Skip to content

Instantly share code, notes, and snippets.

@jacobkahn
Last active December 3, 2019 22:32
Show Gist options
  • Save jacobkahn/ecf18371f52332dc978c5e713f2b677c to your computer and use it in GitHub Desktop.
Save jacobkahn/ecf18371f52332dc978c5e713f2b677c to your computer and use it in GitHub Desktop.
wav2letter Conv + GLU model with ArrayFire 3.6.4 and master (pre 3.7)
epoch: 1 | lr: 0.800000 | lrcriterion: 0.008000 | runtime: 03:03:23 | bch(ms): 1251.95 | smp(ms): 0.64 | fwd(ms): 663.27 | crit-fwd(ms): 30.43 | bwd(ms): 520.30 | optim(ms): 45.35 | loss: 27.04214 | train-LER: 99.35 | train-WER: 99.79 | dev-clean.lst-loss: 19.76846 | dev-clean.lst-LER: 98.35 | dev-clean.lst-WER: 99.76 | dev-other.lst-loss: 18.46809 | dev-other.lst-LER: 98.13 | dev-other.lst-WER: 99.66 | avg-isz: 1228 | avg-tsz: 206 | max-tsz: 430 | hrs: 959.74 | thrpt(sec/sec): 314.00
epoch: 2 | lr: 0.800000 | lrcriterion: 0.008000 | runtime: 03:04:39 | bch(ms): 1260.59 | smp(ms): 0.62 | fwd(ms): 672.51 | crit-fwd(ms): 31.58 | bwd(ms): 519.22 | optim(ms): 45.44 | loss: 10.13135 | train-LER: 27.68 | train-WER: 58.20 | dev-clean.lst-loss: 2.91011 | dev-clean.lst-LER: 10.81 | dev-clean.lst-WER: 29.63 | dev-other.lst-loss: 5.84437 | dev-other.lst-LER: 23.03 | dev-other.lst-WER: 50.14 | avg-isz: 1228 | avg-tsz: 206 | max-tsz: 430 | hrs: 959.74 | thrpt(sec/sec): 311.85
epoch: 3 | lr: 0.800000 | lrcriterion: 0.008000 | runtime: 03:04:41 | bch(ms): 1260.81 | smp(ms): 0.63 | fwd(ms): 669.45 | crit-fwd(ms): 31.75 | bwd(ms): 522.41 | optim(ms): 45.44 | loss: 4.99938 | train-LER: 13.42 | train-WER: 34.03 | dev-clean.lst-loss: 2.04508 | dev-clean.lst-LER: 7.84 | dev-clean.lst-WER: 21.67 | dev-other.lst-loss: 4.83067 | dev-other.lst-LER: 19.46 | dev-other.lst-WER: 42.46 | avg-isz: 1228 | avg-tsz: 206 | max-tsz: 430 | hrs: 959.74 | thrpt(sec/sec): 311.79
epoch: 4 | lr: 0.800000 | lrcriterion: 0.008000 | runtime: 03:04:44 | bch(ms): 1261.24 | smp(ms): 0.63 | fwd(ms): 672.73 | crit-fwd(ms): 32.21 | bwd(ms): 519.28 | optim(ms): 45.59 | loss: 3.86850 | train-LER: 10.45 | train-WER: 27.25 | dev-clean.lst-loss: 1.61716 | dev-clean.lst-LER: 6.17 | dev-clean.lst-WER: 17.62 | dev-other.lst-loss: 4.14080 | dev-other.lst-LER: 16.65 | dev-other.lst-WER: 37.21 | avg-isz: 1228 | avg-tsz: 206 | max-tsz: 430 | hrs: 959.74 | thrpt(sec/sec): 311.69
