{ "best_metric": 0.4694591760635376, "best_model_checkpoint": "ai-light-dance_singing3_ft_wav2vec2-large-xlsr-53-v1-5gram/checkpoint-4464", "epoch": 99.99653979238754, "global_step": 7200, "is_hyper_param_search": false, "is_local_process_zero": true, "is_world_process_zero": true, "log_history": [ { "epoch": 0.14, "learning_rate": 5e-06, "loss": 0.3608, "step": 10 }, { "epoch": 0.28, "learning_rate": 1e-05, "loss": 0.3819, "step": 20 }, { "epoch": 0.42, "learning_rate": 1.5e-05, "loss": 0.3643, "step": 30 }, { "epoch": 0.55, "learning_rate": 2e-05, "loss": 0.3776, "step": 40 }, { "epoch": 0.69, "learning_rate": 2.5e-05, "loss": 0.3628, "step": 50 }, { "epoch": 0.83, "learning_rate": 3e-05, "loss": 0.3659, "step": 60 }, { "epoch": 0.97, "learning_rate": 3.5e-05, "loss": 0.3993, "step": 70 }, { "epoch": 1.0, "eval_loss": 0.5761275887489319, "eval_runtime": 129.3716, "eval_samples_per_second": 4.112, "eval_steps_per_second": 0.518, "eval_wer": 0.26778941955389857, "step": 72 }, { "epoch": 1.11, "learning_rate": 4e-05, "loss": 0.3983, "step": 80 }, { "epoch": 1.25, "learning_rate": 4.5e-05, "loss": 0.3997, "step": 90 }, { "epoch": 1.39, "learning_rate": 5e-05, "loss": 0.3924, "step": 100 }, { "epoch": 1.53, "learning_rate": 4.992957746478874e-05, "loss": 0.4145, "step": 110 }, { "epoch": 1.66, "learning_rate": 4.9859154929577466e-05, "loss": 0.4261, "step": 120 }, { "epoch": 1.8, "learning_rate": 4.97887323943662e-05, "loss": 0.458, "step": 130 }, { "epoch": 1.94, "learning_rate": 4.972535211267606e-05, "loss": 0.4252, "step": 140 }, { "epoch": 2.0, "eval_loss": 0.5669388175010681, "eval_runtime": 129.6387, "eval_samples_per_second": 4.104, "eval_steps_per_second": 0.517, "eval_wer": 0.2987079771164106, "step": 144 }, { "epoch": 2.08, "learning_rate": 4.965492957746479e-05, "loss": 0.4345, "step": 150 }, { "epoch": 2.22, "learning_rate": 4.9584507042253526e-05, "loss": 0.4184, "step": 160 }, { "epoch": 2.36, "learning_rate": 4.951408450704226e-05, "loss": 0.4452, "step": 170 }, { "epoch": 2.5, "learning_rate": 4.944366197183099e-05, "loss": 0.4376, "step": 180 }, { "epoch": 2.64, "learning_rate": 4.9373239436619725e-05, "loss": 0.4491, "step": 190 }, { "epoch": 2.78, "learning_rate": 4.930281690140845e-05, "loss": 0.4416, "step": 200 }, { "epoch": 2.91, "learning_rate": 4.923239436619719e-05, "loss": 0.4309, "step": 210 }, { "epoch": 3.0, "eval_loss": 0.553219199180603, "eval_runtime": 128.929, "eval_samples_per_second": 4.126, "eval_steps_per_second": 0.52, "eval_wer": 0.2913800861348589, "step": 216 }, { "epoch": 3.06, "learning_rate": 4.916197183098592e-05, "loss": 0.4487, "step": 220 }, { "epoch": 3.19, "learning_rate": 4.909154929577465e-05, "loss": 0.4235, "step": 230 }, { "epoch": 3.33, "learning_rate": 4.902112676056339e-05, "loss": 0.4433, "step": 240 }, { "epoch": 3.47, "learning_rate": 4.8950704225352116e-05, "loss": 0.4383, "step": 250 }, { "epoch": 3.61, "learning_rate": 4.888028169014085e-05, "loss": 0.4362, "step": 260 }, { "epoch": 3.75, "learning_rate": 4.880985915492958e-05, "loss": 0.4127, "step": 270 }, { "epoch": 3.89, "learning_rate": 4.873943661971831e-05, "loss": 0.4151, "step": 280 }, { "epoch": 4.0, "eval_loss": 0.5906756520271301, "eval_runtime": 130.7857, "eval_samples_per_second": 4.068, "eval_steps_per_second": 0.512, "eval_wer": 0.27730282188082533, "step": 288 }, { "epoch": 4.03, "learning_rate": 4.866901408450704e-05, "loss": 0.4483, "step": 290 }, { "epoch": 4.17, "learning_rate": 4.859859154929577e-05, "loss": 0.4168, "step": 300 }, { "epoch": 4.3, "learning_rate": 4.852816901408451e-05, "loss": 0.4305, "step": 310 }, { "epoch": 4.44, "learning_rate": 4.845774647887324e-05, "loss": 0.4207, "step": 320 }, { "epoch": 4.58, "learning_rate": 4.838732394366197e-05, "loss": 0.4307, "step": 330 }, { "epoch": 4.72, "learning_rate": 4.8316901408450706e-05, "loss": 0.4278, "step": 340 }, { "epoch": 4.86, "learning_rate": 4.8246478873239435e-05, "loss": 0.4152, "step": 350 }, { "epoch": 5.0, "learning_rate": 4.817605633802817e-05, "loss": 0.4229, "step": 360 }, { "epoch": 5.0, "eval_loss": 0.5352545976638794, "eval_runtime": 128.166, "eval_samples_per_second": 4.151, "eval_steps_per_second": 0.523, "eval_wer": 0.2502410490454458, "step": 360 }, { "epoch": 5.14, "learning_rate": 4.8105633802816905e-05, "loss": 0.4146, "step": 370 }, { "epoch": 5.28, "learning_rate": 4.8035211267605634e-05, "loss": 0.4246, "step": 380 }, { "epoch": 5.42, "learning_rate": 4.796478873239437e-05, "loss": 0.4175, "step": 390 }, { "epoch": 5.55, "learning_rate": 4.78943661971831e-05, "loss": 0.4417, "step": 400 }, { "epoch": 5.69, "learning_rate": 4.782394366197183e-05, "loss": 0.4303, "step": 410 }, { "epoch": 5.83, "learning_rate": 4.775352112676056e-05, "loss": 0.4343, "step": 420 }, { "epoch": 5.97, "learning_rate": 4.7683098591549296e-05, "loss": 0.4229, "step": 430 }, { "epoch": 6.0, "eval_loss": 0.5441553592681885, "eval_runtime": 128.9208, "eval_samples_per_second": 4.127, "eval_steps_per_second": 0.52, "eval_wer": 0.2606543678087035, "step": 432 }, { "epoch": 6.11, "learning_rate": 4.761267605633803e-05, "loss": 0.4585, "step": 440 }, { "epoch": 6.25, "learning_rate": 4.754225352112676e-05, "loss": 0.4034, "step": 450 }, { "epoch": 6.39, "learning_rate": 4.7471830985915495e-05, "loss": 0.4069, "step": 460 }, { "epoch": 6.53, "learning_rate": 4.7401408450704224e-05, "loss": 0.4411, "step": 470 }, { "epoch": 6.66, "learning_rate": 4.733098591549296e-05, "loss": 0.4241, "step": 480 }, { "epoch": 6.8, "learning_rate": 4.7260563380281694e-05, "loss": 0.427, "step": 490 }, { "epoch": 6.94, "learning_rate": 4.719014084507042e-05, "loss": 0.4106, "step": 500 }, { "epoch": 7.0, "eval_loss": 0.5387436747550964, "eval_runtime": 128.5459, "eval_samples_per_second": 4.139, "eval_steps_per_second": 0.521, "eval_wer": 0.2513338047181333, "step": 504 }, { "epoch": 7.08, "learning_rate": 4.711971830985916e-05, "loss": 0.4533, "step": 510 }, { "epoch": 7.22, "learning_rate": 4.704929577464789e-05, "loss": 0.4188, "step": 520 }, { "epoch": 7.36, "learning_rate": 4.697887323943662e-05, "loss": 0.4313, "step": 530 }, { "epoch": 7.5, "learning_rate": 4.690845070422536e-05, "loss": 0.4281, "step": 540 }, { "epoch": 7.64, "learning_rate": 4.6838028169014086e-05, "loss": 0.4031, "step": 550 }, { "epoch": 7.78, "learning_rate": 4.676760563380282e-05, "loss": 0.4223, "step": 560 }, { "epoch": 7.91, "learning_rate": 4.669718309859155e-05, "loss": 0.4161, "step": 570 }, { "epoch": 8.0, "eval_loss": 0.5407251715660095, "eval_runtime": 127.4676, "eval_samples_per_second": 4.174, "eval_steps_per_second": 0.526, "eval_wer": 0.2591759336633027, "step": 576 }, { "epoch": 8.06, "learning_rate": 4.6626760563380285e-05, "loss": 0.4399, "step": 580 }, { "epoch": 8.19, "learning_rate": 4.655633802816901e-05, "loss": 0.4198, "step": 590 }, { "epoch": 8.33, "learning_rate": 4.648591549295775e-05, "loss": 0.4121, "step": 600 }, { "epoch": 8.47, "learning_rate": 4.6415492957746484e-05, "loss": 0.4238, "step": 610 }, { "epoch": 8.61, "learning_rate": 4.634507042253521e-05, "loss": 0.4315, "step": 620 }, { "epoch": 8.75, "learning_rate": 4.627464788732395e-05, "loss": 0.4255, "step": 630 }, { "epoch": 8.89, "learning_rate": 4.6204225352112676e-05, "loss": 0.4072, "step": 640 }, { "epoch": 9.0, "eval_loss": 0.5012030601501465, "eval_runtime": 128.1858, "eval_samples_per_second": 4.15, "eval_steps_per_second": 0.523, "eval_wer": 0.24934113260911486, "step": 648 }, { "epoch": 9.03, "learning_rate": 4.613380281690141e-05, "loss": 0.4213, "step": 650 }, { "epoch": 9.17, "learning_rate": 4.6063380281690146e-05, "loss": 0.4047, "step": 660 }, { "epoch": 9.3, "learning_rate": 4.5992957746478875e-05, "loss": 0.437, "step": 670 }, { "epoch": 9.44, "learning_rate": 4.592253521126761e-05, "loss": 0.4155, "step": 680 }, { "epoch": 9.58, "learning_rate": 4.585211267605634e-05, "loss": 0.4113, "step": 690 }, { "epoch": 9.72, "learning_rate": 4.5781690140845074e-05, "loss": 0.4107, "step": 700 }, { "epoch": 9.86, "learning_rate": 4.571126760563381e-05, "loss": 0.4069, "step": 710 }, { "epoch": 10.0, "learning_rate": 4.564084507042254e-05, "loss": 0.4301, "step": 720 }, { "epoch": 10.0, "eval_loss": 0.5448527932167053, "eval_runtime": 128.8981, "eval_samples_per_second": 4.127, "eval_steps_per_second": 0.52, "eval_wer": 0.2541621135180305, "step": 720 }, { "epoch": 10.14, "learning_rate": 4.557042253521127e-05, "loss": 0.402, "step": 730 }, { "epoch": 10.28, "learning_rate": 4.55e-05, "loss": 0.4327, "step": 740 }, { "epoch": 10.42, "learning_rate": 4.542957746478874e-05, "loss": 0.4174, "step": 750 }, { "epoch": 10.55, "learning_rate": 4.535915492957747e-05, "loss": 0.4304, "step": 760 }, { "epoch": 10.69, "learning_rate": 4.52887323943662e-05, "loss": 0.4234, "step": 770 }, { "epoch": 10.83, "learning_rate": 4.5218309859154936e-05, "loss": 0.4194, "step": 780 }, { "epoch": 10.97, "learning_rate": 4.5147887323943664e-05, "loss": 0.419, "step": 790 }, { "epoch": 11.0, "eval_loss": 0.5024985671043396, "eval_runtime": 127.8059, "eval_samples_per_second": 4.163, "eval_steps_per_second": 0.524, "eval_wer": 0.25576910715433565, "step": 792 }, { "epoch": 11.11, "learning_rate": 4.50774647887324e-05, "loss": 0.4251, "step": 800 }, { "epoch": 11.25, "learning_rate": 4.500704225352113e-05, "loss": 0.4161, "step": 810 }, { "epoch": 11.39, "learning_rate": 4.493661971830986e-05, "loss": 0.4053, "step": 820 }, { "epoch": 11.53, "learning_rate": 4.48661971830986e-05, "loss": 0.444, "step": 830 }, { "epoch": 11.66, "learning_rate": 4.479577464788733e-05, "loss": 0.3947, "step": 840 }, { "epoch": 11.8, "learning_rate": 4.472535211267606e-05, "loss": 0.4309, "step": 850 }, { "epoch": 11.94, "learning_rate": 4.465492957746479e-05, "loss": 0.4336, "step": 860 }, { "epoch": 12.0, "eval_loss": 0.5266738533973694, "eval_runtime": 128.1272, "eval_samples_per_second": 4.152, "eval_steps_per_second": 0.523, "eval_wer": 0.24509866940926914, "step": 864 }, { "epoch": 12.08, "learning_rate": 4.4584507042253526e-05, "loss": 0.4158, "step": 870 }, { "epoch": 12.22, "learning_rate": 4.4514084507042254e-05, "loss": 0.3992, "step": 880 }, { "epoch": 12.36, "learning_rate": 4.444366197183098e-05, "loss": 0.4215, "step": 890 }, { "epoch": 12.5, "learning_rate": 4.437323943661972e-05, "loss": 0.4284, "step": 900 }, { "epoch": 12.64, "learning_rate": 4.4302816901408453e-05, "loss": 0.3886, "step": 910 }, { "epoch": 12.78, "learning_rate": 4.423239436619718e-05, "loss": 0.4216, "step": 920 }, { "epoch": 