{ "best_metric": 0.43458130955696106, "best_model_checkpoint": "ai-light-dance_drums_ft_pretrain_wav2vec2-base-new_onset-idmt-mdb-2/checkpoint-638", "epoch": 99.97777777777777, "global_step": 1100, "is_hyper_param_search": false, "is_local_process_zero": true, "is_world_process_zero": true, "log_history": [ { "epoch": 0.89, "learning_rate": 3.9999999999999996e-05, "loss": 22.7802, "step": 10 }, { "epoch": 0.98, "eval_loss": 60.154850006103516, "eval_runtime": 2.4092, "eval_samples_per_second": 7.056, "eval_steps_per_second": 2.075, "eval_wer": 0.9882005899705014, "step": 11 }, { "epoch": 1.8, "learning_rate": 0.00014, "loss": 13.7635, "step": 20 }, { "epoch": 1.98, "eval_loss": 18.182188034057617, "eval_runtime": 2.5023, "eval_samples_per_second": 6.794, "eval_steps_per_second": 1.998, "eval_wer": 0.9985250737463127, "step": 22 }, { "epoch": 2.71, "learning_rate": 0.00023999999999999998, "loss": 3.4364, "step": 30 }, { "epoch": 2.98, "eval_loss": 1.2338517904281616, "eval_runtime": 2.5133, "eval_samples_per_second": 6.764, "eval_steps_per_second": 1.989, "eval_wer": 0.7315634218289085, "step": 33 }, { "epoch": 3.62, "learning_rate": 0.00029887850467289715, "loss": 1.0479, "step": 40 }, { "epoch": 3.98, "eval_loss": 0.8432507514953613, "eval_runtime": 2.5729, "eval_samples_per_second": 6.607, "eval_steps_per_second": 1.943, "eval_wer": 0.4085545722713864, "step": 44 }, { "epoch": 4.53, "learning_rate": 0.00029607476635514017, "loss": 0.739, "step": 50 }, { "epoch": 4.98, "eval_loss": 0.7656970620155334, "eval_runtime": 2.5779, "eval_samples_per_second": 6.595, "eval_steps_per_second": 1.94, "eval_wer": 0.30973451327433627, "step": 55 }, { "epoch": 5.44, "learning_rate": 0.00029327102803738313, "loss": 0.6492, "step": 60 }, { "epoch": 5.98, "eval_loss": 0.8034406900405884, "eval_runtime": 2.5383, "eval_samples_per_second": 6.697, "eval_steps_per_second": 1.97, "eval_wer": 0.2994100294985251, "step": 66 }, { "epoch": 6.36, "learning_rate": 0.00029046728971962615, "loss": 0.6044, "step": 70 }, { "epoch": 6.98, "eval_loss": 0.6401463150978088, "eval_runtime": 2.5648, "eval_samples_per_second": 6.628, "eval_steps_per_second": 1.949, "eval_wer": 0.3333333333333333, "step": 77 }, { "epoch": 7.27, "learning_rate": 0.0002876635514018691, "loss": 0.5662, "step": 80 }, { "epoch": 7.98, "eval_loss": 0.7298213243484497, "eval_runtime": 2.5349, "eval_samples_per_second": 6.706, "eval_steps_per_second": 1.972, "eval_wer": 0.2610619469026549, "step": 88 }, { "epoch": 8.18, "learning_rate": 0.00028485981308411213, "loss": 0.5321, "step": 90 }, { "epoch": 8.98, "eval_loss": 0.8125590085983276, "eval_runtime": 2.5294, "eval_samples_per_second": 6.721, "eval_steps_per_second": 1.977, "eval_wer": 0.29793510324483774, "step": 99 }, { "epoch": 9.09, "learning_rate": 0.0002820560747663551, "loss": 0.5423, "step": 100 }, { "epoch": 9.98, "learning_rate": 0.0002792523364485981, "loss": 0.5037, "step": 110 }, { "epoch": 9.98, "eval_loss": 0.7135093212127686, "eval_runtime": 2.5473, "eval_samples_per_second": 6.674, "eval_steps_per_second": 1.963, "eval_wer": 0.2994100294985251, "step": 110 }, { "epoch": 10.89, "learning_rate": 0.0002764485981308411, "loss": 0.4823, "step": 120 }, { "epoch": 10.98, "eval_loss": 0.5975809097290039, "eval_runtime": 2.4821, "eval_samples_per_second": 6.849, "eval_steps_per_second": 2.014, "eval_wer": 0.26548672566371684, "step": 121 }, { "epoch": 11.8, "learning_rate": 0.0002736448598130841, "loss": 0.4622, "step": 130 }, { "epoch": 11.98, "eval_loss": 0.6875410079956055, "eval_runtime": 2.4654, "eval_samples_per_second": 6.895, "eval_steps_per_second": 2.028, "eval_wer": 0.2448377581120944, "step": 