 | |
| }, | |
| { | |
| "epoch": 40.0, | |
| "eval_cer": 0.0878773010019152, | |
| "eval_loss": 0.6229643821716309, | |
| "eval_runtime": 29.6603, | |
| "eval_samples_per_second": 27.849, | |
| "eval_steps_per_second": 6.979, | |
| "eval_wer": 0.41071001673440116, | |
| "step": 10320 | |
| }, | |
| { | |
| "epoch": 41.0, | |
| "grad_norm": 0.006813856773078442, | |
| "learning_rate": 6.0185770750988146e-05, | |
| "loss": 0.0133, | |
| "step": 10578 | |
| }, | |
| { | |
| "epoch": 41.0, | |
| "eval_cer": 0.08781398882540084, | |
| "eval_loss": 0.6123417615890503, | |
| "eval_runtime": 29.6582, | |
| "eval_samples_per_second": 27.851, | |
| "eval_steps_per_second": 6.98, | |
| "eval_wer": 0.4066459478842936, | |
| "step": 10578 | |
| }, | |
| { | |
| "epoch": 42.0, | |
| "grad_norm": 0.0054159704595804214, | |
| "learning_rate": 5.9166007905138345e-05, | |
| "loss": 0.0131, | |
| "step": 10836 | |
| }, | |
| { | |
| "epoch": 42.0, | |
| "eval_cer": 0.08721252314851453, | |
| "eval_loss": 0.6532118916511536, | |
| "eval_runtime": 29.9909, | |
| "eval_samples_per_second": 27.542, | |
| "eval_steps_per_second": 6.902, | |
| "eval_wer": 0.40365766196509684, | |
| "step": 10836 | |
| }, | |
| { | |
| "epoch": 43.0, | |
| "grad_norm": 0.07446209341287613, | |
| "learning_rate": 5.814624505928854e-05, | |
| "loss": 0.0122, | |
| "step": 11094 | |
| }, | |
| { | |
| "epoch": 43.0, | |
| "eval_cer": 0.0865160892068567, | |
| "eval_loss": 0.6612946391105652, | |
| "eval_runtime": 29.2687, | |
| "eval_samples_per_second": 28.221, | |
| "eval_steps_per_second": 7.072, | |
| "eval_wer": 0.40640688501075783, | |
| "step": 11094 | |
| }, | |
| { | |
| "epoch": 44.0, | |
| "grad_norm": 5.483983993530273, | |
| "learning_rate": 5.712648221343874e-05, | |
| "loss": 0.0127, | |
| "step": 11352 | |
| }, | |
| { | |
| "epoch": 44.0, | |
| "eval_cer": 0.08488580066161225, | |
| "eval_loss": 0.627931535243988, | |
| "eval_runtime": 30.3672, | |
| "eval_samples_per_second": 27.2, | |
| "eval_steps_per_second": 6.817, | |
| "eval_wer": 0.40234281616065026, | |
| "step": 11352 | |
| }, | |
| { | |
| "epoch": 45.0, | |
| "grad_norm": 0.0012587299570441246, | |
| "learning_rate": 5.6106719367588936e-05, | |
| "loss": 0.0115, | |
| "step": 11610 | |
| }, | |
| { | |
| "epoch": 45.0, | |
| "eval_cer": 0.08453758369078333, | |
| "eval_loss": 0.6949998140335083, | |
| "eval_runtime": 29.5908, | |
| "eval_samples_per_second": 27.914, | |
| "eval_steps_per_second": 6.995, | |
| "eval_wer": 0.3993545302414535, | |
| "step": 11610 | |
| }, | |
| { | |
| "epoch": 46.0, | |
| "grad_norm": 0.0007135935593396425, | |
| "learning_rate": 5.508695652173913e-05, | |
| "loss": 0.01, | |
| "step": 11868 | |
| }, | |
| { | |
| "epoch": 46.0, | |
| "eval_cer": 0.08490162870574083, | |
| "eval_loss": 0.7085043787956238, | |
| "eval_runtime": 29.5898, | |
| "eval_samples_per_second": 27.915, | |
| "eval_steps_per_second": 6.996, | |
| "eval_wer": 0.39242170690891703, | |
| "step": 11868 | |
| }, | |
| { | |
| "epoch": 47.0, | |
| "grad_norm": 0.0029387420509010553, | |
| "learning_rate": 5.406719367588933e-05, | |
| "loss": 0.0093, | |
| "step": 12126 | |
| }, | |
| { | |
| "epoch": 47.0, | |