12.91, "learning_rate": 4.416197183098592e-05, "loss": 0.4188, "step": 930 }, { "epoch": 13.0, "eval_loss": 0.5068269371986389, "eval_runtime": 127.9442, "eval_samples_per_second": 4.158, "eval_steps_per_second": 0.524, "eval_wer": 0.2582117374815196, "step": 936 }, { "epoch": 13.06, "learning_rate": 4.4091549295774646e-05, "loss": 0.4384, "step": 940 }, { "epoch": 13.19, "learning_rate": 4.402112676056338e-05, "loss": 0.4001, "step": 950 }, { "epoch": 13.33, "learning_rate": 4.395070422535211e-05, "loss": 0.4263, "step": 960 }, { "epoch": 13.47, "learning_rate": 4.3880281690140845e-05, "loss": 0.4052, "step": 970 }, { "epoch": 13.61, "learning_rate": 4.380985915492958e-05, "loss": 0.3927, "step": 980 }, { "epoch": 13.75, "learning_rate": 4.373943661971831e-05, "loss": 0.428, "step": 990 }, { "epoch": 13.89, "learning_rate": 4.3669014084507044e-05, "loss": 0.4005, "step": 1000 }, { "epoch": 14.0, "eval_loss": 0.5392433404922485, "eval_runtime": 128.4034, "eval_samples_per_second": 4.143, "eval_steps_per_second": 0.522, "eval_wer": 0.2528765186089863, "step": 1008 }, { "epoch": 14.03, "learning_rate": 4.359859154929577e-05, "loss": 0.4283, "step": 1010 }, { "epoch": 14.17, "learning_rate": 4.352816901408451e-05, "loss": 0.4031, "step": 1020 }, { "epoch": 14.3, "learning_rate": 4.345774647887324e-05, "loss": 0.4168, "step": 1030 }, { "epoch": 14.44, "learning_rate": 4.338732394366197e-05, "loss": 0.3987, "step": 1040 }, { "epoch": 14.58, "learning_rate": 4.3316901408450707e-05, "loss": 0.4338, "step": 1050 }, { "epoch": 14.72, "learning_rate": 4.3246478873239435e-05, "loss": 0.4093, "step": 1060 }, { "epoch": 14.86, "learning_rate": 4.317605633802817e-05, "loss": 0.4004, "step": 1070 }, { "epoch": 15.0, "learning_rate": 4.3105633802816906e-05, "loss": 0.4213, "step": 1080 }, { "epoch": 15.0, "eval_loss": 0.5082538723945618, "eval_runtime": 127.3218, "eval_samples_per_second": 4.178, "eval_steps_per_second": 0.526, "eval_wer": 0.24921257311821046, "step": 1080 }, { "epoch": 15.14, "learning_rate": 4.3035211267605634e-05, "loss": 0.4228, "step": 1090 }, { "epoch": 15.28, "learning_rate": 4.296478873239437e-05, "loss": 0.4163, "step": 1100 }, { "epoch": 15.42, "learning_rate": 4.28943661971831e-05, "loss": 0.3912, "step": 1110 }, { "epoch": 15.55, "learning_rate": 4.282394366197183e-05, "loss": 0.4152, "step": 1120 }, { "epoch": 15.69, "learning_rate": 4.275352112676057e-05, "loss": 0.3893, "step": 1130 }, { "epoch": 15.83, "learning_rate": 4.26830985915493e-05, "loss": 0.4163, "step": 1140 }, { "epoch": 15.97, "learning_rate": 4.261267605633803e-05, "loss": 0.4123, "step": 1150 }, { "epoch": 16.0, "eval_loss": 0.4909058213233948, "eval_runtime": 127.0485, "eval_samples_per_second": 4.187, "eval_steps_per_second": 0.527, "eval_wer": 0.24811981744552292, "step": 1152 }, { "epoch": 16.11, "learning_rate": 4.254225352112676e-05, "loss": 0.4027, "step": 1160 }, { "epoch": 16.25, "learning_rate": 4.2471830985915496e-05, "loss": 0.4021, "step": 1170 }, { "epoch": 16.39, "learning_rate": 4.2401408450704224e-05, "loss": 0.3876, "step": 1180 }, { "epoch": 16.53, "learning_rate": 4.233098591549296e-05, "loss": 0.4161, "step": 1190 }, { "epoch": 16.66, "learning_rate": 4.2260563380281695e-05, "loss": 0.3975, "step": 1200 }, { "epoch": 16.8, "learning_rate": 4.219014084507042e-05, "loss": 0.3785, "step": 1210 }, { "epoch": 16.94, "learning_rate": 4.211971830985916e-05, "loss": 0.3854, "step": 1220 }, { "epoch": 17.0, "eval_loss": 0.5269584655761719, "eval_runtime": 128.6883, "eval_samples_per_second": 4.134, "eval_steps_per_second": 0.521, "eval_wer": 0.2593044931542071, "step": 1224 }, { "epoch": 17.08, "learning_rate": 4.204929577464789e-05, "loss": 0.4326, "step": 1230 }, { "epoch": 17.22, "learning_rate": 4.197887323943662e-05, "loss": 0.3852, "step": 1240 }, { "epoch": 17.36, "learning_rate": 4.190845070422536e-05, "loss": 0.3947, "step": 1250 }, { "epoch": 17.5, "learning_rate": 4.1838028169014086e-05, "loss": 0.4118, "step": 1260 }, { "epoch": 17.64, "learning_rate": 4.176760563380282e-05, "loss": 0.3857, "step": 1270 }, { "epoch": 17.78, "learning_rate": 4.169718309859155e-05, "loss": 0.4201, "step": 1280 }, { "epoch": 17.91, "learning_rate": 4.1626760563380285e-05, "loss": 0.3887, "step": 1290 }, { "epoch": 18.0, "eval_loss": 0.50589519739151, "eval_runtime": 126.713, "eval_samples_per_second": 4.198, "eval_steps_per_second": 0.529, "eval_wer": 0.250305328790898, "step": 1296 }, { "epoch": 18.06, "learning_rate": 4.155633802816902e-05, "loss": 0.4153, "step": 1300 }, { "epoch": 18.19, "learning_rate": 4.148591549295775e-05, "loss": 0.4241, "step": 1310 }, { "epoch": 18.33, "learning_rate": 4.1415492957746484e-05, "loss": 0.4298, "step": 1320 }, { "epoch": 18.47, "learning_rate": 4.134507042253521e-05, "loss": 0.4065, "step": 1330 }, { "epoch": 18.61, "learning_rate": 4.127464788732395e-05, "loss": 0.4064, "step": 1340 }, { "epoch": 18.75, "learning_rate": 4.120422535211268e-05, "loss": 0.398, "step": 1350 }, { "epoch": 18.89, "learning_rate": 4.113380281690141e-05, "loss": 0.3934, "step": 1360 }, { "epoch": 19.0, "eval_loss": 0.4928908944129944, "eval_runtime": 126.8145, "eval_samples_per_second": 4.195, "eval_steps_per_second": 0.528, "eval_wer": 0.2352638683550813, "step": 1368 }, { "epoch": 19.03, "learning_rate": 4.106338028169015e-05, "loss": 0.4086, "step": 1370 }, { "epoch": 19.17, "learning_rate": 4.0992957746478875e-05, "loss": 0.3826, "step": 1380 }, { "epoch": 19.3, "learning_rate": 4.092253521126761e-05, "loss": 0.4094, "step": 1390 }, { "epoch": 19.44, "learning_rate": 4.085211267605634e-05, "loss": 0.3973, "step": 1400 }, { "epoch": 19.58, "learning_rate": 4.0781690140845074e-05, "loss": 0.4066, "step": 1410 }, { "epoch": 19.72, "learning_rate": 4.071126760563381e-05, "loss": 0.3961, "step": 1420 }, { "epoch": 19.86, "learning_rate": 4.064084507042254e-05, "loss": 0.402, "step": 1430 }, { "epoch": 20.0, "learning_rate": 4.057042253521127e-05, "loss": 0.4005, "step": 1440 }, { "epoch": 20.0, "eval_loss": 0.5170095562934875, "eval_runtime": 127.4875, "eval_samples_per_second": 4.173, "eval_steps_per_second": 0.526, "eval_wer": 0.24439159220929485, "step": 1440 }, { "epoch": 20.14, "learning_rate": 4.05e-05, "loss": 0.3946, "step": 1450 }, { "epoch": 20.28, "learning_rate": 4.042957746478874e-05, "loss": 0.3916, "step": 1460 }, { "epoch": 20.42, "learning_rate": 4.0359154929577466e-05, "loss": 0.382, "step": 1470 }, { "epoch": 20.55, "learning_rate": 4.0288732394366194e-05, "loss": 0.4081, "step": 1480 }, { "epoch": 20.69, "learning_rate": 4.021830985915493e-05, "loss": 0.3917, "step": 1490 }, { "epoch": 20.83, "learning_rate": 4.0147887323943665e-05, "loss": 0.4083, "step": 1500 }, { "epoch": 20.97, "learning_rate": 4.007746478873239e-05, "loss": 0.4018, "step": 1510 }, { "epoch": 21.0, "eval_loss": 0.4931071102619171, "eval_runtime": 127.824, "eval_samples_per_second": 4.162, "eval_steps_per_second": 0.524, "eval_wer": 0.24059908722761458, "step": 1512 }, { "epoch": 21.11, "learning_rate": 4.000704225352113e-05, "loss": 0.3987, "step": 1520 }, { "epoch": 21.25, "learning_rate": 3.993661971830986e-05, "loss": 0.4043, "step": 1530 }, { "epoch": 21.39, "learning_rate": 3.986619718309859e-05, "loss": 0.398, "step": 1540 }, { "epoch": 21.53, "learning_rate": 3.979577464788732e-05, "loss": 0.406, "step": 1550 }, { "epoch": 21.66, "learning_rate": 3.9725352112676056e-05, "loss": 0.3786, "step": 1560 }, { "epoch": 21.8, "learning_rate": 3.965492957746479e-05, "loss": 0.3908, "step": 1570 }, { "epoch": 21.94, "learning_rate": 3.958450704225352e-05, "loss": 0.3981, "step": 1580 }, { "epoch": 22.0, "eval_loss": 0.4988018870353699, "eval_runtime": 129.0602, "eval_samples_per_second": 4.122, "eval_steps_per_second": 0.519, "eval_wer": 0.23449251140965482, "step": 1584 }, { "epoch": 22.08, "learning_rate": 3.9514084507042255e-05, "loss": 0.4121, "step": 1590 }, { "epoch": 22.22, "learning_rate": 3.9443661971830983e-05, "loss": 0.3837, "step": 1600 }, { "epoch": 22.36, "learning_rate": 3.937323943661972e-05, "loss": 0.3951, "step": 1610 }, { "epoch": 22.5, "learning_rate": 3.9302816901408454e-05, "loss": 0.4119, "step": 1620 }, { "epoch": 22.64, "learning_rate": 3.923239436619718e-05, "loss": 0.3912, "step": 1630 }, { "epoch": 22.78, "learning_rate": 3.916197183098592e-05, "loss": 0.3787, "step": 1640 }, { "epoch": 22.91, "learning_rate": 3.9091549295774646e-05, "loss": 0.3645, "step": 1650 }, { "epoch": 23.0, "eval_loss": 0.5220637321472168, "eval_runtime": 127.4377, "eval_samples_per_second": 4.175, "eval_steps_per_second": 0.526, "eval_wer": 0.2494696921000193, "step": 1656 }, { "epoch": 23.06, "learning_rate": 3.902112676056338e-05, "loss": 0.403, "step": 1660 }, { "epoch": 23.19, "learning_rate": 3.895070422535212e-05, "loss": 0.3992, "step": 1670 }, { "epoch": 23.33, "learning_rate": 3.8880281690140845e-05, "loss": 0.3958, "step": 1680 }, { "epoch": 23.47, "learning_rate": 3.880985915492958e-05, "loss": 0.4091, "step": 1690 }, { "epoch": 23.61, "learning_rate": 3.873943661971831e-05, "loss": 0.4006, "step": 1700 }, { "epoch": 23.75, "learning_rate": 3.8669014084507044e-05, "loss": 0.3856, "step": 1710 }, { "epoch": 23.89, "learning_rate": 3.859859154929578e-05, "loss": 0.3727, "step": 1720 }, { "epoch": 24.0, "eval_loss": 0.5192346572875977, "eval_runtime": 128.6258, "eval_samples_per_second": 4.136, "eval_steps_per_second": 0.521, "eval_wer": 0.23609950504596003, "step": 1728 }, { "epoch": 24.03, "learning_rate": 3.852816901408451e-05, "loss": 0.3787, "step": 1730 }, { "epoch": 24.17, "learning_rate": 3.845774647887324e-05, "loss": 0.3807, "step": 1740 }, { "epoch": 24.3, "learning_rate": 3.838732394366197e-05, "loss": 0.3858, "step": 1750 }, { "epoch": 24.44, "learning_rate": 3.831690140845071e-05, "loss": 0.3864, "step": 1760 }, { "epoch": 24.58, "learning_rate": 3.8246478873239435e-05, "loss": 0.3966, "step": 1770 }, { "epoch": 24.72, "learning_rate": 3.817605633802817e-05, "loss": 0.3775, "step": 1780 }, { "epoch": 24.86, "learning_rate": 3.8105633802816906e-05, "loss": 0.3813, "step": 1790 }, { "epoch": 25.0, "learning_rate": 3.8035211267605634e-05, "loss": 0.3751, "step": 1800 }, { "epoch": 25.0, "eval_loss": 0.4995923638343811, "eval_runtime": 127.8241, "eval_samples_per_second": 4.162, "eval_steps_per_second": 0.524, "eval_wer": 0.24027768850035353, "step": 1800 }, { "epoch": 25.14, "learning_rate": 3.796478873239437e-05, "loss": 0.3957, "step": 1810 }, { "epoch": 25.28, "learning_rate": 3.78943661971831e-05, "loss": 0.3909, "step": 