132 }, { "epoch": 12.71, "learning_rate": 0.00027084112149532705, "loss": 0.4761, "step": 140 }, { "epoch": 12.98, "eval_loss": 0.6402226090431213, "eval_runtime": 2.4808, "eval_samples_per_second": 6.853, "eval_steps_per_second": 2.015, "eval_wer": 0.24631268436578171, "step": 143 }, { "epoch": 13.62, "learning_rate": 0.00026803738317757007, "loss": 0.4296, "step": 150 }, { "epoch": 13.98, "eval_loss": 0.82173752784729, "eval_runtime": 2.4283, "eval_samples_per_second": 7.001, "eval_steps_per_second": 2.059, "eval_wer": 0.2448377581120944, "step": 154 }, { "epoch": 14.53, "learning_rate": 0.0002652336448598131, "loss": 0.4655, "step": 160 }, { "epoch": 14.98, "eval_loss": 0.7825008630752563, "eval_runtime": 2.4304, "eval_samples_per_second": 6.995, "eval_steps_per_second": 2.057, "eval_wer": 0.2551622418879056, "step": 165 }, { "epoch": 15.44, "learning_rate": 0.00026242990654205605, "loss": 0.4122, "step": 170 }, { "epoch": 15.98, "eval_loss": 0.712116539478302, "eval_runtime": 2.4336, "eval_samples_per_second": 6.986, "eval_steps_per_second": 2.055, "eval_wer": 0.2448377581120944, "step": 176 }, { "epoch": 16.36, "learning_rate": 0.00025962616822429907, "loss": 0.4234, "step": 180 }, { "epoch": 16.98, "eval_loss": 0.8301025032997131, "eval_runtime": 2.4483, "eval_samples_per_second": 6.944, "eval_steps_per_second": 2.042, "eval_wer": 0.2669616519174041, "step": 187 }, { "epoch": 17.27, "learning_rate": 0.00025682242990654204, "loss": 0.441, "step": 190 }, { "epoch": 17.98, "eval_loss": 0.7342628240585327, "eval_runtime": 2.4341, "eval_samples_per_second": 6.984, "eval_steps_per_second": 2.054, "eval_wer": 0.2640117994100295, "step": 198 }, { "epoch": 18.18, "learning_rate": 0.00025401869158878505, "loss": 0.4781, "step": 200 }, { "epoch": 18.98, "eval_loss": 0.7387515902519226, "eval_runtime": 2.4267, "eval_samples_per_second": 7.005, "eval_steps_per_second": 2.06, "eval_wer": 0.21386430678466076, "step": 209 }, { "epoch": 19.09, "learning_rate": 0.000251214953271028, "loss": 0.3873, "step": 210 }, { "epoch": 19.98, "learning_rate": 0.00024841121495327103, "loss": 0.4006, "step": 220 }, { "epoch": 19.98, "eval_loss": 0.6700401306152344, "eval_runtime": 2.427, "eval_samples_per_second": 7.005, "eval_steps_per_second": 2.06, "eval_wer": 0.252212389380531, "step": 220 }, { "epoch": 20.89, "learning_rate": 0.000245607476635514, "loss": 0.42, "step": 230 }, { "epoch": 20.98, "eval_loss": 0.5540403127670288, "eval_runtime": 2.4476, "eval_samples_per_second": 6.946, "eval_steps_per_second": 2.043, "eval_wer": 0.24926253687315633, "step": 231 }, { "epoch": 21.8, "learning_rate": 0.000242803738317757, "loss": 0.4289, "step": 240 }, { "epoch": 21.98, "eval_loss": 0.9949514865875244, "eval_runtime": 2.4445, "eval_samples_per_second": 6.954, "eval_steps_per_second": 2.045, "eval_wer": 0.24926253687315633, "step": 242 }, { "epoch": 22.71, "learning_rate": 0.00023999999999999998, "loss": 0.4014, "step": 250 }, { "epoch": 22.98, "eval_loss": 0.7282938957214355, "eval_runtime": 2.4265, "eval_samples_per_second": 7.006, "eval_steps_per_second": 2.061, "eval_wer": 0.252212389380531, "step": 253 }, { "epoch": 23.62, "learning_rate": 0.00023719626168224297, "loss": 0.3397, "step": 260 }, { "epoch": 23.98, "eval_loss": 0.8326579928398132, "eval_runtime": 2.4497, "eval_samples_per_second": 6.94, "eval_steps_per_second": 2.041, "eval_wer": 0.26548672566371684, "step": 264 }, { "epoch": 24.53, "learning_rate": 0.00023439252336448596, "loss": 0.3879, "step": 270 }, { "epoch": 24.98, "eval_loss": 0.9388102889060974, "eval_runtime": 2.5119, "eval_samples_per_second": 6.768, "eval_steps_per_second": 