| "eval_cer": 0.08711755488374301, | |
| "eval_loss": 0.672882080078125, | |
| "eval_runtime": 29.3653, | |
| "eval_samples_per_second": 28.128, | |
| "eval_steps_per_second": 7.049, | |
| "eval_wer": 0.41310064546975855, | |
| "step": 12126 | |
| }, | |
| { | |
| "epoch": 48.0, | |
| "grad_norm": 0.0017440381925553083, | |
| "learning_rate": 5.304743083003953e-05, | |
| "loss": 0.0106, | |
| "step": 12384 | |
| }, | |
| { | |
| "epoch": 48.0, | |
| "eval_cer": 0.08915937257633075, | |
| "eval_loss": 0.6983070373535156, | |
| "eval_runtime": 29.8557, | |
| "eval_samples_per_second": 27.666, | |
| "eval_steps_per_second": 6.933, | |
| "eval_wer": 0.40963423380349034, | |
| "step": 12384 | |
| }, | |
| { | |
| "epoch": 49.0, | |
| "grad_norm": 0.0008035471546463668, | |
| "learning_rate": 5.202766798418972e-05, | |
| "loss": 0.0087, | |
| "step": 12642 | |
| }, | |
| { | |
| "epoch": 49.0, | |
| "eval_cer": 0.08846293863467292, | |
| "eval_loss": 0.678435742855072, | |
| "eval_runtime": 30.1041, | |
| "eval_samples_per_second": 27.438, | |
| "eval_steps_per_second": 6.876, | |
| "eval_wer": 0.4139373655271336, | |
| "step": 12642 | |
| }, | |
| { | |
| "epoch": 50.0, | |
| "grad_norm": 0.002892086748033762, | |
| "learning_rate": 5.100790513833993e-05, | |
| "loss": 0.0093, | |
| "step": 12900 | |
| }, | |
| { | |
| "epoch": 50.0, | |
| "eval_cer": 0.08513904936766964, | |
| "eval_loss": 0.6546027064323425, | |
| "eval_runtime": 29.6129, | |
| "eval_samples_per_second": 27.893, | |
| "eval_steps_per_second": 6.99, | |
| "eval_wer": 0.4024623475974181, | |
| "step": 12900 | |
| }, | |
| { | |
| "epoch": 51.0, | |
| "grad_norm": 0.03883075714111328, | |
| "learning_rate": 4.998814229249012e-05, | |
| "loss": 0.0088, | |
| "step": 13158 | |
| }, | |
| { | |
| "epoch": 51.0, | |
| "eval_cer": 0.08455341173491192, | |
| "eval_loss": 0.6771977543830872, | |
| "eval_runtime": 29.4829, | |
| "eval_samples_per_second": 28.016, | |
| "eval_steps_per_second": 7.021, | |
| "eval_wer": 0.39983265598852497, | |
| "step": 13158 | |
| }, | |
| { | |
| "epoch": 52.0, | |
| "grad_norm": 0.0004216069064568728, | |
| "learning_rate": 4.8972332015810276e-05, | |
| "loss": 0.0086, | |
| "step": 13416 | |
| }, | |
| { | |
| "epoch": 52.0, | |
| "eval_cer": 0.0862786685449279, | |
| "eval_loss": 0.6763046979904175, | |
| "eval_runtime": 29.8208, | |
| "eval_samples_per_second": 27.699, | |
| "eval_steps_per_second": 6.941, | |
| "eval_wer": 0.39911546736791775, | |
| "step": 13416 | |
| }, | |
| { | |
| "epoch": 53.0, | |
| "grad_norm": 0.006624546833336353, | |
| "learning_rate": 4.7952569169960475e-05, | |
| "loss": 0.0075, | |
| "step": 13674 | |
| }, | |
| { | |
| "epoch": 53.0, | |
| "eval_cer": 0.08407857041105431, | |
| "eval_loss": 0.6990380883216858, | |
| "eval_runtime": 29.3013, | |
| "eval_samples_per_second": 28.19, | |
| "eval_steps_per_second": 7.065, | |
| "eval_wer": 0.39517092995457803, | |
| "step": 13674 | |
| }, | |
| { | |
| "epoch": 54.0, | |
| "grad_norm": 0.002384037710726261, | |
| "learning_rate": 4.6932806324110674e-05, | |
| "loss": 0.0062, | |
| "step": 13932 | |
| }, | |
| { | |
| "epoch": 54.0, | |
| "eval_cer": 0.08322385602811061, | |