epoch: 5 | lr: 0.800000 | lrcriterion: 0.008000 | runtime: 03:04:58 | bch(ms): 1262.77 | smp(ms): 0.62 | fwd(ms): 671.08 | crit-fwd(ms): 32.90 | bwd(ms): 522.78 | optim(ms): 45.10 | loss: 3.27247 | train-LER: 8.90 | train-WER: 23.60 | dev-clean.lst-loss: 1.41295 | dev-clean.lst-LER: 5.19 | dev-clean.lst-WER: 14.61 | dev-other.lst-loss: 3.77897 | dev-other.lst-LER: 14.79 | dev-other.lst-WER: 33.28 | avg-isz: 1228 | avg-tsz: 206 | max-tsz: 430 | hrs: 959.74 | thrpt(sec/sec): 311.31
epoch: 6 | lr: 0.800000 | lrcriterion: 0.008000 | runtime: 03:04:13 | bch(ms): 1257.62 | smp(ms): 0.62 | fwd(ms): 665.01 | crit-fwd(ms): 31.51 | bwd(ms): 523.39 | optim(ms): 45.04 | loss: 2.89158 | train-LER: 7.92 | train-WER: 21.29 | dev-clean.lst-loss: 1.26482 | dev-clean.lst-LER: 4.61 | dev-clean.lst-WER: 13.22 | dev-other.lst-loss: 3.59863 | dev-other.lst-LER: 13.78 | dev-other.lst-WER: 31.21 | avg-isz: 1228 | avg-tsz: 206 | max-tsz: 430 | hrs: 959.74 | thrpt(sec/sec): 312.58
epoch: 7 | lr: 0.800000 | lrcriterion: 0.008000 | runtime: 03:04:16 | bch(ms): 1258.00 | smp(ms): 0.62 | fwd(ms): 671.29 | crit-fwd(ms): 32.88 | bwd(ms): 517.52 | optim(ms): 45.23 | loss: 2.61798 | train-LER: 7.21 | train-WER: 19.59 | dev-clean.lst-loss: 1.17970 | dev-clean.lst-LER: 4.37 | dev-clean.lst-WER: 12.55 | dev-other.lst-loss: 3.39625 | dev-other.lst-LER: 13.18 | dev-other.lst-WER: 29.92 | avg-isz: 1228 | avg-tsz: 206 | max-tsz: 430 | hrs: 959.74 | thrpt(sec/sec): 312.49
epoch: 8 | lr: 0.800000 | lrcriterion: 0.008000 | runtime: 03:03:41 | bch(ms): 1254.00 | smp(ms): 0.62 | fwd(ms): 665.50 | crit-fwd(ms): 32.34 | bwd(ms): 519.94 | optim(ms): 44.99 | loss: 2.40901 | train-LER: 6.68 | train-WER: 18.31 | dev-clean.lst-loss: 1.12136 | dev-clean.lst-LER: 4.12 | dev-clean.lst-WER: 11.73 | dev-other.lst-loss: 3.28954 | dev-other.lst-LER: 12.69 | dev-other.lst-WER: 28.77 | avg-isz: 1228 | avg-tsz: 206 | max-tsz: 430 | hrs: 959.74 | thrpt(sec/sec): 313.49
epoch: 9 | lr: 0.800000 | lrcriterion: 0.008000 | runtime: 03:04:14 | bch(ms): 1257.77 | smp(ms): 0.62 | fwd(ms): 668.32 | crit-fwd(ms): 31.76 | bwd(ms): 520.69 | optim(ms): 45.15 | loss: 2.24253 | train-LER: 6.25 | train-WER: 17.28 | dev-clean.lst-loss: 1.07121 | dev-clean.lst-LER: 3.89 | dev-clean.lst-WER: 11.34 | dev-other.lst-loss: 3.19273 | dev-other.lst-LER: 12.06 | dev-other.lst-WER: 27.46 | avg-isz: 1228 | avg-tsz: 206 | max-tsz: 430 | hrs: 959.74 | thrpt(sec/sec): 312.55
epoch: 10 | lr: 0.800000 | lrcriterion: 0.008000 | runtime: 03:04:20 | bch(ms): 1258.44 | smp(ms): 0.62 | fwd(ms): 667.23 | crit-fwd(ms): 31.99 | bwd(ms): 522.26 | optim(ms): 45.13 | loss: 2.10431 | train-LER: 5.90 | train-WER: 16.42 | dev-clean.lst-loss: 1.02946 | dev-clean.lst-LER: 3.74 | dev-clean.lst-WER: 10.93 | dev-other.lst-loss: 3.22510 | dev-other.lst-LER: 12.23 | dev-other.lst-WER: 27.47 | avg-isz: 1228 | avg-tsz: 206 | max-tsz: 430 | hrs: 959.74 | thrpt(sec/sec): 312.38