1.991, "eval_wer": 0.29056047197640117, "step": 275 }, { "epoch": 25.44, "learning_rate": 0.00023158878504672895, "loss": 0.3445, "step": 280 }, { "epoch": 25.98, "eval_loss": 0.7623295783996582, "eval_runtime": 2.4514, "eval_samples_per_second": 6.935, "eval_steps_per_second": 2.04, "eval_wer": 0.252212389380531, "step": 286 }, { "epoch": 26.36, "learning_rate": 0.00022878504672897194, "loss": 0.3933, "step": 290 }, { "epoch": 26.98, "eval_loss": 0.9125231504440308, "eval_runtime": 2.4489, "eval_samples_per_second": 6.942, "eval_steps_per_second": 2.042, "eval_wer": 0.24188790560471976, "step": 297 }, { "epoch": 27.27, "learning_rate": 0.00022598130841121493, "loss": 0.3173, "step": 300 }, { "epoch": 27.98, "eval_loss": 0.7446789741516113, "eval_runtime": 2.4479, "eval_samples_per_second": 6.945, "eval_steps_per_second": 2.043, "eval_wer": 0.2448377581120944, "step": 308 }, { "epoch": 28.18, "learning_rate": 0.00022317757009345792, "loss": 0.3734, "step": 310 }, { "epoch": 28.98, "eval_loss": 0.6601439118385315, "eval_runtime": 2.4288, "eval_samples_per_second": 6.999, "eval_steps_per_second": 2.059, "eval_wer": 0.2935103244837758, "step": 319 }, { "epoch": 29.09, "learning_rate": 0.0002203738317757009, "loss": 0.3681, "step": 320 }, { "epoch": 29.98, "learning_rate": 0.0002175700934579439, "loss": 0.3347, "step": 330 }, { "epoch": 29.98, "eval_loss": 0.7021532654762268, "eval_runtime": 2.4714, "eval_samples_per_second": 6.879, "eval_steps_per_second": 2.023, "eval_wer": 0.26991150442477874, "step": 330 }, { "epoch": 30.89, "learning_rate": 0.0002147663551401869, "loss": 0.3564, "step": 340 }, { "epoch": 30.98, "eval_loss": 0.748757004737854, "eval_runtime": 2.5209, "eval_samples_per_second": 6.744, "eval_steps_per_second": 1.983, "eval_wer": 0.2920353982300885, "step": 341 }, { "epoch": 31.8, "learning_rate": 0.00021196261682242989, "loss": 0.3371, "step": 350 }, { "epoch": 31.98, "eval_loss": 0.6413426995277405, "eval_runtime": 2.4848, "eval_samples_per_second": 6.842, "eval_steps_per_second": 2.012, "eval_wer": 0.2581120943952802, "step": 352 }, { "epoch": 32.71, "learning_rate": 0.00020915887850467288, "loss": 0.355, "step": 360 }, { "epoch": 32.98, "eval_loss": 0.5131234526634216, "eval_runtime": 2.4307, "eval_samples_per_second": 6.994, "eval_steps_per_second": 2.057, "eval_wer": 0.2374631268436578, "step": 363 }, { "epoch": 33.62, "learning_rate": 0.00020635514018691587, "loss": 0.3648, "step": 370 }, { "epoch": 33.98, "eval_loss": 0.5807757377624512, "eval_runtime": 2.4713, "eval_samples_per_second": 6.879, "eval_steps_per_second": 2.023, "eval_wer": 0.22861356932153393, "step": 374 }, { "epoch": 34.53, "learning_rate": 0.00020355140186915886, "loss": 0.3209, "step": 380 }, { "epoch": 34.98, "eval_loss": 0.5391702651977539, "eval_runtime": 2.4562, "eval_samples_per_second": 6.921, "eval_steps_per_second": 2.036, "eval_wer": 0.22566371681415928, "step": 385 }, { "epoch": 35.44, "learning_rate": 0.00020074766355140185, "loss": 0.3522, "step": 390 }, { "epoch": 35.98, "eval_loss": 0.4410792887210846, "eval_runtime": 2.4622, "eval_samples_per_second": 6.904, "eval_steps_per_second": 2.031, "eval_wer": 0.22271386430678466, "step": 396 }, { "epoch": 36.36, "learning_rate": 0.00019794392523364484, "loss": 0.3252, "step": 400 }, { "epoch": 36.98, "eval_loss": 0.46930572390556335, "eval_runtime": 2.419, "eval_samples_per_second": 7.028, "eval_steps_per_second": 2.067, "eval_wer": 0.21091445427728614, "step": 407 }, { "epoch": 37.27, "learning_rate": 0.00019514018691588783, "loss": 0.3216, "step": 410 }, { "epoch": 37.98, "eval_loss": 0.4621430039405823, "eval_runtime": 2.4503, "eval_samples_per_second": 