| "eval_loss": 0.6648118495941162, | |
| "eval_runtime": 30.4578, | |
| "eval_samples_per_second": 27.119, | |
| "eval_steps_per_second": 6.796, | |
| "eval_wer": 0.39361702127659576, | |
| "step": 13932 | |
| }, | |
| { | |
| "epoch": 55.0, | |
| "grad_norm": 0.0031123559456318617, | |
| "learning_rate": 4.591304347826087e-05, | |
| "loss": 0.0072, | |
| "step": 14190 | |
| }, | |
| { | |
| "epoch": 55.0, | |
| "eval_cer": 0.08659522942749964, | |
| "eval_loss": 0.7061810493469238, | |
| "eval_runtime": 29.6721, | |
| "eval_samples_per_second": 27.838, | |
| "eval_steps_per_second": 6.976, | |
| "eval_wer": 0.4115467367917762, | |
| "step": 14190 | |
| }, | |
| { | |
| "epoch": 56.0, | |
| "grad_norm": 0.0030562193132936954, | |
| "learning_rate": 4.4893280632411066e-05, | |
| "loss": 0.0066, | |
| "step": 14448 | |
| }, | |
| { | |
| "epoch": 56.0, | |
| "eval_cer": 0.08675350986878551, | |
| "eval_loss": 0.6818866729736328, | |
| "eval_runtime": 30.078, | |
| "eval_samples_per_second": 27.462, | |
| "eval_steps_per_second": 6.882, | |
| "eval_wer": 0.40437485058570405, | |
| "step": 14448 | |
| }, | |
| { | |
| "epoch": 57.0, | |
| "grad_norm": 0.0037516027223318815, | |
| "learning_rate": 4.387351778656127e-05, | |
| "loss": 0.0053, | |
| "step": 14706 | |
| }, | |
| { | |
| "epoch": 57.0, | |
| "eval_cer": 0.085930451574099, | |
| "eval_loss": 0.70534348487854, | |
| "eval_runtime": 29.3554, | |
| "eval_samples_per_second": 28.138, | |
| "eval_steps_per_second": 7.052, | |
| "eval_wer": 0.40437485058570405, | |
| "step": 14706 | |
| }, | |
| { | |
| "epoch": 58.0, | |
| "grad_norm": 0.0011016507633030415, | |
| "learning_rate": 4.2853754940711464e-05, | |
| "loss": 0.004, | |
| "step": 14964 | |
| }, | |
| { | |
| "epoch": 58.0, | |
| "eval_cer": 0.08333465233701072, | |
| "eval_loss": 0.6890330910682678, | |
| "eval_runtime": 29.1941, | |
| "eval_samples_per_second": 28.293, | |
| "eval_steps_per_second": 7.09, | |
| "eval_wer": 0.3966053071957925, | |
| "step": 14964 | |
| }, | |
| { | |
| "epoch": 59.0, | |
| "grad_norm": 0.0015181864146143198, | |
| "learning_rate": 4.1833992094861664e-05, | |
| "loss": 0.0038, | |
| "step": 15222 | |
| }, | |
| { | |
| "epoch": 59.0, | |
| "eval_cer": 0.08499659697051236, | |
| "eval_loss": 0.7095320820808411, | |
| "eval_runtime": 29.474, | |
| "eval_samples_per_second": 28.025, | |
| "eval_steps_per_second": 7.023, | |
| "eval_wer": 0.40090843891943584, | |
| "step": 15222 | |
| }, | |
| { | |
| "epoch": 60.0, | |
| "grad_norm": 0.0011176789412274957, | |
| "learning_rate": 4.0814229249011856e-05, | |
| "loss": 0.005, | |
| "step": 15480 | |
| }, | |
| { | |
| "epoch": 60.0, | |
| "eval_cer": 0.08502825305876953, | |
| "eval_loss": 0.6998753547668457, | |
| "eval_runtime": 29.5712, | |
| "eval_samples_per_second": 27.933, | |
| "eval_steps_per_second": 7.0, | |
| "eval_wer": 0.39433420989720297, | |
| "step": 15480 | |
| }, | |
| { | |
| "epoch": 61.0, | |
| "grad_norm": 0.00044457780313678086, | |
| "learning_rate": 3.9794466403162055e-05, | |
| "loss": 0.0055, | |
| "step": 15738 | |
| }, | |
| { | |
| "epoch": 61.0, | |
| "eval_cer": 0.08455341173491192, | |
| "eval_loss": 0.7265011072158813, | |