epoch: 11 | lr: 0.800000 | lrcriterion: 0.008000 | runtime: 03:05:08 | bch(ms): 1263.90 | smp(ms): 0.62 | fwd(ms): 673.68 | crit-fwd(ms): 31.88 | bwd(ms): 521.04 | optim(ms): 45.49 | loss: 1.98603 | train-LER: 5.59 | train-WER: 15.68 | dev-clean.lst-loss: 1.01398 | dev-clean.lst-LER: 3.62 | dev-clean.lst-WER: 10.34 | dev-other.lst-loss: 3.13359 | dev-other.lst-LER: 11.46 | dev-other.lst-WER: 25.96 | avg-isz: 1228 | avg-tsz: 206 | max-tsz: 430 | hrs: 959.74 | thrpt(sec/sec): 311.03
epoch: 12 | lr: 0.800000 | lrcriterion: 0.008000 | runtime: 03:04:10 | bch(ms): 1257.35 | smp(ms): 0.62 | fwd(ms): 669.35 | crit-fwd(ms): 32.28 | bwd(ms): 519.25 | optim(ms): 44.87 | loss: 1.88676 | train-LER: 5.35 | train-WER: 15.07 | dev-clean.lst-loss: 0.97230 | dev-clean.lst-LER: 3.56 | dev-clean.lst-WER: 10.31 | dev-other.lst-loss: 3.04260 | dev-other.lst-LER: 11.29 | dev-other.lst-WER: 25.68 | avg-isz: 1228 | avg-tsz: 206 | max-tsz: 430 | hrs: 959.74 | thrpt(sec/sec): 312.65
epoch: 13 | lr: 0.800000 | lrcriterion: 0.008000 | runtime: 03:04:16 | bch(ms): 1257.94 | smp(ms): 0.62 | fwd(ms): 672.31 | crit-fwd(ms): 33.85 | bwd(ms): 516.59 | optim(ms): 45.17 | loss: 1.80164 | train-LER: 5.12 | train-WER: 14.52 | dev-clean.lst-loss: 0.92885 | dev-clean.lst-LER: 3.39 | dev-clean.lst-WER: 9.79 | dev-other.lst-loss: 3.03917 | dev-other.lst-LER: 11.10 | dev-other.lst-WER: 25.27 | avg-isz: 1228 | avg-tsz: 206 | max-tsz: 430 | hrs: 959.74 | thrpt(sec/sec): 312.51
epoch: 14 | lr: 0.800000 | lrcriterion: 0.008000 | runtime: 03:03:33 | bch(ms): 1253.10 | smp(ms): 0.62 | fwd(ms): 666.19 | crit-fwd(ms): 31.69 | bwd(ms): 518.64 | optim(ms): 44.76 | loss: 1.72611 | train-LER: 4.92 | train-WER: 14.02 | dev-clean.lst-loss: 0.91112 | dev-clean.lst-LER: 3.34 | dev-clean.lst-WER: 9.68 | dev-other.lst-loss: 3.01247 | dev-other.lst-LER: 11.13 | dev-other.lst-WER: 25.28 | avg-isz: 1228 | avg-tsz: 206 | max-tsz: 430 | hrs: 959.74 | thrpt(sec/sec): 313.71
epoch: 15 | lr: 0.800000 | lrcriterion: 0.008000 | runtime: 03:03:52 | bch(ms): 1255.27 | smp(ms): 0.63 | fwd(ms): 669.28 | crit-fwd(ms): 32.56 | bwd(ms): 517.39 | optim(ms): 44.98 | loss: 1.65832 | train-LER: 4.74 | train-WER: 13.59 | dev-clean.lst-loss: 0.90917 | dev-clean.lst-LER: 3.31 | dev-clean.lst-WER: 9.73 | dev-other.lst-loss: 2.94840 | dev-other.lst-LER: 10.98 | dev-other.lst-WER: 25.17 | avg-isz: 1228 | avg-tsz: 206 | max-tsz: 430 | hrs: 959.74 | thrpt(sec/sec): 313.17