6.938, "eval_steps_per_second": 2.041, "eval_wer": 0.20648967551622419, "step": 418 }, { "epoch": 38.18, "learning_rate": 0.00019233644859813082, "loss": 0.3119, "step": 420 }, { "epoch": 38.98, "eval_loss": 0.5093832612037659, "eval_runtime": 2.4699, "eval_samples_per_second": 6.883, "eval_steps_per_second": 2.024, "eval_wer": 0.2168141592920354, "step": 429 }, { "epoch": 39.09, "learning_rate": 0.0001895327102803738, "loss": 0.3266, "step": 430 }, { "epoch": 39.98, "learning_rate": 0.0001867289719626168, "loss": 0.3247, "step": 440 }, { "epoch": 39.98, "eval_loss": 0.4897085130214691, "eval_runtime": 2.4522, "eval_samples_per_second": 6.933, "eval_steps_per_second": 2.039, "eval_wer": 0.23156342182890854, "step": 440 }, { "epoch": 40.89, "learning_rate": 0.0001839252336448598, "loss": 0.3246, "step": 450 }, { "epoch": 40.98, "eval_loss": 0.6471104621887207, "eval_runtime": 2.42, "eval_samples_per_second": 7.025, "eval_steps_per_second": 2.066, "eval_wer": 0.22123893805309736, "step": 451 }, { "epoch": 41.8, "learning_rate": 0.00018112149532710278, "loss": 0.2997, "step": 460 }, { "epoch": 41.98, "eval_loss": 0.5569115877151489, "eval_runtime": 2.4575, "eval_samples_per_second": 6.918, "eval_steps_per_second": 2.035, "eval_wer": 0.2153392330383481, "step": 462 }, { "epoch": 42.71, "learning_rate": 0.00017831775700934577, "loss": 0.2969, "step": 470 }, { "epoch": 42.98, "eval_loss": 0.47663480043411255, "eval_runtime": 2.444, "eval_samples_per_second": 6.956, "eval_steps_per_second": 2.046, "eval_wer": 0.20943952802359883, "step": 473 }, { "epoch": 43.62, "learning_rate": 0.00017551401869158876, "loss": 0.3202, "step": 480 }, { "epoch": 43.98, "eval_loss": 0.4978266954421997, "eval_runtime": 2.4301, "eval_samples_per_second": 6.996, "eval_steps_per_second": 2.058, "eval_wer": 0.23156342182890854, "step": 484 }, { "epoch": 44.53, "learning_rate": 0.00017271028037383175, "loss": 0.3093, "step": 490 }, { "epoch": 44.98, "eval_loss": 0.4776196777820587, "eval_runtime": 2.4293, "eval_samples_per_second": 6.998, "eval_steps_per_second": 2.058, "eval_wer": 0.2182890855457227, "step": 495 }, { "epoch": 45.44, "learning_rate": 0.00016990654205607475, "loss": 0.298, "step": 500 }, { "epoch": 45.98, "eval_loss": 0.5007936954498291, "eval_runtime": 2.452, "eval_samples_per_second": 6.933, "eval_steps_per_second": 2.039, "eval_wer": 0.21976401179941002, "step": 506 }, { "epoch": 46.36, "learning_rate": 0.00016710280373831774, "loss": 0.3151, "step": 510 }, { "epoch": 46.98, "eval_loss": 0.48113784193992615, "eval_runtime": 2.4479, "eval_samples_per_second": 6.945, "eval_steps_per_second": 2.043, "eval_wer": 0.2079646017699115, "step": 517 }, { "epoch": 47.27, "learning_rate": 0.00016429906542056073, "loss": 0.2824, "step": 520 }, { "epoch": 47.98, "eval_loss": 0.5011029839515686, "eval_runtime": 2.4404, "eval_samples_per_second": 6.966, "eval_steps_per_second": 2.049, "eval_wer": 0.20648967551622419, "step": 528 }, { "epoch": 48.18, "learning_rate": 0.00016149532710280372, "loss": 0.3089, "step": 530 }, { "epoch": 48.98, "eval_loss": 0.5131256580352783, "eval_runtime": 2.4575, "eval_samples_per_second": 6.917, "eval_steps_per_second": 2.035, "eval_wer": 0.21386430678466076, "step": 539 }, { "epoch": 49.09, "learning_rate": 0.0001586915887850467, "loss": 0.2795, "step": 540 }, { "epoch": 49.98, "learning_rate": 0.0001558878504672897, "loss": 0.3064, "step": 550 }, { "epoch": 49.98, "eval_loss": 0.47494271397590637, "eval_runtime": 2.4456, "eval_samples_per_second": 6.951, "eval_steps_per_second": 2.045, "eval_wer": 0.22271386430678466, "step": 550 }, { "epoch": 50.89, "learning_rate": 0.0001530841121495327, "loss": 