1820 }, { "epoch": 25.42, "learning_rate": 3.7823943661971833e-05, "loss": 0.3882, "step": 1830 }, { "epoch": 25.55, "learning_rate": 3.775352112676057e-05, "loss": 0.391, "step": 1840 }, { "epoch": 25.69, "learning_rate": 3.76830985915493e-05, "loss": 0.3897, "step": 1850 }, { "epoch": 25.83, "learning_rate": 3.761267605633803e-05, "loss": 0.3926, "step": 1860 }, { "epoch": 25.97, "learning_rate": 3.754225352112676e-05, "loss": 0.3839, "step": 1870 }, { "epoch": 26.0, "eval_loss": 0.4840199053287506, "eval_runtime": 127.8495, "eval_samples_per_second": 4.161, "eval_steps_per_second": 0.524, "eval_wer": 0.22272931799190074, "step": 1872 }, { "epoch": 26.11, "learning_rate": 3.7471830985915496e-05, "loss": 0.3878, "step": 1880 }, { "epoch": 26.25, "learning_rate": 3.740140845070423e-05, "loss": 0.4001, "step": 1890 }, { "epoch": 26.39, "learning_rate": 3.733098591549296e-05, "loss": 0.3706, "step": 1900 }, { "epoch": 26.53, "learning_rate": 3.7260563380281695e-05, "loss": 0.3816, "step": 1910 }, { "epoch": 26.66, "learning_rate": 3.7190140845070424e-05, "loss": 0.3735, "step": 1920 }, { "epoch": 26.8, "learning_rate": 3.711971830985916e-05, "loss": 0.3979, "step": 1930 }, { "epoch": 26.94, "learning_rate": 3.704929577464789e-05, "loss": 0.3739, "step": 1940 }, { "epoch": 27.0, "eval_loss": 0.4790037274360657, "eval_runtime": 129.2383, "eval_samples_per_second": 4.116, "eval_steps_per_second": 0.518, "eval_wer": 0.2264575432281288, "step": 1944 }, { "epoch": 27.08, "learning_rate": 3.697887323943662e-05, "loss": 0.3946, "step": 1950 }, { "epoch": 27.22, "learning_rate": 3.690845070422536e-05, "loss": 0.3866, "step": 1960 }, { "epoch": 27.36, "learning_rate": 3.6838028169014087e-05, "loss": 0.3747, "step": 1970 }, { "epoch": 27.5, "learning_rate": 3.676760563380282e-05, "loss": 0.3778, "step": 1980 }, { "epoch": 27.64, "learning_rate": 3.669718309859155e-05, "loss": 0.3772, "step": 1990 }, { "epoch": 27.78, "learning_rate": 3.6626760563380286e-05, "loss": 0.3872, "step": 2000 }, { "epoch": 27.91, "learning_rate": 3.655633802816902e-05, "loss": 0.3781, "step": 2010 }, { "epoch": 28.0, "eval_loss": 0.5081353783607483, "eval_runtime": 127.0954, "eval_samples_per_second": 4.186, "eval_steps_per_second": 0.527, "eval_wer": 0.2247219901009192, "step": 2016 }, { "epoch": 28.06, "learning_rate": 3.648591549295775e-05, "loss": 0.389, "step": 2020 }, { "epoch": 28.19, "learning_rate": 3.6415492957746485e-05, "loss": 0.3737, "step": 2030 }, { "epoch": 28.33, "learning_rate": 3.634507042253521e-05, "loss": 0.4003, "step": 2040 }, { "epoch": 28.47, "learning_rate": 3.627464788732395e-05, "loss": 0.3863, "step": 2050 }, { "epoch": 28.61, "learning_rate": 3.620422535211268e-05, "loss": 0.395, "step": 2060 }, { "epoch": 28.75, "learning_rate": 3.6133802816901405e-05, "loss": 0.3946, "step": 2070 }, { "epoch": 28.89, "learning_rate": 3.606338028169014e-05, "loss": 0.3655, "step": 2080 }, { "epoch": 29.0, "eval_loss": 0.4903033375740051, "eval_runtime": 127.2191, "eval_samples_per_second": 4.182, "eval_steps_per_second": 0.527, "eval_wer": 0.2262647039917722, "step": 2088 }, { "epoch": 29.03, "learning_rate": 3.5992957746478876e-05, "loss": 0.3724, "step": 2090 }, { "epoch": 29.17, "learning_rate": 3.5922535211267604e-05, "loss": 0.3667, "step": 2100 }, { "epoch": 29.3, "learning_rate": 3.585211267605634e-05, "loss": 0.3724, "step": 2110 }, { "epoch": 29.44, "learning_rate": 3.578169014084507e-05, "loss": 0.3749, "step": 2120 }, { "epoch": 29.58, "learning_rate": 3.57112676056338e-05, "loss": 0.3839, "step": 2130 }, { "epoch": 29.72, "learning_rate": 3.564084507042253e-05, "loss": 0.3756, "step": 2140 }, { "epoch": 29.86, "learning_rate": 3.557042253521127e-05, "loss": 0.3799, "step": 2150 }, { "epoch": 30.0, "learning_rate": 3.55e-05, "loss": 0.3832, "step": 2160 }, { "epoch": 30.0, "eval_loss": 0.5195701718330383, "eval_runtime": 128.2447, "eval_samples_per_second": 4.148, "eval_steps_per_second": 0.522, "eval_wer": 0.2382850163913351, "step": 2160 }, { "epoch": 30.14, "learning_rate": 3.543661971830986e-05, "loss": 0.3806, "step": 2170 }, { "epoch": 30.28, "learning_rate": 3.536619718309859e-05, "loss": 0.379, "step": 2180 }, { "epoch": 30.42, "learning_rate": 3.5295774647887326e-05, "loss": 0.369, "step": 2190 }, { "epoch": 30.55, "learning_rate": 3.5225352112676055e-05, "loss": 0.3781, "step": 2200 }, { "epoch": 30.69, "learning_rate": 3.515492957746479e-05, "loss": 0.3635, "step": 2210 }, { "epoch": 30.83, "learning_rate": 3.5084507042253525e-05, "loss": 0.3843, "step": 2220 }, { "epoch": 30.97, "learning_rate": 3.5014084507042254e-05, "loss": 0.3662, "step": 2230 }, { "epoch": 31.0, "eval_loss": 0.5286476612091064, "eval_runtime": 127.7591, "eval_samples_per_second": 4.164, "eval_steps_per_second": 0.524, "eval_wer": 0.23629234428231663, "step": 2232 }, { "epoch": 31.11, "learning_rate": 3.494366197183099e-05, "loss": 0.3803, "step": 2240 }, { "epoch": 31.25, "learning_rate": 3.487323943661972e-05, "loss": 0.3676, "step": 2250 }, { "epoch": 31.39, "learning_rate": 3.480281690140845e-05, "loss": 0.3443, "step": 2260 }, { "epoch": 31.53, "learning_rate": 3.473239436619719e-05, "loss": 0.3662, "step": 2270 }, { "epoch": 31.66, "learning_rate": 3.466197183098592e-05, "loss": 0.3751, "step": 2280 }, { "epoch": 31.8, "learning_rate": 3.459154929577465e-05, "loss": 0.3838, "step": 2290 }, { "epoch": 31.94, "learning_rate": 3.452112676056338e-05, "loss": 0.3692, "step": 2300 }, { "epoch": 32.0, "eval_loss": 0.5283333659172058, "eval_runtime": 128.3957, "eval_samples_per_second": 4.143, "eval_steps_per_second": 0.522, "eval_wer": 0.24599858584560005, "step": 2304 }, { "epoch": 32.08, "learning_rate": 3.4450704225352116e-05, "loss": 0.3885, "step": 2310 }, { "epoch": 32.22, "learning_rate": 3.4380281690140844e-05, "loss": 0.3639, "step": 2320 }, { "epoch": 32.36, "learning_rate": 3.430985915492958e-05, "loss": 0.3875, "step": 2330 }, { "epoch": 32.5, "learning_rate": 3.4239436619718315e-05, "loss": 0.3779, "step": 2340 }, { "epoch": 32.64, "learning_rate": 3.416901408450704e-05, "loss": 0.3504, "step": 2350 }, { "epoch": 32.78, "learning_rate": 3.409859154929578e-05, "loss": 0.3638, "step": 2360 }, { "epoch": 32.91, "learning_rate": 3.402816901408451e-05, "loss": 0.3484, "step": 2370 }, { "epoch": 33.0, "eval_loss": 0.49058952927589417, "eval_runtime": 127.7529, "eval_samples_per_second": 4.164, "eval_steps_per_second": 0.524, "eval_wer": 0.22665038246448543, "step": 2376 }, { "epoch": 33.06, "learning_rate": 3.395774647887324e-05, "loss": 0.387, "step": 2380 }, { "epoch": 33.19, "learning_rate": 3.388732394366198e-05, "loss": 0.3559, "step": 2390 }, { "epoch": 33.33, "learning_rate": 3.3816901408450706e-05, "loss": 0.3691, "step": 2400 }, { "epoch": 33.47, "learning_rate": 3.374647887323944e-05, "loss": 0.371, "step": 2410 }, { "epoch": 33.61, "learning_rate": 3.367605633802817e-05, "loss": 0.3688, "step": 2420 }, { "epoch": 33.75, "learning_rate": 3.3605633802816905e-05, "loss": 0.3602, "step": 2430 }, { "epoch": 33.89, "learning_rate": 3.353521126760564e-05, "loss": 0.3493, "step": 2440 }, { "epoch": 34.0, "eval_loss": 0.4982738196849823, "eval_runtime": 129.337, "eval_samples_per_second": 4.113, "eval_steps_per_second": 0.518, "eval_wer": 0.2214437230828566, "step": 2448 }, { "epoch": 34.03, "learning_rate": 3.346478873239437e-05, "loss": 0.3825, "step": 2450 }, { "epoch": 34.17, "learning_rate": 3.3394366197183104e-05, "loss": 0.3668, "step": 2460 }, { "epoch": 34.3, "learning_rate": 3.332394366197183e-05, "loss": 0.3679, "step": 2470 }, { "epoch": 34.44, "learning_rate": 3.325352112676057e-05, "loss": 0.3697, "step": 2480 }, { "epoch": 34.58, "learning_rate": 3.31830985915493e-05, "loss": 0.3699, "step": 2490 }, { "epoch": 34.72, "learning_rate": 3.311267605633803e-05, "loss": 0.3794, "step": 2500 }, { "epoch": 34.86, "learning_rate": 3.304225352112677e-05, "loss": 0.3838, "step": 2510 }, { "epoch": 35.0, "learning_rate": 3.2971830985915495e-05, "loss": 0.3783, "step": 2520 }, { "epoch": 35.0, "eval_loss": 0.5358861088752747, "eval_runtime": 128.4601, "eval_samples_per_second": 4.141, "eval_steps_per_second": 0.522, "eval_wer": 0.2319213215915665, "step": 2520 }, { "epoch": 35.14, "learning_rate": 3.2901408450704224e-05, "loss": 0.3732, "step": 2530 }, { "epoch": 35.28, "learning_rate": 3.283098591549296e-05, "loss": 0.3765, "step": 2540 }, { "epoch": 35.42, "learning_rate": 3.276056338028169e-05, "loss": 0.3705, "step": 2550 }, { "epoch": 35.55, "learning_rate": 3.269014084507042e-05, "loss": 0.3824, "step": 2560 }, { "epoch": 35.69, "learning_rate": 3.261971830985915e-05, "loss": 0.368, "step": 2570 }, { "epoch": 35.83, "learning_rate": 3.2549295774647887e-05, "loss": 0.3926, "step": 2580 }, { "epoch": 35.97, "learning_rate": 3.247887323943662e-05, "loss": 0.3604, "step": 2590 }, { "epoch": 36.0, "eval_loss": 0.5129276514053345, "eval_runtime": 127.2766, "eval_samples_per_second": 4.18, "eval_steps_per_second": 0.526, "eval_wer": 0.23706370122774315, "step": 2592 }, { "epoch": 36.11, "learning_rate": 3.240845070422535e-05, "loss": 0.379, "step": 2600 }, { "epoch": 36.25, "learning_rate": 3.2338028169014086e-05, "loss": 0.3653, "step": 2610 }, { "epoch": 36.39, "learning_rate": 3.2267605633802814e-05, "loss": 0.3473, "step": 2620 }, { "epoch": 36.53, "learning_rate": 3.219718309859155e-05, "loss": 0.3891, "step": 2630 }, { "epoch": 36.66, "learning_rate": 3.2126760563380285e-05, "loss": 0.3686, "step": 2640 }, { "epoch": 36.8, "learning_rate": 3.205633802816901e-05, "loss": 0.3736, "step": 2650 }, { "epoch": 36.94, "learning_rate": 3.198591549295775e-05, "loss": 0.3796, "step": 2660 }, { "epoch": 37.0, "eval_loss": 0.4993954300880432, "eval_runtime": 129.0338, "eval_samples_per_second": 4.123, "eval_steps_per_second": 0.519, "eval_wer": 0.2320498810824709, "step": 2664 }, { "epoch": 37.08, "learning_rate": 3.191549295774648e-05, "loss": 0.3864, "step": 2670 }, { "epoch": 37.22, "learning_rate": 3.184507042253521e-05, "loss": 0.3758, "step": 2680 }, { "epoch": 37.36, "learning_rate": 3.177464788732394e-05, "loss": 0.3665, "step": 2690 }, { "epoch": 37.5, "learning_rate": 3.1704225352112676e-05, "loss": 0.3618, "step": 2700 }, { "epoch": 37.64, "learning_rate": 3.163380281690141e-05, "loss": 0.3579, "step": 2710 }, { "epoch": 37.78, "learning_rate": 3.156338028169014e-05, "loss": 0.3909, "step": 2720 }, { "epoch": 37.91, "learning_rate": 3.1492957746478875e-05, "loss": 0.3629, "step": 2730 }, { "epoch": 38.0, "eval_loss": 