epoch: 16 | lr: 0.800000 | lrcriterion: 0.008000 | runtime: 03:04:02 | bch(ms): 1256.43 | smp(ms): 0.62 | fwd(ms): 669.84 | crit-fwd(ms): 32.11 | bwd(ms): 517.97 | optim(ms): 45.00 | loss: 1.59856 | train-LER: 4.60 | train-WER: 13.23 | dev-clean.lst-loss: 0.90703 | dev-clean.lst-LER: 3.29 | dev-clean.lst-WER: 9.48 | dev-other.lst-loss: 2.90936 | dev-other.lst-LER: 10.79 | dev-other.lst-WER: 24.53 | avg-isz: 1228 | avg-tsz: 206 | max-tsz: 430 | hrs: 959.74 | thrpt(sec/sec): 312.88
epoch: 17 | lr: 0.800000 | lrcriterion: 0.008000 | runtime: 03:04:35 | bch(ms): 1260.13 | smp(ms): 0.62 | fwd(ms): 670.75 | crit-fwd(ms): 33.55 | bwd(ms): 520.39 | optim(ms): 45.14 | loss: 1.54446 | train-LER: 4.45 | train-WER: 12.86 | dev-clean.lst-loss: 0.88833 | dev-clean.lst-LER: 3.17 | dev-clean.lst-WER: 9.17 | dev-other.lst-loss: 2.94638 | dev-other.lst-LER: 10.68 | dev-other.lst-WER: 24.36 | avg-isz: 1228 | avg-tsz: 206 | max-tsz: 430 | hrs: 959.74 | thrpt(sec/sec): 311.96
epoch: 18 | lr: 0.800000 | lrcriterion: 0.008000 | runtime: 03:04:21 | bch(ms): 1258.61 | smp(ms): 0.63 | fwd(ms): 670.32 | crit-fwd(ms): 32.01 | bwd(ms): 519.56 | optim(ms): 44.89 | loss: 1.49296 | train-LER: 4.32 | train-WER: 12.53 | dev-clean.lst-loss: 0.87383 | dev-clean.lst-LER: 3.07 | dev-clean.lst-WER: 8.74 | dev-other.lst-loss: 2.89175 | dev-other.lst-LER: 10.38 | dev-other.lst-WER: 23.65 | avg-isz: 1228 | avg-tsz: 206 | max-tsz: 430 | hrs: 959.74 | thrpt(sec/sec): 312.34
epoch: 19 | lr: 0.800000 | lrcriterion: 0.008000 | runtime: 03:03:49 | bch(ms): 1254.96 | smp(ms): 0.63 | fwd(ms): 669.52 | crit-fwd(ms): 32.51 | bwd(ms): 516.70 | optim(ms): 45.21 | loss: 1.44668 | train-LER: 4.20 | train-WER: 12.22 | dev-clean.lst-loss: 0.85853 | dev-clean.lst-LER: 3.02 | dev-clean.lst-WER: 8.75 | dev-other.lst-loss: 2.94300 | dev-other.lst-LER: 10.55 | dev-other.lst-WER: 23.87 | avg-isz: 1228 | avg-tsz: 206 | max-tsz: 430 | hrs: 959.74 | thrpt(sec/sec): 313.25
epoch: 20 | lr: 0.800000 | lrcriterion: 0.008000 | runtime: 03:04:18 | bch(ms): 1258.20 | smp(ms): 0.62 | fwd(ms): 667.41 | crit-fwd(ms): 31.91 | bwd(ms): 521.96 | optim(ms): 45.01 | loss: 1.40701 | train-LER: 4.10 | train-WER: 11.96 | dev-clean.lst-loss: 0.86443 | dev-clean.lst-LER: 3.03 | dev-clean.lst-WER: 8.66 | dev-other.lst-loss: 2.93447 | dev-other.lst-LER: 10.40 | dev-other.lst-WER: 23.62 | avg-isz: 1228 | avg-tsz: 206 | max-tsz: 430 | hrs: 959.74 | thrpt(sec/sec): 312.44