0.2734, "step": 560 }, { "epoch": 50.98, "eval_loss": 0.5396826863288879, "eval_runtime": 2.4349, "eval_samples_per_second": 6.982, "eval_steps_per_second": 2.053, "eval_wer": 0.2079646017699115, "step": 561 }, { "epoch": 51.8, "learning_rate": 0.00015028037383177568, "loss": 0.2911, "step": 570 }, { "epoch": 51.98, "eval_loss": 0.4975402355194092, "eval_runtime": 2.4591, "eval_samples_per_second": 6.913, "eval_steps_per_second": 2.033, "eval_wer": 0.20353982300884957, "step": 572 }, { "epoch": 52.71, "learning_rate": 0.00014747663551401867, "loss": 0.2889, "step": 580 }, { "epoch": 52.98, "eval_loss": 0.46327924728393555, "eval_runtime": 2.4538, "eval_samples_per_second": 6.928, "eval_steps_per_second": 2.038, "eval_wer": 0.2168141592920354, "step": 583 }, { "epoch": 53.62, "learning_rate": 0.00014467289719626166, "loss": 0.2523, "step": 590 }, { "epoch": 53.98, "eval_loss": 0.4589496850967407, "eval_runtime": 2.4386, "eval_samples_per_second": 6.971, "eval_steps_per_second": 2.05, "eval_wer": 0.22418879056047197, "step": 594 }, { "epoch": 54.53, "learning_rate": 0.00014186915887850465, "loss": 0.272, "step": 600 }, { "epoch": 54.98, "eval_loss": 0.4855732023715973, "eval_runtime": 2.4564, "eval_samples_per_second": 6.921, "eval_steps_per_second": 2.035, "eval_wer": 0.21238938053097345, "step": 605 }, { "epoch": 55.44, "learning_rate": 0.00013906542056074764, "loss": 0.2733, "step": 610 }, { "epoch": 55.98, "eval_loss": 0.4474070966243744, "eval_runtime": 2.4335, "eval_samples_per_second": 6.986, "eval_steps_per_second": 2.055, "eval_wer": 0.22418879056047197, "step": 616 }, { "epoch": 56.36, "learning_rate": 0.00013626168224299063, "loss": 0.2856, "step": 620 }, { "epoch": 56.98, "eval_loss": 0.4533959627151489, "eval_runtime": 2.4476, "eval_samples_per_second": 6.946, "eval_steps_per_second": 2.043, "eval_wer": 0.22713864306784662, "step": 627 }, { "epoch": 57.27, "learning_rate": 0.00013345794392523362, "loss": 0.2402, "step": 630 }, { "epoch": 57.98, "eval_loss": 0.43458130955696106, "eval_runtime": 2.4576, "eval_samples_per_second": 6.917, "eval_steps_per_second": 2.034, "eval_wer": 0.22418879056047197, "step": 638 }, { "epoch": 58.18, "learning_rate": 0.00013065420560747661, "loss": 0.2567, "step": 640 }, { "epoch": 58.98, "eval_loss": 0.501366376876831, "eval_runtime": 2.4401, "eval_samples_per_second": 6.967, "eval_steps_per_second": 2.049, "eval_wer": 0.22861356932153393, "step": 649 }, { "epoch": 59.09, "learning_rate": 0.0001278504672897196, "loss": 0.294, "step": 650 }, { "epoch": 59.98, "learning_rate": 0.0001250467289719626, "loss": 0.28, "step": 660 }, { "epoch": 59.98, "eval_loss": 0.44283339381217957, "eval_runtime": 2.4528, "eval_samples_per_second": 6.931, "eval_steps_per_second": 2.038, "eval_wer": 0.2182890855457227, "step": 660 }, { "epoch": 60.89, "learning_rate": 0.00012224299065420559, "loss": 0.2541, "step": 670 }, { "epoch": 60.98, "eval_loss": 0.4875503480434418, "eval_runtime": 2.4588, "eval_samples_per_second": 6.914, "eval_steps_per_second": 2.033, "eval_wer": 0.22271386430678466, "step": 671 }, { "epoch": 61.8, "learning_rate": 0.00011943925233644859, "loss": 0.2544, "step": 680 }, { "epoch": 61.98, "eval_loss": 0.4705180823802948, "eval_runtime": 2.4501, "eval_samples_per_second": 6.939, "eval_steps_per_second": 2.041, "eval_wer": 0.20501474926253688, "step": 682 }, { "epoch": 62.71, "learning_rate": 0.00011663551401869158, "loss": 0.2786, "step": 690 }, { "epoch": 62.98, "eval_loss": 0.44488102197647095, "eval_runtime": 2.4756, "eval_samples_per_second": 6.867, "eval_steps_per_second": 2.02, "eval_wer": 0.20206489675516223, "step": 693 }, { "epoch": 