0.47256624698638916, "eval_runtime": 126.9577, "eval_samples_per_second": 4.19, "eval_steps_per_second": 0.528, "eval_wer": 0.22260075850099634, "step": 2736 }, { "epoch": 38.06, "learning_rate": 3.14225352112676e-05, "loss": 0.38, "step": 2740 }, { "epoch": 38.19, "learning_rate": 3.135211267605634e-05, "loss": 0.3617, "step": 2750 }, { "epoch": 38.33, "learning_rate": 3.1281690140845074e-05, "loss": 0.3469, "step": 2760 }, { "epoch": 38.47, "learning_rate": 3.12112676056338e-05, "loss": 0.3528, "step": 2770 }, { "epoch": 38.61, "learning_rate": 3.114084507042254e-05, "loss": 0.3533, "step": 2780 }, { "epoch": 38.75, "learning_rate": 3.1070422535211266e-05, "loss": 0.3649, "step": 2790 }, { "epoch": 38.89, "learning_rate": 3.1e-05, "loss": 0.3428, "step": 2800 }, { "epoch": 39.0, "eval_loss": 0.5130040645599365, "eval_runtime": 127.749, "eval_samples_per_second": 4.164, "eval_steps_per_second": 0.524, "eval_wer": 0.22253647875554414, "step": 2808 }, { "epoch": 39.03, "learning_rate": 3.092957746478874e-05, "loss": 0.3717, "step": 2810 }, { "epoch": 39.17, "learning_rate": 3.0859154929577465e-05, "loss": 0.36, "step": 2820 }, { "epoch": 39.3, "learning_rate": 3.07887323943662e-05, "loss": 0.3595, "step": 2830 }, { "epoch": 39.44, "learning_rate": 3.071830985915493e-05, "loss": 0.3378, "step": 2840 }, { "epoch": 39.58, "learning_rate": 3.0647887323943664e-05, "loss": 0.3543, "step": 2850 }, { "epoch": 39.72, "learning_rate": 3.05774647887324e-05, "loss": 0.3628, "step": 2860 }, { "epoch": 39.86, "learning_rate": 3.0507042253521128e-05, "loss": 0.3473, "step": 2870 }, { "epoch": 40.0, "learning_rate": 3.043661971830986e-05, "loss": 0.375, "step": 2880 }, { "epoch": 40.0, "eval_loss": 0.5154946446418762, "eval_runtime": 127.9888, "eval_samples_per_second": 4.157, "eval_steps_per_second": 0.523, "eval_wer": 0.22851449508259947, "step": 2880 }, { "epoch": 40.14, "learning_rate": 3.0366197183098595e-05, "loss": 0.375, "step": 2890 }, { "epoch": 40.28, "learning_rate": 3.0295774647887327e-05, "loss": 0.3668, "step": 2900 }, { "epoch": 40.42, "learning_rate": 3.022535211267606e-05, "loss": 0.3508, "step": 2910 }, { "epoch": 40.55, "learning_rate": 3.015492957746479e-05, "loss": 0.3602, "step": 2920 }, { "epoch": 40.69, "learning_rate": 3.0084507042253523e-05, "loss": 0.3435, "step": 2930 }, { "epoch": 40.83, "learning_rate": 3.0014084507042254e-05, "loss": 0.3787, "step": 2940 }, { "epoch": 40.97, "learning_rate": 2.994366197183099e-05, "loss": 0.3581, "step": 2950 }, { "epoch": 41.0, "eval_loss": 0.4994366765022278, "eval_runtime": 127.9492, "eval_samples_per_second": 4.158, "eval_steps_per_second": 0.524, "eval_wer": 0.22665038246448543, "step": 2952 }, { "epoch": 41.11, "learning_rate": 2.988028169014085e-05, "loss": 0.3779, "step": 2960 }, { "epoch": 41.25, "learning_rate": 2.9809859154929582e-05, "loss": 0.3643, "step": 2970 }, { "epoch": 41.39, "learning_rate": 2.9739436619718314e-05, "loss": 0.3476, "step": 2980 }, { "epoch": 41.53, "learning_rate": 2.9669014084507046e-05, "loss": 0.3483, "step": 2990 }, { "epoch": 41.66, "learning_rate": 2.9598591549295778e-05, "loss": 0.3596, "step": 3000 }, { "epoch": 41.8, "learning_rate": 2.9528169014084506e-05, "loss": 0.3649, "step": 3010 }, { "epoch": 41.94, "learning_rate": 2.9457746478873238e-05, "loss": 0.3394, "step": 3020 }, { "epoch": 42.0, "eval_loss": 0.4989403784275055, "eval_runtime": 129.1484, "eval_samples_per_second": 4.119, "eval_steps_per_second": 0.519, "eval_wer": 0.21765121810117632, "step": 3024 }, { "epoch": 42.08, "learning_rate": 2.938732394366197e-05, "loss": 0.3666, "step": 3030 }, { "epoch": 42.22, "learning_rate": 2.93169014084507e-05, "loss": 0.3416, "step": 3040 }, { "epoch": 42.36, "learning_rate": 2.9246478873239437e-05, "loss": 0.3542, "step": 3050 }, { "epoch": 42.5, "learning_rate": 2.917605633802817e-05, "loss": 0.3639, "step": 3060 }, { "epoch": 42.64, "learning_rate": 2.91056338028169e-05, "loss": 0.3514, "step": 3070 }, { "epoch": 42.78, "learning_rate": 2.9035211267605633e-05, "loss": 0.3658, "step": 3080 }, { "epoch": 42.91, "learning_rate": 2.8964788732394364e-05, "loss": 0.3421, "step": 3090 }, { "epoch": 43.0, "eval_loss": 0.4820369780063629, "eval_runtime": 128.7733, "eval_samples_per_second": 4.131, "eval_steps_per_second": 0.52, "eval_wer": 0.22735745966445972, "step": 3096 }, { "epoch": 43.06, "learning_rate": 2.88943661971831e-05, "loss": 0.3618, "step": 3100 }, { "epoch": 43.19, "learning_rate": 2.882394366197183e-05, "loss": 0.3449, "step": 3110 }, { "epoch": 43.33, "learning_rate": 2.8753521126760563e-05, "loss": 0.3446, "step": 3120 }, { "epoch": 43.47, "learning_rate": 2.8683098591549295e-05, "loss": 0.3456, "step": 3130 }, { "epoch": 43.61, "learning_rate": 2.8612676056338027e-05, "loss": 0.3545, "step": 3140 }, { "epoch": 43.75, "learning_rate": 2.854225352112676e-05, "loss": 0.3493, "step": 3150 }, { "epoch": 43.89, "learning_rate": 2.8471830985915494e-05, "loss": 0.342, "step": 3160 }, { "epoch": 44.0, "eval_loss": 0.48394179344177246, "eval_runtime": 129.9101, "eval_samples_per_second": 4.095, "eval_steps_per_second": 0.516, "eval_wer": 0.21887253326476827, "step": 3168 }, { "epoch": 44.03, "learning_rate": 2.8401408450704226e-05, "loss": 0.3561, "step": 3170 }, { "epoch": 44.17, "learning_rate": 2.8330985915492958e-05, "loss": 0.3402, "step": 3180 }, { "epoch": 44.3, "learning_rate": 2.826056338028169e-05, "loss": 0.3596, "step": 3190 }, { "epoch": 44.44, "learning_rate": 2.8190140845070422e-05, "loss": 0.3485, "step": 3200 }, { "epoch": 44.58, "learning_rate": 2.8119718309859157e-05, "loss": 0.3435, "step": 3210 }, { "epoch": 44.72, "learning_rate": 2.804929577464789e-05, "loss": 0.332, "step": 3220 }, { "epoch": 44.86, "learning_rate": 2.797887323943662e-05, "loss": 0.3339, "step": 3230 }, { "epoch": 45.0, "learning_rate": 2.7908450704225353e-05, "loss": 0.3482, "step": 3240 }, { "epoch": 45.0, "eval_loss": 0.4978794753551483, "eval_runtime": 129.1828, "eval_samples_per_second": 4.118, "eval_steps_per_second": 0.519, "eval_wer": 0.21199460050138202, "step": 3240 }, { "epoch": 45.14, "learning_rate": 2.7838028169014085e-05, "loss": 0.3325, "step": 3250 }, { "epoch": 45.28, "learning_rate": 2.7767605633802816e-05, "loss": 0.3409, "step": 3260 }, { "epoch": 45.42, "learning_rate": 2.7697183098591552e-05, "loss": 0.3368, "step": 3270 }, { "epoch": 45.55, "learning_rate": 2.7626760563380284e-05, "loss": 0.35, "step": 3280 }, { "epoch": 45.69, "learning_rate": 2.7556338028169015e-05, "loss": 0.3433, "step": 3290 }, { "epoch": 45.83, "learning_rate": 2.7485915492957747e-05, "loss": 0.3441, "step": 3300 }, { "epoch": 45.97, "learning_rate": 2.741549295774648e-05, "loss": 0.3494, "step": 3310 }, { "epoch": 46.0, "eval_loss": 0.5068593621253967, "eval_runtime": 128.5981, "eval_samples_per_second": 4.137, "eval_steps_per_second": 0.521, "eval_wer": 0.21861541428295944, "step": 3312 }, { "epoch": 46.11, "learning_rate": 2.734507042253521e-05, "loss": 0.3578, "step": 3320 }, { "epoch": 46.25, "learning_rate": 2.7274647887323946e-05, "loss": 0.3451, "step": 3330 }, { "epoch": 46.39, "learning_rate": 2.7204225352112678e-05, "loss": 0.3417, "step": 3340 }, { "epoch": 46.53, "learning_rate": 2.713380281690141e-05, "loss": 0.3444, "step": 3350 }, { "epoch": 46.66, "learning_rate": 2.7063380281690142e-05, "loss": 0.3409, "step": 3360 }, { "epoch": 46.8, "learning_rate": 2.6992957746478874e-05, "loss": 0.3477, "step": 3370 }, { "epoch": 46.94, "learning_rate": 2.692253521126761e-05, "loss": 0.345, "step": 3380 }, { "epoch": 47.0, "eval_loss": 0.49980834126472473, "eval_runtime": 128.798, "eval_samples_per_second": 4.13, "eval_steps_per_second": 0.52, "eval_wer": 0.23211416082792313, "step": 3384 }, { "epoch": 47.08, "learning_rate": 2.685211267605634e-05, "loss": 0.3616, "step": 3390 }, { "epoch": 47.22, "learning_rate": 2.6781690140845073e-05, "loss": 0.3455, "step": 3400 }, { "epoch": 47.36, "learning_rate": 2.6711267605633805e-05, "loss": 0.3323, "step": 3410 }, { "epoch": 47.5, "learning_rate": 2.6640845070422537e-05, "loss": 0.3476, "step": 3420 }, { "epoch": 47.64, "learning_rate": 2.657042253521127e-05, "loss": 0.3346, "step": 3430 }, { "epoch": 47.78, "learning_rate": 2.6500000000000004e-05, "loss": 0.3439, "step": 3440 }, { "epoch": 47.91, "learning_rate": 2.6429577464788736e-05, "loss": 0.327, "step": 3450 }, { "epoch": 48.0, "eval_loss": 0.49777674674987793, "eval_runtime": 127.6593, "eval_samples_per_second": 4.167, "eval_steps_per_second": 0.525, "eval_wer": 0.21893681301022047, "step": 3456 }, { "epoch": 48.06, "learning_rate": 2.6359154929577468e-05, "loss": 0.3601, "step": 3460 }, { "epoch": 48.19, "learning_rate": 2.62887323943662e-05, "loss": 0.3493, "step": 3470 }, { "epoch": 48.33, "learning_rate": 2.621830985915493e-05, "loss": 0.3506, "step": 3480 }, { "epoch": 48.47, "learning_rate": 2.6147887323943667e-05, "loss": 0.3293, "step": 3490 }, { "epoch": 48.61, "learning_rate": 2.60774647887324e-05, "loss": 0.3398, "step": 3500 }, { "epoch": 48.75, "learning_rate": 2.600704225352113e-05, "loss": 0.3399, "step": 3510 }, { "epoch": 48.89, "learning_rate": 2.5936619718309862e-05, "loss": 0.3266, "step": 3520 }, { "epoch": 49.0, "eval_loss": 0.5078500509262085, "eval_runtime": 127.4612, "eval_samples_per_second": 4.174, "eval_steps_per_second": 0.526, "eval_wer": 0.21874397377386384, "step": 3528 }, { "epoch": 49.03, "learning_rate": 2.5866197183098594e-05, "loss": 0.3485, "step": 3530 }, { "epoch": 49.17, "learning_rate": 2.5795774647887326e-05, "loss": 0.3352, "step": 3540 }, { "epoch": 49.3, "learning_rate": 2.572535211267606e-05, "loss": 0.3424, "step": 3550 }, { "epoch": 49.44, "learning_rate": 2.5654929577464793e-05, "loss": 0.3218, "step": 3560 }, { "epoch": 49.58, "learning_rate": 2.5584507042253525e-05, "loss": 0.3407, "step": 3570 }, { "epoch": 49.72, "learning_rate": 2.5514084507042257e-05, "loss": 0.3429, "step": 3580 }, { "epoch": 49.86, "learning_rate": 2.544366197183099e-05, "loss": 0.333, "step": 3590 }, { "epoch": 50.0, "learning_rate": 2.5373239436619717e-05, "loss": 0.3353, "step": 3600 }, { "epoch": 50.0, "eval_loss": 0.5302192568778992, "eval_runtime": 129.7397, "eval_samples_per_second": 4.101, "eval_steps_per_second": 0.516, "eval_wer": 0.22350067493732725, "step": 3600 }, { "epoch": 50.14, "learning_rate": 2.530281690140845e-05, "loss": 0.3433, "step": 3610 }, { "epoch": 50.28, "learning_rate": 2.523239436619718e-05, "loss": 0.3373, "step": 3620 }, { "epoch": 50.42, "learning_rate": 2.5161971830985913e-05, "loss": 