epoch: 21 | lr: 0.800000 | lrcriterion: 0.008000 | runtime: 03:04:40 | bch(ms): 1260.68 | smp(ms): 0.62 | fwd(ms): 674.08 | crit-fwd(ms): 32.03 | bwd(ms): 517.41 | optim(ms): 45.40 | loss: 1.36579 | train-LER: 3.99 | train-WER: 11.70 | dev-clean.lst-loss: 0.83327 | dev-clean.lst-LER: 3.04 | dev-clean.lst-WER: 8.75 | dev-other.lst-loss: 2.88455 | dev-other.lst-LER: 10.46 | dev-other.lst-WER: 23.64 | avg-isz: 1228 | avg-tsz: 206 | max-tsz: 430 | hrs: 959.74 | thrpt(sec/sec): 311.83
epoch: 22 | lr: 0.800000 | lrcriterion: 0.008000 | runtime: 03:04:08 | bch(ms): 1257.12 | smp(ms): 0.62 | fwd(ms): 668.44 | crit-fwd(ms): 31.57 | bwd(ms): 519.78 | optim(ms): 45.29 | loss: 1.33089 | train-LER: 3.90 | train-WER: 11.46 | dev-clean.lst-loss: 0.84243 | dev-clean.lst-LER: 2.95 | dev-clean.lst-WER: 8.58 | dev-other.lst-loss: 2.80818 | dev-other.lst-LER: 10.15 | dev-other.lst-WER: 23.15 | avg-isz: 1228 | avg-tsz: 206 | max-tsz: 430 | hrs: 959.74 | thrpt(sec/sec): 312.71
epoch: 23 | lr: 0.800000 | lrcriterion: 0.008000 | runtime: 03:04:34 | bch(ms): 1260.00 | smp(ms): 0.62 | fwd(ms): 672.96 | crit-fwd(ms): 32.78 | bwd(ms): 518.29 | optim(ms): 45.25 | loss: 1.29956 | train-LER: 3.82 | train-WER: 11.26 | dev-clean.lst-loss: 0.83576 | dev-clean.lst-LER: 2.90 | dev-clean.lst-WER: 8.52 | dev-other.lst-loss: 2.83318 | dev-other.lst-LER: 10.04 | dev-other.lst-WER: 22.98 | avg-isz: 1228 | avg-tsz: 206 | max-tsz: 430 | hrs: 959.74 | thrpt(sec/sec): 311.99
epoch: 1 | lr: 0.800000 | lrcriterion: 0.008000 | runtime: 03:01:26 | bch(ms): 1238.63 | smp(ms): 0.64 | fwd(ms): 656.24 | crit-fwd(ms): 30.26 | bwd(ms): 513.76 | optim(ms): 45.58 | loss: 27.02391 | train-LER: 99.30 | train-WER: 99.79 | dev-clean.lst-loss: 19.64907 | dev-clean.lst-LER: 98.10 | dev-clean.lst-WER: 99.83 | dev-other.lst-loss: 18.35306 | dev-other.lst-LER: 97.89 | dev-other.lst-WER: 99.85 | avg-isz: 1228 | avg-tsz: 206 | max-tsz: 430 | hrs: 959.74 | thrpt(sec/sec): 317.38
epoch: 2 | lr: 0.800000 | lrcriterion: 0.008000 | runtime: 03:02:40 | bch(ms): 1247.10 | smp(ms): 0.64 | fwd(ms): 664.95 | crit-fwd(ms): 31.58 | bwd(ms): 512.98 | optim(ms): 45.64 | loss: 9.92500 | train-LER: 27.08 | train-WER: 57.24 | dev-clean.lst-loss: 2.82845 | dev-clean.lst-LER: 10.58 | dev-clean.lst-WER: 29.06 | dev-other.lst-loss: 5.92070 | dev-other.lst-LER: 23.74 | dev-other.lst-WER: 50.92 | avg-isz: 1228 | avg-tsz: 206 | max-tsz: 430 | hrs: 959.74 | thrpt(sec/sec): 315.22