63.62, "learning_rate": 0.00011383177570093457, "loss": 0.2524, "step": 700 }, { "epoch": 63.98, "eval_loss": 0.558536946773529, "eval_runtime": 2.4336, "eval_samples_per_second": 6.986, "eval_steps_per_second": 2.055, "eval_wer": 0.20943952802359883, "step": 704 }, { "epoch": 64.53, "learning_rate": 0.00011102803738317756, "loss": 0.2524, "step": 710 }, { "epoch": 64.98, "eval_loss": 0.5178933143615723, "eval_runtime": 2.4703, "eval_samples_per_second": 6.882, "eval_steps_per_second": 2.024, "eval_wer": 0.21091445427728614, "step": 715 }, { "epoch": 65.44, "learning_rate": 0.00010822429906542055, "loss": 0.2852, "step": 720 }, { "epoch": 65.98, "eval_loss": 0.5062990188598633, "eval_runtime": 2.427, "eval_samples_per_second": 7.004, "eval_steps_per_second": 2.06, "eval_wer": 0.21976401179941002, "step": 726 }, { "epoch": 66.36, "learning_rate": 0.00010542056074766354, "loss": 0.2393, "step": 730 }, { "epoch": 66.98, "eval_loss": 0.4768075942993164, "eval_runtime": 2.4351, "eval_samples_per_second": 6.981, "eval_steps_per_second": 2.053, "eval_wer": 0.19911504424778761, "step": 737 }, { "epoch": 67.27, "learning_rate": 0.00010261682242990653, "loss": 0.2522, "step": 740 }, { "epoch": 67.98, "eval_loss": 0.44732481241226196, "eval_runtime": 2.4715, "eval_samples_per_second": 6.878, "eval_steps_per_second": 2.023, "eval_wer": 0.19321533923303835, "step": 748 }, { "epoch": 68.18, "learning_rate": 9.981308411214952e-05, "loss": 0.2768, "step": 750 }, { "epoch": 68.98, "eval_loss": 0.4713572859764099, "eval_runtime": 2.4588, "eval_samples_per_second": 6.914, "eval_steps_per_second": 2.034, "eval_wer": 0.19911504424778761, "step": 759 }, { "epoch": 69.09, "learning_rate": 9.700934579439252e-05, "loss": 0.2484, "step": 760 }, { "epoch": 69.98, "learning_rate": 9.42056074766355e-05, "loss": 0.2463, "step": 770 }, { "epoch": 69.98, "eval_loss": 0.4947798550128937, "eval_runtime": 2.5096, "eval_samples_per_second": 6.774, "eval_steps_per_second": 1.992, "eval_wer": 0.19469026548672566, "step": 770 }, { "epoch": 70.89, "learning_rate": 9.14018691588785e-05, "loss": 0.2379, "step": 780 }, { "epoch": 70.98, "eval_loss": 0.4977743923664093, "eval_runtime": 2.4614, "eval_samples_per_second": 6.907, "eval_steps_per_second": 2.031, "eval_wer": 0.19321533923303835, "step": 781 }, { "epoch": 71.8, "learning_rate": 8.859813084112149e-05, "loss": 0.2343, "step": 790 }, { "epoch": 71.98, "eval_loss": 0.45256489515304565, "eval_runtime": 2.4455, "eval_samples_per_second": 6.952, "eval_steps_per_second": 2.045, "eval_wer": 0.1902654867256637, "step": 792 }, { "epoch": 72.71, "learning_rate": 8.579439252336448e-05, "loss": 0.3377, "step": 800 }, { "epoch": 72.98, "eval_loss": 0.451813668012619, "eval_runtime": 2.4509, "eval_samples_per_second": 6.936, "eval_steps_per_second": 2.04, "eval_wer": 0.19616519174041297, "step": 803 }, { "epoch": 73.62, "learning_rate": 8.299065420560747e-05, "loss": 0.2683, "step": 810 }, { "epoch": 73.98, "eval_loss": 0.44569432735443115, "eval_runtime": 2.4492, "eval_samples_per_second": 6.941, "eval_steps_per_second": 2.041, "eval_wer": 0.21091445427728614, "step": 814 }, { "epoch": 74.53, "learning_rate": 8.018691588785046e-05, "loss": 0.2371, "step": 820 }, { "epoch": 74.98, "eval_loss": 0.4563811719417572, "eval_runtime": 2.4581, "eval_samples_per_second": 6.916, "eval_steps_per_second": 2.034, "eval_wer": 0.20206489675516223, "step": 825 }, { "epoch": 75.44, "learning_rate": 7.738317757009345e-05, "loss": 0.2438, "step": 830 }, { "epoch": 75.98, "eval_loss": 0.48755523562431335, "eval_runtime": 2.4767, "eval_samples_per_second": 6.864, "eval_steps_per_second": 2.019, "eval_wer": 