| "eval_runtime": 31.1453, | |
| "eval_samples_per_second": 26.521, | |
| "eval_steps_per_second": 6.646, | |
| "eval_wer": 0.3957685871384174, | |
| "step": 15738 | |
| }, | |
| { | |
| "epoch": 62.0, | |
| "grad_norm": 0.0023276652209460735, | |
| "learning_rate": 3.8774703557312254e-05, | |
| "loss": 0.0043, | |
| "step": 15996 | |
| }, | |
| { | |
| "epoch": 62.0, | |
| "eval_cer": 0.08366704126371104, | |
| "eval_loss": 0.7266702055931091, | |
| "eval_runtime": 29.5135, | |
| "eval_samples_per_second": 27.987, | |
| "eval_steps_per_second": 7.014, | |
| "eval_wer": 0.3926607697824528, | |
| "step": 15996 | |
| }, | |
| { | |
| "epoch": 63.0, | |
| "grad_norm": 0.011850759387016296, | |
| "learning_rate": 3.775494071146245e-05, | |
| "loss": 0.0038, | |
| "step": 16254 | |
| }, | |
| { | |
| "epoch": 63.0, | |
| "eval_cer": 0.08369869735196822, | |
| "eval_loss": 0.7014022469520569, | |
| "eval_runtime": 29.8354, | |
| "eval_samples_per_second": 27.685, | |
| "eval_steps_per_second": 6.938, | |
| "eval_wer": 0.38680372938082713, | |
| "step": 16254 | |
| }, | |
| { | |
| "epoch": 64.0, | |
| "grad_norm": 2.36328125, | |
| "learning_rate": 3.673913043478261e-05, | |
| "loss": 0.0026, | |
| "step": 16512 | |
| }, | |
| { | |
| "epoch": 64.0, | |
| "eval_cer": 0.08346127669003942, | |
| "eval_loss": 0.7608654499053955, | |
| "eval_runtime": 30.3452, | |
| "eval_samples_per_second": 27.22, | |
| "eval_steps_per_second": 6.822, | |
| "eval_wer": 0.3909873296677026, | |
| "step": 16512 | |
| }, | |
| { | |
| "epoch": 65.0, | |
| "grad_norm": 0.000907552894204855, | |
| "learning_rate": 3.5719367588932805e-05, | |
| "loss": 0.0024, | |
| "step": 16770 | |
| }, | |
| { | |
| "epoch": 65.0, | |
| "eval_cer": 0.08748159989870052, | |
| "eval_loss": 0.7435780167579651, | |
| "eval_runtime": 30.3356, | |
| "eval_samples_per_second": 27.229, | |
| "eval_steps_per_second": 6.824, | |
| "eval_wer": 0.4052115706430791, | |
| "step": 16770 | |
| }, | |
| { | |
| "epoch": 66.0, | |
| "grad_norm": 0.0005879841046407819, | |
| "learning_rate": 3.4699604743083004e-05, | |
| "loss": 0.0044, | |
| "step": 17028 | |
| }, | |
| { | |
| "epoch": 66.0, | |
| "eval_cer": 0.082749014704253, | |
| "eval_loss": 0.7609582543373108, | |
| "eval_runtime": 29.8655, | |
| "eval_samples_per_second": 27.657, | |
| "eval_steps_per_second": 6.931, | |
| "eval_wer": 0.38489122639254125, | |
| "step": 17028 | |
| } | |
| ], | |
| "logging_steps": 500, | |
| "max_steps": 25800, | |
| "num_input_tokens_seen": 0, | |
| "num_train_epochs": 100, | |
| "save_steps": 500, | |
| "stateful_callbacks": { | |
| "EarlyStoppingCallback": { | |
| "args": { | |
| "early_stopping_patience": 20, | |
| "early_stopping_threshold": 0.001 | |
| }, | |
| "attributes": { | |
| "early_stopping_patience_counter": 0 | |
| } | |
| }, | |
| "TrainerControl": { | |
| "args": { | |
| "should_epoch_stop": false, | |
| "should_evaluate": false, | |
| "should_log": false, | |
| "should_save": true, | |
| "should_training_stop": false | |
| }, | |
| "attributes": {} | |
| } | |
| }, | |
| "total_flos": 3.605498760517867e+19, | |
| "train_batch_size": 8, | |
| "trial_name": null, | |
| "trial_params": null | |
| } | |