epoch: 3 | lr: 0.800000 | lrcriterion: 0.008000 | runtime: 03:03:03 | bch(ms): 1249.70 | smp(ms): 0.63 | fwd(ms): 663.72 | crit-fwd(ms): 31.59 | bwd(ms): 516.59 | optim(ms): 45.82 | loss: 4.94224 | train-LER: 13.25 | train-WER: 33.59 | dev-clean.lst-loss: 1.96384 | dev-clean.lst-LER: 7.36 | dev-clean.lst-WER: 20.51 | dev-other.lst-loss: 4.76895 | dev-other.lst-LER: 18.43 | dev-other.lst-WER: 40.61 | avg-isz: 1228 | avg-tsz: 206 | max-tsz: 430 | hrs: 959.74 | thrpt(sec/sec): 314.56
epoch: 4 | lr: 0.800000 | lrcriterion: 0.008000 | runtime: 03:02:41 | bch(ms): 1247.19 | smp(ms): 0.63 | fwd(ms): 664.76 | crit-fwd(ms): 31.80 | bwd(ms): 512.62 | optim(ms): 46.15 | loss: 3.83265 | train-LER: 10.35 | train-WER: 26.97 | dev-clean.lst-loss: 1.58043 | dev-clean.lst-LER: 6.09 | dev-clean.lst-WER: 17.49 | dev-other.lst-loss: 4.11756 | dev-other.lst-LER: 16.64 | dev-other.lst-WER: 37.22 | avg-isz: 1228 | avg-tsz: 206 | max-tsz: 430 | hrs: 959.74 | thrpt(sec/sec): 315.20
epoch: 5 | lr: 0.800000 | lrcriterion: 0.008000 | runtime: 03:03:00 | bch(ms): 1249.30 | smp(ms): 0.62 | fwd(ms): 662.76 | crit-fwd(ms): 32.79 | bwd(ms): 517.22 | optim(ms): 45.61 | loss: 3.24541 | train-LER: 8.83 | train-WER: 23.43 | dev-clean.lst-loss: 1.39817 | dev-clean.lst-LER: 5.17 | dev-clean.lst-WER: 14.65 | dev-other.lst-loss: 3.77073 | dev-other.lst-LER: 14.52 | dev-other.lst-WER: 32.68 | avg-isz: 1228 | avg-tsz: 206 | max-tsz: 430 | hrs: 959.74 | thrpt(sec/sec): 314.67
epoch: 6 | lr: 0.800000 | lrcriterion: 0.008000 | runtime: 03:02:32 | bch(ms): 1246.21 | smp(ms): 0.63 | fwd(ms): 658.60 | crit-fwd(ms): 31.59 | bwd(ms): 518.11 | optim(ms): 45.33 | loss: 2.86858 | train-LER: 7.86 | train-WER: 21.12 | dev-clean.lst-loss: 1.26669 | dev-clean.lst-LER: 4.80 | dev-clean.lst-WER: 13.82 | dev-other.lst-loss: 3.56284 | dev-other.lst-LER: 14.15 | dev-other.lst-WER: 32.08 | avg-isz: 1228 | avg-tsz: 206 | max-tsz: 430 | hrs: 959.74 | thrpt(sec/sec): 315.45
epoch: 7 | lr: 0.800000 | lrcriterion: 0.008000 | runtime: 03:02:24 | bch(ms): 1245.21 | smp(ms): 0.63 | fwd(ms): 663.33 | crit-fwd(ms): 32.46 | bwd(ms): 512.13 | optim(ms): 45.74 | loss: 2.59583 | train-LER: 7.16 | train-WER: 19.45 | dev-clean.lst-loss: 1.17400 | dev-clean.lst-LER: 4.48 | dev-clean.lst-WER: 12.76 | dev-other.lst-loss: 3.40651 | dev-other.lst-LER: 13.49 | dev-other.lst-WER: 30.45 | avg-isz: 1228 | avg-tsz: 206 | max-tsz: 430 | hrs: 959.74 | thrpt(sec/sec): 315.70
epoch: 8 | lr: 0.800000 | lrcriterion: 0.008000 | runtime: 03:01:56 | bch(ms): 1242.03 | smp(ms): 0.63 | fwd(ms): 658.54 | crit-fwd(ms): 32.21 | bwd(ms): 514.34 | optim(ms): 45.57 | loss: 2.38860 | train-LER: 6.63 | train-WER: 18.17 | dev-clean.lst-loss: 1.12483 | dev-clean.lst-LER: 4.16 | dev-clean.lst-WER: 11.97 | dev-other.lst-loss: 3.29756 | dev-other.lst-LER: 12.82 | dev-other.lst-WER: 29.01 | avg-isz: 1228 | avg-tsz: 206 | max-tsz: 430 | hrs: 959.74 | thrpt(sec/sec): 316.51