0.20943952802359883, "step": 836 }, { "epoch": 76.36, "learning_rate": 7.457943925233644e-05, "loss": 0.2408, "step": 840 }, { "epoch": 76.98, "eval_loss": 0.43862077593803406, "eval_runtime": 2.4583, "eval_samples_per_second": 6.915, "eval_steps_per_second": 2.034, "eval_wer": 0.20206489675516223, "step": 847 }, { "epoch": 77.27, "learning_rate": 7.177570093457943e-05, "loss": 0.2323, "step": 850 }, { "epoch": 77.98, "eval_loss": 0.4512799382209778, "eval_runtime": 2.4462, "eval_samples_per_second": 6.95, "eval_steps_per_second": 2.044, "eval_wer": 0.19911504424778761, "step": 858 }, { "epoch": 78.18, "learning_rate": 6.897196261682242e-05, "loss": 0.271, "step": 860 }, { "epoch": 78.98, "eval_loss": 0.487374484539032, "eval_runtime": 2.4852, "eval_samples_per_second": 6.841, "eval_steps_per_second": 2.012, "eval_wer": 0.20206489675516223, "step": 869 }, { "epoch": 79.09, "learning_rate": 6.616822429906541e-05, "loss": 0.2406, "step": 870 }, { "epoch": 79.98, "learning_rate": 6.33644859813084e-05, "loss": 0.229, "step": 880 }, { "epoch": 79.98, "eval_loss": 0.4882209002971649, "eval_runtime": 2.454, "eval_samples_per_second": 6.927, "eval_steps_per_second": 2.037, "eval_wer": 0.20648967551622419, "step": 880 }, { "epoch": 80.89, "learning_rate": 6.05607476635514e-05, "loss": 0.224, "step": 890 }, { "epoch": 80.98, "eval_loss": 0.49814578890800476, "eval_runtime": 2.4571, "eval_samples_per_second": 6.919, "eval_steps_per_second": 2.035, "eval_wer": 0.19911504424778761, "step": 891 }, { "epoch": 81.8, "learning_rate": 5.775700934579439e-05, "loss": 0.2442, "step": 900 }, { "epoch": 81.98, "eval_loss": 0.5448421835899353, "eval_runtime": 2.4539, "eval_samples_per_second": 6.928, "eval_steps_per_second": 2.038, "eval_wer": 0.20206489675516223, "step": 902 }, { "epoch": 82.71, "learning_rate": 5.495327102803738e-05, "loss": 0.2075, "step": 910 }, { "epoch": 82.98, "eval_loss": 0.4626116156578064, "eval_runtime": 2.4651, "eval_samples_per_second": 6.896, "eval_steps_per_second": 2.028, "eval_wer": 0.19911504424778761, "step": 913 }, { "epoch": 83.62, "learning_rate": 5.214953271028037e-05, "loss": 0.2314, "step": 920 }, { "epoch": 83.98, "eval_loss": 0.47063305974006653, "eval_runtime": 2.4549, "eval_samples_per_second": 6.925, "eval_steps_per_second": 2.037, "eval_wer": 0.20648967551622419, "step": 924 }, { "epoch": 84.53, "learning_rate": 4.934579439252336e-05, "loss": 0.2208, "step": 930 }, { "epoch": 84.98, "eval_loss": 0.5073422789573669, "eval_runtime": 2.4571, "eval_samples_per_second": 6.919, "eval_steps_per_second": 2.035, "eval_wer": 0.20353982300884957, "step": 935 }, { "epoch": 85.44, "learning_rate": 4.6542056074766354e-05, "loss": 0.2547, "step": 940 }, { "epoch": 85.98, "eval_loss": 0.48184993863105774, "eval_runtime": 2.4595, "eval_samples_per_second": 6.912, "eval_steps_per_second": 2.033, "eval_wer": 0.19616519174041297, "step": 946 }, { "epoch": 86.36, "learning_rate": 4.3738317757009345e-05, "loss": 0.2895, "step": 950 }, { "epoch": 86.98, "eval_loss": 0.49310269951820374, "eval_runtime": 2.4502, "eval_samples_per_second": 6.938, "eval_steps_per_second": 2.041, "eval_wer": 0.19911504424778761, "step": 957 }, { "epoch": 87.27, "learning_rate": 4.0934579439252335e-05, "loss": 0.1988, "step": 960 }, { "epoch": 87.98, "eval_loss": 0.47016218304634094, "eval_runtime": 2.4764, "eval_samples_per_second": 6.865, "eval_steps_per_second": 2.019, "eval_wer": 0.20058997050147492, "step": 968 }, { "epoch": 88.18, "learning_rate": 3.8130841121495326e-05, "loss": 0.2383, "step": 970 }, { "epoch": 88.98, "eval_loss": 0.4681670665740967, "eval_runtime": 2.4685, "eval_samples_per_second": 