0.3393, "step": 3630 }, { "epoch": 50.55, "learning_rate": 2.5091549295774648e-05, "loss": 0.3473, "step": 3640 }, { "epoch": 50.69, "learning_rate": 2.502112676056338e-05, "loss": 0.3399, "step": 3650 }, { "epoch": 50.83, "learning_rate": 2.4950704225352115e-05, "loss": 0.3421, "step": 3660 }, { "epoch": 50.97, "learning_rate": 2.4880281690140847e-05, "loss": 0.3384, "step": 3670 }, { "epoch": 51.0, "eval_loss": 0.4837527275085449, "eval_runtime": 128.8469, "eval_samples_per_second": 4.129, "eval_steps_per_second": 0.52, "eval_wer": 0.21167320177412097, "step": 3672 }, { "epoch": 51.11, "learning_rate": 2.480985915492958e-05, "loss": 0.3385, "step": 3680 }, { "epoch": 51.25, "learning_rate": 2.473943661971831e-05, "loss": 0.3374, "step": 3690 }, { "epoch": 51.39, "learning_rate": 2.4669014084507046e-05, "loss": 0.3305, "step": 3700 }, { "epoch": 51.53, "learning_rate": 2.4598591549295778e-05, "loss": 0.3462, "step": 3710 }, { "epoch": 51.66, "learning_rate": 2.452816901408451e-05, "loss": 0.3221, "step": 3720 }, { "epoch": 51.8, "learning_rate": 2.4457746478873242e-05, "loss": 0.3276, "step": 3730 }, { "epoch": 51.94, "learning_rate": 2.4387323943661974e-05, "loss": 0.3259, "step": 3740 }, { "epoch": 52.0, "eval_loss": 0.5022213459014893, "eval_runtime": 129.1853, "eval_samples_per_second": 4.118, "eval_steps_per_second": 0.519, "eval_wer": 0.21694414090120204, "step": 3744 }, { "epoch": 52.08, "learning_rate": 2.4316901408450705e-05, "loss": 0.3393, "step": 3750 }, { "epoch": 52.22, "learning_rate": 2.4246478873239437e-05, "loss": 0.3306, "step": 3760 }, { "epoch": 52.36, "learning_rate": 2.417605633802817e-05, "loss": 0.3342, "step": 3770 }, { "epoch": 52.5, "learning_rate": 2.41056338028169e-05, "loss": 0.3341, "step": 3780 }, { "epoch": 52.64, "learning_rate": 2.4035211267605633e-05, "loss": 0.3149, "step": 3790 }, { "epoch": 52.78, "learning_rate": 2.3964788732394365e-05, "loss": 0.3452, "step": 3800 }, { "epoch": 52.91, "learning_rate": 2.38943661971831e-05, "loss": 0.336, "step": 3810 }, { "epoch": 53.0, "eval_loss": 0.4834235608577728, "eval_runtime": 128.5245, "eval_samples_per_second": 4.139, "eval_steps_per_second": 0.521, "eval_wer": 0.2214437230828566, "step": 3816 }, { "epoch": 53.06, "learning_rate": 2.3823943661971832e-05, "loss": 0.3509, "step": 3820 }, { "epoch": 53.19, "learning_rate": 2.3753521126760564e-05, "loss": 0.326, "step": 3830 }, { "epoch": 53.33, "learning_rate": 2.3683098591549296e-05, "loss": 0.3393, "step": 3840 }, { "epoch": 53.47, "learning_rate": 2.3612676056338028e-05, "loss": 0.3369, "step": 3850 }, { "epoch": 53.61, "learning_rate": 2.3542253521126763e-05, "loss": 0.3317, "step": 3860 }, { "epoch": 53.75, "learning_rate": 2.3471830985915495e-05, "loss": 0.3275, "step": 3870 }, { "epoch": 53.89, "learning_rate": 2.3401408450704227e-05, "loss": 0.3123, "step": 3880 }, { "epoch": 54.0, "eval_loss": 0.5113204121589661, "eval_runtime": 130.2117, "eval_samples_per_second": 4.086, "eval_steps_per_second": 0.515, "eval_wer": 0.21642990293758438, "step": 3888 }, { "epoch": 54.03, "learning_rate": 2.333098591549296e-05, "loss": 0.3572, "step": 3890 }, { "epoch": 54.17, "learning_rate": 2.326056338028169e-05, "loss": 0.3197, "step": 3900 }, { "epoch": 54.3, "learning_rate": 2.3190140845070422e-05, "loss": 0.3348, "step": 3910 }, { "epoch": 54.44, "learning_rate": 2.3119718309859158e-05, "loss": 0.3166, "step": 3920 }, { "epoch": 54.58, "learning_rate": 2.304929577464789e-05, "loss": 0.3311, "step": 3930 }, { "epoch": 54.72, "learning_rate": 2.297887323943662e-05, "loss": 0.3262, "step": 3940 }, { "epoch": 54.86, "learning_rate": 2.2908450704225353e-05, "loss": 0.3358, "step": 3950 }, { "epoch": 55.0, "learning_rate": 2.2838028169014085e-05, "loss": 0.3306, "step": 3960 }, { "epoch": 55.0, "eval_loss": 0.47700902819633484, "eval_runtime": 129.6952, "eval_samples_per_second": 4.102, "eval_steps_per_second": 0.517, "eval_wer": 0.2111589638105033, "step": 3960 }, { "epoch": 55.14, "learning_rate": 2.276760563380282e-05, "loss": 0.3319, "step": 3970 }, { "epoch": 55.28, "learning_rate": 2.2697183098591552e-05, "loss": 0.3313, "step": 3980 }, { "epoch": 55.42, "learning_rate": 2.2626760563380284e-05, "loss": 0.3325, "step": 3990 }, { "epoch": 55.55, "learning_rate": 2.2556338028169016e-05, "loss": 0.3255, "step": 4000 }, { "epoch": 55.69, "learning_rate": 2.2485915492957748e-05, "loss": 0.3299, "step": 4010 }, { "epoch": 55.83, "learning_rate": 2.241549295774648e-05, "loss": 0.3321, "step": 4020 }, { "epoch": 55.97, "learning_rate": 2.2345070422535215e-05, "loss": 0.3344, "step": 4030 }, { "epoch": 56.0, "eval_loss": 0.4752146601676941, "eval_runtime": 128.9657, "eval_samples_per_second": 4.125, "eval_steps_per_second": 0.52, "eval_wer": 0.21302307642861734, "step": 4032 }, { "epoch": 56.11, "learning_rate": 2.2274647887323943e-05, "loss": 0.3311, "step": 4040 }, { "epoch": 56.25, "learning_rate": 2.2204225352112675e-05, "loss": 0.3314, "step": 4050 }, { "epoch": 56.39, "learning_rate": 2.2133802816901407e-05, "loss": 0.3128, "step": 4060 }, { "epoch": 56.53, "learning_rate": 2.2063380281690142e-05, "loss": 0.3471, "step": 4070 }, { "epoch": 56.66, "learning_rate": 2.1992957746478874e-05, "loss": 0.3281, "step": 4080 }, { "epoch": 56.8, "learning_rate": 2.1922535211267606e-05, "loss": 0.3371, "step": 4090 }, { "epoch": 56.94, "learning_rate": 2.1852112676056338e-05, "loss": 0.3279, "step": 4100 }, { "epoch": 57.0, "eval_loss": 0.4953095316886902, "eval_runtime": 129.8414, "eval_samples_per_second": 4.097, "eval_steps_per_second": 0.516, "eval_wer": 0.21231599922864305, "step": 4104 }, { "epoch": 57.08, "learning_rate": 2.178169014084507e-05, "loss": 0.3253, "step": 4110 }, { "epoch": 57.22, "learning_rate": 2.1711267605633802e-05, "loss": 0.33, "step": 4120 }, { "epoch": 57.36, "learning_rate": 2.1640845070422537e-05, "loss": 0.3386, "step": 4130 }, { "epoch": 57.5, "learning_rate": 2.157042253521127e-05, "loss": 0.3367, "step": 4140 }, { "epoch": 57.64, "learning_rate": 2.15e-05, "loss": 0.3266, "step": 4150 }, { "epoch": 57.78, "learning_rate": 2.1429577464788733e-05, "loss": 0.3329, "step": 4160 }, { "epoch": 57.91, "learning_rate": 2.1359154929577465e-05, "loss": 0.3067, "step": 4170 }, { "epoch": 58.0, "eval_loss": 0.4822605550289154, "eval_runtime": 127.9027, "eval_samples_per_second": 4.159, "eval_steps_per_second": 0.524, "eval_wer": 0.2175226586102719, "step": 4176 }, { "epoch": 58.06, "learning_rate": 2.12887323943662e-05, "loss": 0.3345, "step": 4180 }, { "epoch": 58.19, "learning_rate": 2.1218309859154932e-05, "loss": 0.3205, "step": 4190 }, { "epoch": 58.33, "learning_rate": 2.1147887323943664e-05, "loss": 0.3362, "step": 4200 }, { "epoch": 58.47, "learning_rate": 2.1077464788732395e-05, "loss": 0.323, "step": 4210 }, { "epoch": 58.61, "learning_rate": 2.1007042253521127e-05, "loss": 0.3101, "step": 4220 }, { "epoch": 58.75, "learning_rate": 2.093661971830986e-05, "loss": 0.3338, "step": 4230 }, { "epoch": 58.89, "learning_rate": 2.0866197183098594e-05, "loss": 0.3131, "step": 4240 }, { "epoch": 59.0, "eval_loss": 0.48572954535484314, "eval_runtime": 127.6235, "eval_samples_per_second": 4.169, "eval_steps_per_second": 0.525, "eval_wer": 0.21366587388313943, "step": 4248 }, { "epoch": 59.03, "learning_rate": 2.0795774647887326e-05, "loss": 0.3319, "step": 4250 }, { "epoch": 59.17, "learning_rate": 2.0725352112676058e-05, "loss": 0.3056, "step": 4260 }, { "epoch": 59.3, "learning_rate": 2.065492957746479e-05, "loss": 0.3301, "step": 4270 }, { "epoch": 59.44, "learning_rate": 2.0584507042253522e-05, "loss": 0.333, "step": 4280 }, { "epoch": 59.58, "learning_rate": 2.0514084507042254e-05, "loss": 0.3269, "step": 4290 }, { "epoch": 59.72, "learning_rate": 2.044366197183099e-05, "loss": 0.3316, "step": 4300 }, { "epoch": 59.86, "learning_rate": 2.037323943661972e-05, "loss": 0.3096, "step": 4310 }, { "epoch": 60.0, "learning_rate": 2.0302816901408453e-05, "loss": 0.3178, "step": 4320 }, { "epoch": 60.0, "eval_loss": 0.49693578481674194, "eval_runtime": 128.0482, "eval_samples_per_second": 4.155, "eval_steps_per_second": 0.523, "eval_wer": 0.21540142701034903, "step": 4320 }, { "epoch": 60.14, "learning_rate": 2.0232394366197185e-05, "loss": 0.3303, "step": 4330 }, { "epoch": 60.28, "learning_rate": 2.0161971830985917e-05, "loss": 0.3226, "step": 4340 }, { "epoch": 60.42, "learning_rate": 2.009154929577465e-05, "loss": 0.3207, "step": 4350 }, { "epoch": 60.55, "learning_rate": 2.002112676056338e-05, "loss": 0.3228, "step": 4360 }, { "epoch": 60.69, "learning_rate": 1.9950704225352112e-05, "loss": 0.3258, "step": 4370 }, { "epoch": 60.83, "learning_rate": 1.9880281690140844e-05, "loss": 0.3223, "step": 4380 }, { "epoch": 60.97, "learning_rate": 1.9809859154929576e-05, "loss": 0.3306, "step": 4390 }, { "epoch": 61.0, "eval_loss": 0.4816165566444397, "eval_runtime": 128.2312, "eval_samples_per_second": 4.149, "eval_steps_per_second": 0.522, "eval_wer": 0.21655846242848878, "step": 4392 }, { "epoch": 61.11, "learning_rate": 1.973943661971831e-05, "loss": 0.3323, "step": 4400 }, { "epoch": 61.25, "learning_rate": 1.9669014084507043e-05, "loss": 0.3327, "step": 4410 }, { "epoch": 61.39, "learning_rate": 1.9598591549295775e-05, "loss": 0.3154, "step": 4420 }, { "epoch": 61.53, "learning_rate": 1.9528169014084507e-05, "loss": 0.3179, "step": 4430 }, { "epoch": 61.66, "learning_rate": 1.945774647887324e-05, "loss": 0.305, "step": 4440 }, { "epoch": 61.8, "learning_rate": 1.9387323943661974e-05, "loss": 0.331, "step": 4450 }, { "epoch": 61.94, "learning_rate": 1.9316901408450706e-05, "loss": 0.3152, "step": 4460 }, { "epoch": 62.0, "eval_loss": 0.4694591760635376, "eval_runtime": 128.0603, "eval_samples_per_second": 4.154, "eval_steps_per_second": 0.523, "eval_wer": 0.21064472584688565, "step": 4464 }, { "epoch": 62.08, "learning_rate": 1.9246478873239438e-05, "loss": 0.3376, "step": 4470 }, { "epoch": 62.22, "learning_rate": 1.917605633802817e-05, "loss": 0.3176, "step": 4480 }, { "epoch": 62.36, "learning_rate": 1.91056338028169e-05, "loss": 0.3249, "step": 4490 }, { "epoch": 62.5, "learning_rate": 1.9035211267605633e-05, "loss": 0.3214, "step": 4500 }, { "epoch": 62.64, "learning_rate": 1.896478873239437e-05, "loss": 0.3153, "step": 4510 }, { "epoch": 62.78, "learning_rate": 1.88943661971831e-05, "loss": 0.3308, "step": 4520 }, { "epoch": 62.91, "learning_rate": 1.8823943661971832e-05, "loss": 0.3119, "step": 4530 }, { "epoch": 63.0, "eval_loss": 0.5111212730407715, "eval_runtime": 128.7341, "eval_samples_per_second": 4.133, "eval_steps_per_second": 0.52, "eval_wer": 0.21058044610143345, "step": 4536 }, { "epoch": 63.06, "learning_rate": 1.8753521126760564e-05, "loss": 0.3313, "step": 4540 }, { "epoch": 63.19, "learning_rate": 1.8683098591549296e-05, "loss": 0.3123, "step": 4550 }, { "epoch": 63.33, "learning_rate": 1.861267605633803e-05, "loss": 0.3206, "step": 4560 }, { "epoch": 63.47, "learning_rate": 1.8542253521126763e-05, "loss": 0.3139, "step": 4570 }, { "epoch": 63.61, "learning_rate": 1.8471830985915495e-05, "loss": 0.3169, "step": 4580 }, { "epoch": 63.75, "learning_rate": 1.8401408450704227e-05, "loss": 0.3155, "step": 4590 }, { "epoch": 63.89, "learning_rate": 1.833098591549296e-05, "loss": 0.3124, "step": 4600 }, { "epoch": 64.0, "eval_loss": 0.501674473285675, "eval_runtime": 127.9707, "eval_samples_per_second": 4.157, "eval_steps_per_second": 0.524, "eval_wer": 0.2118660410104776, "step": 4608 }, { "epoch": 64.03, "learning_rate": 1.826056338028169e-05, "loss": 0.3236, "step": 4610 }, { "epoch": 64.17, "learning_rate": 1.8190140845070426e-05, "loss": 0.3077, "step": 4620 }, { "epoch": 64.3, "learning_rate": 1.8119718309859155e-05, "loss": 0.321, "step": 4630 }, { "epoch": 64.44, "learning_rate": 1.8049295774647886e-05, "loss": 0.3019, "step": 4640 }, { "epoch": 64.58, "learning_rate": 1.797887323943662e-05, "loss": 0.3181, "step": 4650 }, { "epoch": 64.72, "learning_rate": 1.790845070422535e-05, "loss": 0.3113, "step": 4660 }, { "epoch": 64.86, "learning_rate": 1.7838028169014085e-05, "loss": 0.3191, "step": 4670 }, { "epoch": 65.0, "learning_rate": 1.7767605633802817e-05, "loss": 0.317, "step": 4680 }, { "epoch": 65.0, "eval_loss": 0.5071455836296082, "eval_runtime": 129.2281, "eval_samples_per_second": 4.117, "eval_steps_per_second": 0.518, "eval_wer": 0.21649418268303658, "step": 4680 }, { "epoch": 65.14, "learning_rate": 1.769718309859155e-05, "loss": 0.3253, "step": 4690 }, { "epoch": 65.28, "learning_rate": 1.762676056338028e-05, "loss": 0.3272, "step": 4700 }, { "epoch": 65.42, "learning_rate": 1.7556338028169013e-05, "loss": 0.3143, "step": 4710 }, { "epoch": 65.55, "learning_rate": 1.7485915492957748e-05, "loss": 0.3108, "step": 4720 }, { "epoch": 65.69, "learning_rate": 1.741549295774648e-05, "loss": 0.3075, "step": 4730 }, { "epoch": 65.83, "learning_rate": 1.7345070422535212e-05, "loss": 0.3167, "step": 4740 }, { "epoch": 65.97, "learning_rate": 1.7274647887323944e-05, "loss": 0.2934, "step": 4750 }, { "epoch": 66.0, "eval_loss": 0.4819434583187103, "eval_runtime": 128.1216, "eval_samples_per_second": 4.152, "eval_steps_per_second": 0.523, "eval_wer": 0.2168155814102976, "step": 4752 }, { "epoch": 66.11, "learning_rate": 1.7204225352112676e-05, "loss": 0.3263, "step": 4760 }, { "epoch": 66.25, "learning_rate": 1.7133802816901408e-05, "loss": 0.3156, "step": 4770 }, { "epoch": 66.39, "learning_rate": 1.7063380281690143e-05, "loss": 0.3101, "step": 4780 }, { "epoch": 66.53, "learning_rate": 1.6992957746478875e-05, "loss": 0.3222, "step": 4790 }, { "epoch": 66.66, "learning_rate": 1.6922535211267607e-05, "loss": 0.3104, "step": 4800 }, { "epoch": 66.8, "learning_rate": 1.685211267605634e-05, "loss": 0.3104, "step": 4810 }, { "epoch": 66.94, "learning_rate": 1.678169014084507e-05, "loss": 0.3103, "step": 4820 }, { "epoch": 67.0, "eval_loss": 0.4943528175354004, "eval_runtime": 128.9828, "eval_samples_per_second": 4.125, "eval_steps_per_second": 0.519, "eval_wer": 0.21430867133766152, "step": 4824 }, { "epoch": 67.08, "learning_rate": 1.6711267605633806e-05, "loss": 0.3063, "step": 4830 }, { "epoch": 67.22, "learning_rate": 1.6640845070422538e-05, "loss": 0.3118, "step": 4840 }, { "epoch": 67.36, "learning_rate": 1.657042253521127e-05, "loss": 0.2975, "step": 4850 }, { "epoch": 67.5, "learning_rate": 1.65e-05, "loss": 0.3126, "step": 4860 }, { "epoch": 67.64, "learning_rate": 1.6429577464788733e-05, "loss": 0.3022, "step": 4870 }, { "epoch": 67.78, "learning_rate": 1.6359154929577465e-05, "loss": 0.3155, "step": 4880 }, { "epoch": 67.91, "learning_rate": 1.62887323943662e-05, "loss": 0.3078, "step": 4890 }, { "epoch": 68.0, "eval_loss": 0.5003769993782043, "eval_runtime": 127.2601, "eval_samples_per_second": 4.18, "eval_steps_per_second": 0.526, "eval_wer": 0.2110304043195989, "step": 4896 }, { "epoch": 68.06, "learning_rate": 1.6218309859154932e-05, "loss": 0.3161, "step": 4900 }, { "epoch": 68.19, "learning_rate": 1.6147887323943664e-05, "loss": 0.3094, "step": 4910 }, { "epoch": 68.33, "learning_rate": 1.6077464788732396e-05, "loss": 0.3191, "step": 4920 }, { "epoch": 68.47, "learning_rate": 1.6007042253521128e-05, "loss": 0.3246, "step": 4930 }, { "epoch": 68.61, "learning_rate": 1.593661971830986e-05, "loss": 0.3078, "step": 4940 }, { "epoch": 68.75, "learning_rate": 1.586619718309859e-05, "loss": 0.3044, "step": 4950 }, { "epoch": 68.89, "learning_rate": 1.5795774647887323e-05, "loss": 0.3084, "step": 4960 }, { "epoch": 69.0, "eval_loss": 0.49981018900871277, "eval_runtime": 127.4871, "eval_samples_per_second": 4.173, "eval_steps_per_second": 0.526, "eval_wer": 0.21475862955582695, "step": 4968 }, { "epoch": 69.03, "learning_rate": 1.5725352112676055e-05, "loss": 0.3225, "step": 4970 }, { "epoch": 69.17, "learning_rate": 1.5654929577464787e-05, "loss": 0.3125, "step": 4980 }, { "epoch": 69.3, "learning_rate": 1.5584507042253522e-05, "loss": 0.3189, "step": 4990 }, { "epoch": 69.44, "learning_rate": 1.5514084507042254e-05, "loss": 0.3132, "step": 5000 }, { "epoch": 69.58, "learning_rate": 1.5443661971830986e-05, "loss": 0.319, "step": 5010 }, { "epoch": 69.72, "learning_rate": 1.5373239436619718e-05, "loss": 0.3092, "step": 5020 }, { "epoch": 69.86, "learning_rate": 1.530281690140845e-05, "loss": 0.3108, "step": 5030 }, { "epoch": 70.0, "learning_rate": 1.5232394366197183e-05, "loss": 0.3147, "step": 5040 }, { "epoch": 70.0, "eval_loss": 0.48848283290863037, "eval_runtime": 127.875, "eval_samples_per_second": 4.16, "eval_steps_per_second": 0.524, "eval_wer": 0.20987336890145916, "step": 5040 }, { "epoch": 70.14, "learning_rate": 1.5161971830985915e-05, "loss": 0.312, "step": 5050 }, { "epoch": 70.28, "learning_rate": 1.5091549295774649e-05, "loss": 0.3199, "step": 5060 }, { "epoch": 70.42, "learning_rate": 1.502112676056338e-05, "loss": 0.3058, "step": 5070 }, { "epoch": 70.55, "learning_rate": 1.4950704225352113e-05, "loss": 0.3022, "step": 5080 }, { "epoch": 70.69, "learning_rate": 1.4880281690140846e-05, "loss": 0.3043, "step": 5090 }, { "epoch": 70.83, "learning_rate": 1.4809859154929578e-05, "loss": 0.3183, "step": 5100 }, { "epoch": 70.97, "learning_rate": 1.4739436619718312e-05, "loss": 0.3081, "step": 5110 }, { "epoch": 71.0, "eval_loss": 0.5062239766120911, "eval_runtime": 127.6812, "eval_samples_per_second": 4.167, "eval_steps_per_second": 0.525, "eval_wer": 0.2096162499196503, "step": 5112 }, { "epoch": 71.11, "learning_rate": 1.4669014084507044e-05, "loss": 0.3103, "step": 5120 }, { "epoch": 71.25, "learning_rate": 1.4598591549295775e-05, "loss": 0.2963, "step": 5130 }, { "epoch": 71.39, "learning_rate": 1.4528169014084509e-05, "loss": 0.2978, "step": 5140 }, { "epoch": 71.53, "learning_rate": 1.4457746478873241e-05, "loss": 0.3107, "step": 5150 }, { "epoch": 71.66, "learning_rate": 1.4387323943661973e-05, "loss": 0.3103, "step": 5160 }, { "epoch": 71.8, "learning_rate": 1.4316901408450706e-05, "loss": 0.3113, "step": 5170 }, { "epoch": 71.94, "learning_rate": 1.4246478873239438e-05, "loss": 0.3046, "step": 5180 }, { "epoch": 72.0, "eval_loss": 0.47900620102882385, "eval_runtime": 127.6914, "eval_samples_per_second": 4.166, "eval_steps_per_second": 0.525, "eval_wer": 0.21244455871954748, "step": 5184 }, { "epoch": 72.08, "learning_rate": 1.417605633802817e-05, "loss": 0.3169, "step": 5190 }, { "epoch": 72.22, "learning_rate": 1.4105633802816904e-05, "loss": 0.3112, "step": 5200 }, { "epoch": 72.36, "learning_rate": 1.4035211267605636e-05, "loss": 0.2977, "step": 5210 }, { "epoch": 72.5, "learning_rate": 1.3964788732394367e-05, "loss": 0.312, "step": 5220 }, { "epoch": 72.64, "learning_rate": 1.3894366197183098e-05, "loss": 0.2976, "step": 5230 }, { "epoch": 72.78, "learning_rate": 1.3823943661971831e-05, "loss": 0.2993, "step": 5240 }, { "epoch": 72.91, "learning_rate": 1.3753521126760563e-05, "loss": 0.2993, "step": 5250 }, { "epoch": 73.0, "eval_loss": 0.5045033097267151, "eval_runtime": 127.5139, "eval_samples_per_second": 4.172, "eval_steps_per_second": 0.525, "eval_wer": 0.2096162499196503, "step": 5256 }, { "epoch": 73.06, "learning_rate": 1.3683098591549295e-05, "loss": 0.3093, "step": 5260 }, { "epoch": 73.19, "learning_rate": 1.3612676056338028e-05, "loss": 0.2961, "step": 5270 }, { "epoch": 73.33, "learning_rate": 1.354225352112676e-05, "loss": 0.3014, "step": 5280 }, { "epoch": 73.47, "learning_rate": 1.3471830985915492e-05, "loss": 0.3113, "step": 5290 }, { "epoch": 73.61, "learning_rate": 1.3401408450704226e-05, "loss": 0.2948, "step": 5300 }, { "epoch": 73.75, "learning_rate": 1.3330985915492958e-05, "loss": 0.3022, "step": 5310 }, { "epoch": 73.89, "learning_rate": 1.326056338028169e-05, "loss": 0.2844, "step": 5320 }, { "epoch": 74.0, "eval_loss": 0.4865418076515198, "eval_runtime": 127.6726, "eval_samples_per_second": 4.167, "eval_steps_per_second": 0.525, "eval_wer": 0.20601658417432667, "step": 5328 }, { "epoch": 74.03, "learning_rate": 1.3190140845070423e-05, "loss": 0.3137, "step": 5330 }, { "epoch": 74.17, "learning_rate": 1.3119718309859155e-05, "loss": 0.2849, "step": 5340 }, { "epoch": 74.3, "learning_rate": 1.3049295774647889e-05, "loss": 0.301, "step": 5350 }, { "epoch": 74.44, "learning_rate": 1.297887323943662e-05, "loss": 0.3007, "step": 5360 }, { "epoch": 74.58, "learning_rate": 1.2908450704225352e-05, "loss": 0.3074, "step": 5370 }, { "epoch": 74.72, "learning_rate": 1.2838028169014086e-05, "loss": 0.306, "step": 5380 }, { "epoch": 74.86, "learning_rate": 1.2767605633802818e-05, "loss": 0.3103, "step": 5390 }, { "epoch": 75.0, "learning_rate": 1.269718309859155e-05, "loss": 0.3061, "step": 5400 }, { "epoch": 75.0, "eval_loss": 0.49969518184661865, "eval_runtime": 126.5881, "eval_samples_per_second": 4.203, "eval_steps_per_second": 0.529, "eval_wer": 0.2102590473741724, "step": 5400 }, { "epoch": 75.14, "learning_rate": 1.2626760563380283e-05, "loss": 0.3207, "step": 5410 }, { "epoch": 75.28, "learning_rate": 1.2556338028169015e-05, "loss": 0.3112, "step": 5420 }, { "epoch": 75.42, "learning_rate": 1.2485915492957747e-05, "loss": 0.2981, "step": 5430 }, { "epoch": 75.55, "learning_rate": 1.241549295774648e-05, "loss": 0.3007, "step": 5440 }, { "epoch": 75.69, "learning_rate": 1.234507042253521e-05, "loss": 0.2967, "step": 5450 }, { "epoch": 75.83, "learning_rate": 1.2274647887323944e-05, "loss": 0.2986, "step": 5460 }, { "epoch": 75.97, "learning_rate": 1.2204225352112676e-05, "loss": 0.305, "step": 5470 }, { "epoch": 76.0, "eval_loss": 0.5118193626403809, "eval_runtime": 125.4914, "eval_samples_per_second": 4.239, "eval_steps_per_second": 0.534, "eval_wer": 0.21199460050138202, "step": 5472 }, { "epoch": 76.11, "learning_rate": 1.2133802816901408e-05, "loss": 0.3029, "step": 5480 }, { "epoch": 76.25, "learning_rate": 1.2063380281690142e-05, "loss": 0.3084, "step": 5490 }, { "epoch": 76.39, "learning_rate": 1.1992957746478873e-05, "loss": 0.3031, "step": 5500 }, { "epoch": 76.53, "learning_rate": 1.1922535211267605e-05, "loss": 0.3015, "step": 5510 }, { "epoch": 76.66, "learning_rate": 1.1852112676056339e-05, "loss": 0.2959, "step": 5520 }, { "epoch": 76.8, "learning_rate": 1.178169014084507e-05, "loss": 0.3022, "step": 5530 }, { "epoch": 76.94, "learning_rate": 1.1711267605633804e-05, "loss": 0.3015, "step": 5540 }, { "epoch": 77.0, "eval_loss": 0.489970326423645, "eval_runtime": 125.7413, "eval_samples_per_second": 4.231, "eval_steps_per_second": 0.533, "eval_wer": 0.2095519701741981, "step": 5544 }, { "epoch": 77.08, "learning_rate": 1.1640845070422536e-05, "loss": 0.2974, "step": 5550 }, { "epoch": 77.22, "learning_rate": 1.1570422535211268e-05, "loss": 0.3004, "step": 5560 }, { "epoch": 77.36, "learning_rate": 1.1500000000000002e-05, "loss": 0.2996, "step": 5570 }, { "epoch": 77.5, "learning_rate": 1.1429577464788734e-05, "loss": 0.3014, "step": 5580 }, { "epoch": 77.64, "learning_rate": 1.1359154929577465e-05, "loss": 0.2959, "step": 5590 }, { "epoch": 77.78, "learning_rate": 1.1288732394366197e-05, "loss": 0.3052, "step": 5600 }, { "epoch": 77.91, "learning_rate": 1.121830985915493e-05, "loss": 0.2969, "step": 5610 }, { "epoch": 78.0, "eval_loss": 0.49838951230049133, "eval_runtime": 125.5517, "eval_samples_per_second": 4.237, "eval_steps_per_second": 0.534, "eval_wer": 0.21064472584688565, "step": 5616 }, { "epoch": 78.06, "learning_rate": 1.1147887323943663e-05, "loss": 0.3042, "step": 5620 }, { "epoch": 78.19, "learning_rate": 1.1077464788732395e-05, "loss": 0.2955, "step": 5630 }, { "epoch": 78.33, "learning_rate": 1.1007042253521127e-05, "loss": 0.3023, "step": 5640 }, { "epoch": 78.47, "learning_rate": 1.093661971830986e-05, "loss": 0.2917, "step": 5650 }, { "epoch": 78.61, "learning_rate": 1.0866197183098592e-05, "loss": 0.2865, "step": 5660 }, { "epoch": 78.75, "learning_rate": 1.0795774647887324e-05, "loss": 0.2969, "step": 5670 }, { "epoch": 78.89, "learning_rate": 1.0725352112676057e-05, "loss": 0.2988, "step": 5680 }, { "epoch": 79.0, "eval_loss": 0.494792103767395, "eval_runtime": 125.4092, "eval_samples_per_second": 4.242, "eval_steps_per_second": 0.534, "eval_wer": 0.20653082213794433, "step": 5688 }, { "epoch": 79.03, "learning_rate": 1.065492957746479e-05, "loss": 0.3085, "step": 5690 }, { "epoch": 79.17, "learning_rate": 1.0584507042253521e-05, "loss": 0.2926, "step": 5700 }, { "epoch": 79.3, "learning_rate": 1.0514084507042255e-05, "loss": 0.3023, "step": 5710 }, { "epoch": 79.44, "learning_rate": 1.0443661971830987e-05, "loss": 0.2905, "step": 5720 }, { "epoch": 79.58, "learning_rate": 1.037323943661972e-05, "loss": 0.3016, "step": 5730 }, { "epoch": 79.72, "learning_rate": 1.030281690140845e-05, "loss": 0.301, "step": 5740 }, { "epoch": 79.86, "learning_rate": 1.0232394366197182e-05, "loss": 0.2903, "step": 5750 }, { "epoch": 80.0, "learning_rate": 1.0161971830985916e-05, "loss": 0.3006, "step": 5760 }, { "epoch": 80.0, "eval_loss": 0.48496711254119873, "eval_runtime": 125.2817, "eval_samples_per_second": 4.246, "eval_steps_per_second": 0.535, "eval_wer": 0.20318827537442952, "step": 5760 }, { "epoch": 80.14, "learning_rate": 1.0091549295774648e-05, "loss": 0.2919, "step": 5770 }, { "epoch": 80.28, "learning_rate": 1.0028169014084506e-05, "loss": 0.2934, "step": 5780 }, { "epoch": 80.42, "learning_rate": 9.95774647887324e-06, "loss": 0.2944, "step": 5790 }, { "epoch": 80.55, "learning_rate": 9.887323943661972e-06, "loss": 0.2834, "step": 5800 }, { "epoch": 80.69, "learning_rate": 9.816901408450704e-06, "loss": 0.2948, "step": 5810 }, { "epoch": 80.83, "learning_rate": 9.746478873239437e-06, "loss": 0.3089, "step": 5820 }, { "epoch": 80.97, "learning_rate": 9.676056338028169e-06, "loss": 0.2982, "step": 5830 }, { "epoch": 81.0, "eval_loss": 0.49331343173980713, "eval_runtime": 125.9952, "eval_samples_per_second": 4.222, "eval_steps_per_second": 0.532, "eval_wer": 0.2046667095198303, "step": 5832 }, { "epoch": 81.11, "learning_rate": 9.605633802816901e-06, "loss": 0.3012, "step": 5840 }, { "epoch": 81.25, "learning_rate": 9.535211267605635e-06, "loss": 0.3046, "step": 5850 }, { "epoch": 81.39, "learning_rate": 9.464788732394366e-06, "loss": 0.2831, "step": 5860 }, { "epoch": 81.53, "learning_rate": 9.3943661971831e-06, "loss": 0.2947, "step": 5870 }, { "epoch": 81.66, "learning_rate": 9.323943661971832e-06, "loss": 0.3004, "step": 5880 }, { "epoch": 81.8, "learning_rate": 9.253521126760564e-06, "loss": 0.3017, "step": 5890 }, { "epoch": 81.94, "learning_rate": 9.183098591549297e-06, "loss": 0.2938, "step": 5900 }, { "epoch": 82.0, "eval_loss": 0.4996909201145172, "eval_runtime": 126.3183, "eval_samples_per_second": 4.212, "eval_steps_per_second": 0.53, "eval_wer": 0.20524522722890018, "step": 5904 }, { "epoch": 82.08, "learning_rate": 9.11267605633803e-06, "loss": 0.2938, "step": 5910 }, { "epoch": 82.22, "learning_rate": 9.042253521126761e-06, "loss": 0.2973, "step": 5920 }, { "epoch": 82.36, "learning_rate": 8.971830985915493e-06, "loss": 0.2763, "step": 5930 }, { "epoch": 82.5, "learning_rate": 8.901408450704225e-06, "loss": 0.307, "step": 5940 }, { "epoch": 82.64, "learning_rate": 8.830985915492958e-06, "loss": 0.2967, "step": 5950 }, { "epoch": 82.78, "learning_rate": 8.76056338028169e-06, "loss": 0.2951, "step": 5960 }, { "epoch": 82.91, "learning_rate": 8.690140845070422e-06, "loss": 0.2903, "step": 5970 }, { "epoch": 83.0, "eval_loss": 0.49193274974823, "eval_runtime": 125.3571, "eval_samples_per_second": 4.244, "eval_steps_per_second": 0.534, "eval_wer": 0.20344539435623835, "step": 5976 }, { "epoch": 83.06, "learning_rate": 8.619718309859156e-06, "loss": 0.3021, "step": 5980 }, { "epoch": 83.19, "learning_rate": 8.549295774647888e-06, "loss": 0.294, "step": 5990 }, { "epoch": 83.33, "learning_rate": 8.47887323943662e-06, "loss": 0.2897, "step": 6000 }, { "epoch": 83.47, "learning_rate": 8.408450704225353e-06, "loss": 0.2934, "step": 6010 }, { "epoch": 83.61, "learning_rate": 8.338028169014085e-06, "loss": 0.2776, "step": 6020 }, { "epoch": 83.75, "learning_rate": 8.267605633802817e-06, "loss": 0.2944, "step": 6030 }, { "epoch": 83.89, "learning_rate": 8.19718309859155e-06, "loss": 0.297, "step": 6040 }, { "epoch": 84.0, "eval_loss": 0.4807397127151489, "eval_runtime": 125.5219, "eval_samples_per_second": 4.238, "eval_steps_per_second": 0.534, "eval_wer": 0.20260975766535966, "step": 6048 }, { "epoch": 84.03, "learning_rate": 8.126760563380282e-06, "loss": 0.3071, "step": 6050 }, { "epoch": 84.17, "learning_rate": 8.056338028169016e-06, "loss": 0.2797, "step": 6060 }, { "epoch": 84.3, "learning_rate": 7.985915492957746e-06, "loss": 0.2976, "step": 6070 }, { "epoch": 84.44, "learning_rate": 7.915492957746478e-06, "loss": 0.3014, "step": 6080 }, { "epoch": 84.58, "learning_rate": 7.845070422535211e-06, "loss": 0.2899, "step": 6090 }, { "epoch": 84.72, "learning_rate": 7.774647887323943e-06, "loss": 0.2781, "step": 6100 }, { "epoch": 84.86, "learning_rate": 7.704225352112677e-06, "loss": 0.2898, "step": 6110 }, { "epoch": 85.0, "learning_rate": 7.633802816901409e-06, "loss": 0.2955, "step": 6120 }, { "epoch": 85.0, "eval_loss": 0.4953136146068573, "eval_runtime": 125.5494, "eval_samples_per_second": 4.237, "eval_steps_per_second": 0.534, "eval_wer": 0.20479526901073472, "step": 6120 }, { "epoch": 85.14, "learning_rate": 7.5633802816901415e-06, "loss": 0.2777, "step": 6130 }, { "epoch": 85.28, "learning_rate": 7.492957746478873e-06, "loss": 0.2856, "step": 6140 }, { "epoch": 85.42, "learning_rate": 7.422535211267606e-06, "loss": 0.2983, "step": 6150 }, { "epoch": 85.55, "learning_rate": 7.352112676056339e-06, "loss": 0.2938, "step": 6160 }, { "epoch": 85.69, "learning_rate": 7.281690140845071e-06, "loss": 0.2858, "step": 6170 }, { "epoch": 85.83, "learning_rate": 7.211267605633803e-06, "loss": 0.2948, "step": 6180 }, { "epoch": 85.97, "learning_rate": 7.140845070422536e-06, "loss": 0.2847, "step": 6190 }, { "epoch": 86.0, "eval_loss": 0.4950926899909973, "eval_runtime": 126.5682, "eval_samples_per_second": 4.203, "eval_steps_per_second": 0.529, "eval_wer": 0.2021597994471942, "step": 6192 }, { "epoch": 86.11, "learning_rate": 7.070422535211269e-06, "loss": 0.287, "step": 6200 }, { "epoch": 86.25, "learning_rate": 7.000000000000001e-06, "loss": 0.3001, "step": 6210 }, { "epoch": 86.39, "learning_rate": 6.929577464788732e-06, "loss": 0.2815, "step": 6220 }, { "epoch": 86.53, "learning_rate": 6.8591549295774645e-06, "loss": 0.2909, "step": 6230 }, { "epoch": 86.66, "learning_rate": 6.788732394366197e-06, "loss": 0.2766, "step": 6240 }, { "epoch": 86.8, "learning_rate": 6.71830985915493e-06, "loss": 0.2977, "step": 6250 }, { "epoch": 86.94, "learning_rate": 6.647887323943662e-06, "loss": 0.2891, "step": 6260 }, { "epoch": 87.0, "eval_loss": 0.494158536195755, "eval_runtime": 124.8943, "eval_samples_per_second": 4.26, "eval_steps_per_second": 0.536, "eval_wer": 0.20357395384714277, "step": 6264 }, { "epoch": 87.08, "learning_rate": 6.5774647887323945e-06, "loss": 0.3011, "step": 6270 }, { "epoch": 87.22, "learning_rate": 6.507042253521127e-06, "loss": 0.2975, "step": 6280 }, { "epoch": 87.36, "learning_rate": 6.436619718309859e-06, "loss": 0.2863, "step": 6290 }, { "epoch": 87.5, "learning_rate": 6.366197183098592e-06, "loss": 0.2871, "step": 6300 }, { "epoch": 87.64, "learning_rate": 6.295774647887325e-06, "loss": 0.274, "step": 6310 }, { "epoch": 87.78, "learning_rate": 6.225352112676057e-06, "loss": 0.2818, "step": 6320 }, { "epoch": 87.91, "learning_rate": 6.154929577464788e-06, "loss": 0.2895, "step": 6330 }, { "epoch": 88.0, "eval_loss": 0.48882603645324707, "eval_runtime": 126.4329, "eval_samples_per_second": 4.208, "eval_steps_per_second": 