6.887, "eval_steps_per_second": 2.026, "eval_wer": 0.19911504424778761, "step": 979 }, { "epoch": 89.09, "learning_rate": 3.5327102803738316e-05, "loss": 0.2235, "step": 980 }, { "epoch": 89.98, "learning_rate": 3.252336448598131e-05, "loss": 0.2332, "step": 990 }, { "epoch": 89.98, "eval_loss": 0.4574868381023407, "eval_runtime": 2.4687, "eval_samples_per_second": 6.886, "eval_steps_per_second": 2.025, "eval_wer": 0.20648967551622419, "step": 990 }, { "epoch": 90.89, "learning_rate": 2.9719626168224294e-05, "loss": 0.1983, "step": 1000 }, { "epoch": 90.98, "eval_loss": 0.47059592604637146, "eval_runtime": 2.441, "eval_samples_per_second": 6.964, "eval_steps_per_second": 2.048, "eval_wer": 0.19911504424778761, "step": 1001 }, { "epoch": 91.8, "learning_rate": 2.6915887850467285e-05, "loss": 0.2182, "step": 1010 }, { "epoch": 91.98, "eval_loss": 0.4756470322608948, "eval_runtime": 2.4523, "eval_samples_per_second": 6.932, "eval_steps_per_second": 2.039, "eval_wer": 0.19911504424778761, "step": 1012 }, { "epoch": 92.71, "learning_rate": 2.411214953271028e-05, "loss": 0.2161, "step": 1020 }, { "epoch": 92.98, "eval_loss": 0.4685567617416382, "eval_runtime": 2.5426, "eval_samples_per_second": 6.686, "eval_steps_per_second": 1.967, "eval_wer": 0.19616519174041297, "step": 1023 }, { "epoch": 93.62, "learning_rate": 2.130841121495327e-05, "loss": 0.2215, "step": 1030 }, { "epoch": 93.98, "eval_loss": 0.4688819944858551, "eval_runtime": 2.4691, "eval_samples_per_second": 6.885, "eval_steps_per_second": 2.025, "eval_wer": 0.19321533923303835, "step": 1034 }, { "epoch": 94.53, "learning_rate": 1.850467289719626e-05, "loss": 0.2223, "step": 1040 }, { "epoch": 94.98, "eval_loss": 0.451424241065979, "eval_runtime": 2.4423, "eval_samples_per_second": 6.961, "eval_steps_per_second": 2.047, "eval_wer": 0.1887905604719764, "step": 1045 }, { "epoch": 95.44, "learning_rate": 1.570093457943925e-05, "loss": 0.2068, "step": 1050 }, { "epoch": 95.98, "eval_loss": 0.44822120666503906, "eval_runtime": 2.455, "eval_samples_per_second": 6.925, "eval_steps_per_second": 2.037, "eval_wer": 0.1887905604719764, "step": 1056 }, { "epoch": 96.36, "learning_rate": 1.2897196261682242e-05, "loss": 0.2046, "step": 1060 }, { "epoch": 96.98, "eval_loss": 0.4480588138103485, "eval_runtime": 2.4461, "eval_samples_per_second": 6.95, "eval_steps_per_second": 2.044, "eval_wer": 0.18584070796460178, "step": 1067 }, { "epoch": 97.27, "learning_rate": 1.0093457943925232e-05, "loss": 0.2411, "step": 1070 }, { "epoch": 97.98, "eval_loss": 0.4532119333744049, "eval_runtime": 2.4326, "eval_samples_per_second": 6.988, "eval_steps_per_second": 2.055, "eval_wer": 0.1902654867256637, "step": 1078 }, { "epoch": 98.18, "learning_rate": 7.289719626168224e-06, "loss": 0.2296, "step": 1080 }, { "epoch": 98.98, "eval_loss": 0.4600735306739807, "eval_runtime": 2.4501, "eval_samples_per_second": 6.939, "eval_steps_per_second": 2.041, "eval_wer": 0.19321533923303835, "step": 1089 }, { "epoch": 99.09, "learning_rate": 4.4859813084112145e-06, "loss": 0.2113, "step": 1090 }, { "epoch": 99.98, "learning_rate": 1.6822429906542053e-06, "loss": 0.2211, "step": 1100 }, { "epoch": 99.98, "eval_loss": 0.462471067905426, "eval_runtime": 2.4339, "eval_samples_per_second": 6.985, "eval_steps_per_second": 2.054, "eval_wer": 0.19469026548672566, "step": 1100 }, { "epoch": 99.98, "step": 1100, "total_flos": 1.8082449930602819e+18, "train_loss": 0.679512545520609, "train_runtime": 4168.4023, "train_samples_per_second": 4.294, "train_steps_per_second": 0.264 } ], "max_steps": 1100, "num_train_epochs": 100, "total_flos": 1.8082449930602819e+18, "trial_name": null, "trial_params": null }