0.53, "eval_wer": 0.20203123995628977, "step": 6336 }, { "epoch": 88.06, "learning_rate": 6.084507042253521e-06, "loss": 0.3026, "step": 6340 }, { "epoch": 88.19, "learning_rate": 6.014084507042254e-06, "loss": 0.2863, "step": 6350 }, { "epoch": 88.33, "learning_rate": 5.9436619718309865e-06, "loss": 0.2946, "step": 6360 }, { "epoch": 88.47, "learning_rate": 5.873239436619718e-06, "loss": 0.2898, "step": 6370 }, { "epoch": 88.61, "learning_rate": 5.802816901408451e-06, "loss": 0.2784, "step": 6380 }, { "epoch": 88.75, "learning_rate": 5.732394366197184e-06, "loss": 0.2927, "step": 6390 }, { "epoch": 88.89, "learning_rate": 5.661971830985916e-06, "loss": 0.2823, "step": 6400 }, { "epoch": 89.0, "eval_loss": 0.4967969059944153, "eval_runtime": 125.4743, "eval_samples_per_second": 4.24, "eval_steps_per_second": 0.534, "eval_wer": 0.20325255511988172, "step": 6408 }, { "epoch": 89.03, "learning_rate": 5.5915492957746476e-06, "loss": 0.2929, "step": 6410 }, { "epoch": 89.17, "learning_rate": 5.52112676056338e-06, "loss": 0.2925, "step": 6420 }, { "epoch": 89.3, "learning_rate": 5.450704225352113e-06, "loss": 0.2773, "step": 6430 }, { "epoch": 89.44, "learning_rate": 5.380281690140846e-06, "loss": 0.2801, "step": 6440 }, { "epoch": 89.58, "learning_rate": 5.309859154929578e-06, "loss": 0.2885, "step": 6450 }, { "epoch": 89.72, "learning_rate": 5.23943661971831e-06, "loss": 0.2862, "step": 6460 }, { "epoch": 89.86, "learning_rate": 5.169014084507043e-06, "loss": 0.2866, "step": 6470 }, { "epoch": 90.0, "learning_rate": 5.098591549295775e-06, "loss": 0.2885, "step": 6480 }, { "epoch": 90.0, "eval_loss": 0.5042439103126526, "eval_runtime": 125.2751, "eval_samples_per_second": 4.247, "eval_steps_per_second": 0.535, "eval_wer": 0.203959632319856, "step": 6480 }, { "epoch": 90.14, "learning_rate": 5.028169014084507e-06, "loss": 0.2913, "step": 6490 }, { "epoch": 90.28, "learning_rate": 4.9577464788732395e-06, "loss": 0.2982, "step": 6500 }, { "epoch": 90.42, "learning_rate": 4.887323943661972e-06, "loss": 0.2869, "step": 6510 }, { "epoch": 90.55, "learning_rate": 4.816901408450705e-06, "loss": 0.2898, "step": 6520 }, { "epoch": 90.69, "learning_rate": 4.746478873239437e-06, "loss": 0.2867, "step": 6530 }, { "epoch": 90.83, "learning_rate": 4.67605633802817e-06, "loss": 0.287, "step": 6540 }, { "epoch": 90.97, "learning_rate": 4.6056338028169015e-06, "loss": 0.2876, "step": 6550 }, { "epoch": 91.0, "eval_loss": 0.4943309724330902, "eval_runtime": 125.8105, "eval_samples_per_second": 4.229, "eval_steps_per_second": 0.533, "eval_wer": 0.20254547791990743, "step": 6552 }, { "epoch": 91.11, "learning_rate": 4.535211267605634e-06, "loss": 0.2954, "step": 6560 }, { "epoch": 91.25, "learning_rate": 4.464788732394366e-06, "loss": 0.2929, "step": 6570 }, { "epoch": 91.39, "learning_rate": 4.394366197183099e-06, "loss": 0.2797, "step": 6580 }, { "epoch": 91.53, "learning_rate": 4.3239436619718315e-06, "loss": 0.2857, "step": 6590 }, { "epoch": 91.66, "learning_rate": 4.253521126760563e-06, "loss": 0.2783, "step": 6600 }, { "epoch": 91.8, "learning_rate": 4.183098591549296e-06, "loss": 0.2761, "step": 6610 }, { "epoch": 91.94, "learning_rate": 4.112676056338028e-06, "loss": 0.2765, "step": 6620 }, { "epoch": 92.0, "eval_loss": 0.4961002767086029, "eval_runtime": 124.9062, "eval_samples_per_second": 4.259, "eval_steps_per_second": 0.536, "eval_wer": 0.20035996657453237, "step": 6624 }, { "epoch": 92.08, "learning_rate": 4.042253521126761e-06, "loss": 0.3009, "step": 6630 }, { "epoch": 92.22, "learning_rate": 3.9718309859154926e-06, "loss": 0.2869, "step": 6640 }, { "epoch": 92.36, "learning_rate": 3.901408450704225e-06, "loss": 0.2783, "step": 6650 }, { "epoch": 92.5, "learning_rate": 3.830985915492958e-06, "loss": 0.2872, "step": 6660 }, { "epoch": 92.64, "learning_rate": 3.7605633802816903e-06, "loss": 0.274, "step": 6670 }, { "epoch": 92.78, "learning_rate": 3.690140845070423e-06, "loss": 0.2937, "step": 6680 }, { "epoch": 92.91, "learning_rate": 3.6197183098591553e-06, "loss": 0.282, "step": 6690 }, { "epoch": 93.0, "eval_loss": 0.4846987724304199, "eval_runtime": 125.4194, "eval_samples_per_second": 4.242, "eval_steps_per_second": 0.534, "eval_wer": 0.20170984122902874, "step": 6696 }, { "epoch": 93.06, "learning_rate": 3.549295774647887e-06, "loss": 0.2995, "step": 6700 }, { "epoch": 93.19, "learning_rate": 3.4788732394366195e-06, "loss": 0.2756, "step": 6710 }, { "epoch": 93.33, "learning_rate": 3.4084507042253522e-06, "loss": 0.2797, "step": 6720 }, { "epoch": 93.47, "learning_rate": 3.3380281690140845e-06, "loss": 0.2842, "step": 6730 }, { "epoch": 93.61, "learning_rate": 3.2676056338028173e-06, "loss": 0.2866, "step": 6740 }, { "epoch": 93.75, "learning_rate": 3.1971830985915496e-06, "loss": 0.293, "step": 6750 }, { "epoch": 93.89, "learning_rate": 3.1267605633802823e-06, "loss": 0.277, "step": 6760 }, { "epoch": 94.0, "eval_loss": 0.4849061965942383, "eval_runtime": 124.6049, "eval_samples_per_second": 4.269, "eval_steps_per_second": 0.538, "eval_wer": 0.19939577039274925, "step": 6768 }, { "epoch": 94.03, "learning_rate": 3.056338028169014e-06, "loss": 0.288, "step": 6770 }, { "epoch": 94.17, "learning_rate": 2.985915492957747e-06, "loss": 0.2819, "step": 6780 }, { "epoch": 94.3, "learning_rate": 2.9154929577464788e-06, "loss": 0.2869, "step": 6790 }, { "epoch": 94.44, "learning_rate": 2.8450704225352115e-06, "loss": 0.2735, "step": 6800 }, { "epoch": 94.58, "learning_rate": 2.7746478873239438e-06, "loss": 0.2725, "step": 6810 }, { "epoch": 94.72, "learning_rate": 2.7042253521126765e-06, "loss": 0.2807, "step": 6820 }, { "epoch": 94.86, "learning_rate": 2.6338028169014084e-06, "loss": 0.2826, "step": 6830 }, { "epoch": 95.0, "learning_rate": 2.563380281690141e-06, "loss": 0.2883, "step": 6840 }, { "epoch": 95.0, "eval_loss": 0.48485061526298523, "eval_runtime": 125.1761, "eval_samples_per_second": 4.25, "eval_steps_per_second": 0.535, "eval_wer": 0.20029568682908017, "step": 6840 }, { "epoch": 95.14, "learning_rate": 2.4929577464788734e-06, "loss": 0.2703, "step": 6850 }, { "epoch": 95.28, "learning_rate": 2.422535211267606e-06, "loss": 0.2825, "step": 6860 }, { "epoch": 95.42, "learning_rate": 2.352112676056338e-06, "loss": 0.2778, "step": 6870 }, { "epoch": 95.55, "learning_rate": 2.2816901408450707e-06, "loss": 0.2909, "step": 6880 }, { "epoch": 95.69, "learning_rate": 2.211267605633803e-06, "loss": 0.2784, "step": 6890 }, { "epoch": 95.83, "learning_rate": 2.1408450704225353e-06, "loss": 0.2789, "step": 6900 }, { "epoch": 95.97, "learning_rate": 2.0704225352112676e-06, "loss": 0.2906, "step": 6910 }, { "epoch": 96.0, "eval_loss": 0.4912075102329254, "eval_runtime": 125.8417, "eval_samples_per_second": 4.228, "eval_steps_per_second": 0.532, "eval_wer": 0.20016712733817574, "step": 6912 }, { "epoch": 96.11, "learning_rate": 2.0000000000000003e-06, "loss": 0.2822, "step": 6920 }, { "epoch": 96.25, "learning_rate": 1.9295774647887326e-06, "loss": 0.2836, "step": 6930 }, { "epoch": 96.39, "learning_rate": 1.8591549295774647e-06, "loss": 0.2685, "step": 6940 }, { "epoch": 96.53, "learning_rate": 1.7887323943661972e-06, "loss": 0.2838, "step": 6950 }, { "epoch": 96.66, "learning_rate": 1.7183098591549297e-06, "loss": 0.2775, "step": 6960 }, { "epoch": 96.8, "learning_rate": 1.6478873239436623e-06, "loss": 0.2867, "step": 6970 }, { "epoch": 96.94, "learning_rate": 1.5774647887323943e-06, "loss": 0.2781, "step": 6980 }, { "epoch": 97.0, "eval_loss": 0.4874226152896881, "eval_runtime": 125.3753, "eval_samples_per_second": 4.243, "eval_steps_per_second": 0.534, "eval_wer": 0.2015812817381243, "step": 6984 }, { "epoch": 97.08, "learning_rate": 1.5070422535211269e-06, "loss": 0.2772, "step": 6990 }, { "epoch": 97.22, "learning_rate": 1.4366197183098591e-06, "loss": 0.284, "step": 7000 }, { "epoch": 97.36, "learning_rate": 1.3661971830985917e-06, "loss": 0.2733, "step": 7010 }, { "epoch": 97.5, "learning_rate": 1.295774647887324e-06, "loss": 0.2783, "step": 7020 }, { "epoch": 97.64, "learning_rate": 1.2253521126760565e-06, "loss": 0.2632, "step": 7030 }, { "epoch": 97.78, "learning_rate": 1.1549295774647888e-06, "loss": 0.2841, "step": 7040 }, { "epoch": 97.91, "learning_rate": 1.0845070422535213e-06, "loss": 0.2836, "step": 7050 }, { "epoch": 98.0, "eval_loss": 0.4825427532196045, "eval_runtime": 126.7624, "eval_samples_per_second": 4.197, "eval_steps_per_second": 0.529, "eval_wer": 0.19958860962910588, "step": 7056 }, { "epoch": 98.06, "learning_rate": 1.0140845070422536e-06, "loss": 0.2835, "step": 7060 }, { "epoch": 98.19, "learning_rate": 9.43661971830986e-07, "loss": 0.2791, "step": 7070 }, { "epoch": 98.33, "learning_rate": 8.732394366197183e-07, "loss": 0.2818, "step": 7080 }, { "epoch": 98.47, "learning_rate": 8.028169014084508e-07, "loss": 0.2848, "step": 7090 }, { "epoch": 98.61, "learning_rate": 7.323943661971831e-07, "loss": 0.2723, "step": 7100 }, { "epoch": 98.75, "learning_rate": 6.619718309859155e-07, "loss": 0.2823, "step": 7110 }, { "epoch": 98.89, "learning_rate": 5.915492957746479e-07, "loss": 0.2676, "step": 7120 }, { "epoch": 99.0, "eval_loss": 0.4805923402309418, "eval_runtime": 125.4559, "eval_samples_per_second": 4.241, "eval_steps_per_second": 0.534, "eval_wer": 0.19946005013820145, "step": 7128 }, { "epoch": 99.03, "learning_rate": 5.211267605633803e-07, "loss": 0.3002, "step": 7130 }, { "epoch": 99.17, "learning_rate": 4.507042253521127e-07, "loss": 0.2796, "step": 7140 }, { "epoch": 99.3, "learning_rate": 3.8028169014084507e-07, "loss": 0.2774, "step": 7150 }, { "epoch": 99.44, "learning_rate": 3.0985915492957747e-07, "loss": 0.2858, "step": 7160 }, { "epoch": 99.58, "learning_rate": 2.394366197183099e-07, "loss": 0.2805, "step": 7170 }, { "epoch": 99.72, "learning_rate": 1.6901408450704228e-07, "loss": 0.2736, "step": 7180 }, { "epoch": 99.86, "learning_rate": 9.859154929577465e-08, "loss": 0.2705, "step": 7190 }, { "epoch": 100.0, "learning_rate": 2.8169014084507045e-08, "loss": 0.28, "step": 7200 }, { "epoch": 100.0, "eval_loss": 0.4813876152038574, "eval_runtime": 126.071, "eval_samples_per_second": 4.22, "eval_steps_per_second": 0.531, "eval_wer": 0.19939577039274925, "step": 7200 }, { "epoch": 100.0, "step": 7200, "total_flos": 2.7202925310621965e+20, "train_loss": 0.3457774919602606, "train_runtime": 189878.6422, "train_samples_per_second": 2.432, "train_steps_per_second": 0.038 } ], "max_steps": 7200, "num_train_epochs": 100, "total_flos": 2.7202925310621965e+20, "trial_name": null, "trial_params": null }