epoch: 9 | lr: 0.800000 | lrcriterion: 0.008000 | runtime: 03:02:19 | bch(ms): 1244.67 | smp(ms): 0.63 | fwd(ms): 660.19 | crit-fwd(ms): 31.62 | bwd(ms): 515.31 | optim(ms): 45.57 | loss: 2.22416 | train-LER: 6.21 | train-WER: 17.14 | dev-clean.lst-loss: 1.06163 | dev-clean.lst-LER: 3.92 | dev-clean.lst-WER: 11.27 | dev-other.lst-loss: 3.18382 | dev-other.lst-LER: 12.30 | dev-other.lst-WER: 27.97 | avg-isz: 1228 | avg-tsz: 206 | max-tsz: 430 | hrs: 959.74 | thrpt(sec/sec): 315.84
epoch: 10 | lr: 0.800000 | lrcriterion: 0.008000 | runtime: 03:02:32 | bch(ms): 1246.11 | smp(ms): 0.63 | fwd(ms): 659.94 | crit-fwd(ms): 31.75 | bwd(ms): 516.56 | optim(ms): 45.76 | loss: 2.08812 | train-LER: 5.85 | train-WER: 16.29 | dev-clean.lst-loss: 1.03091 | dev-clean.lst-LER: 3.86 | dev-clean.lst-WER: 11.24 | dev-other.lst-loss: 3.16816 | dev-other.lst-LER: 12.24 | dev-other.lst-WER: 27.66 | avg-isz: 1228 | avg-tsz: 206 | max-tsz: 430 | hrs: 959.74 | thrpt(sec/sec): 315.47
epoch: 11 | lr: 0.800000 | lrcriterion: 0.008000 | runtime: 03:03:28 | bch(ms): 1252.58 | smp(ms): 0.62 | fwd(ms): 666.35 | crit-fwd(ms): 31.76 | bwd(ms): 516.60 | optim(ms): 45.90 | loss: 1.97285 | train-LER: 5.56 | train-WER: 15.58 | dev-clean.lst-loss: 0.99396 | dev-clean.lst-LER: 3.64 | dev-clean.lst-WER: 10.38 | dev-other.lst-loss: 3.14254 | dev-other.lst-LER: 11.72 | dev-other.lst-WER: 26.40 | avg-isz: 1228 | avg-tsz: 206 | max-tsz: 430 | hrs: 959.74 | thrpt(sec/sec): 313.84
epoch: 12 | lr: 0.800000 | lrcriterion: 0.008000 | runtime: 03:02:27 | bch(ms): 1245.64 | smp(ms): 0.63 | fwd(ms): 661.86 | crit-fwd(ms): 32.04 | bwd(ms): 514.24 | optim(ms): 45.62 | loss: 1.87504 | train-LER: 5.31 | train-WER: 14.95 | dev-clean.lst-loss: 0.96631 | dev-clean.lst-LER: 3.51 | dev-clean.lst-WER: 10.03 | dev-other.lst-loss: 3.07437 | dev-other.lst-LER: 11.25 | dev-other.lst-WER: 25.48 | avg-isz: 1228 | avg-tsz: 206 | max-tsz: 430 | hrs: 959.74 | thrpt(sec/sec): 315.59
epoch: 13 | lr: 0.800000 | lrcriterion: 0.008000 | runtime: 03:02:33 | bch(ms): 1246.26 | smp(ms): 0.62 | fwd(ms): 664.84 | crit-fwd(ms): 33.64 | bwd(ms): 511.77 | optim(ms): 45.76 | loss: 1.78786 | train-LER: 5.09 | train-WER: 14.41 | dev-clean.lst-loss: 0.93854 | dev-clean.lst-LER: 3.40 | dev-clean.lst-WER: 9.86 | dev-other.lst-loss: 3.06310 | dev-other.lst-LER: 11.24 | dev-other.lst-WER: 25.49 | avg-isz: 1228 | avg-tsz: 206 | max-tsz: 430 | hrs: 959.74 | thrpt(sec/sec): 315.43
epoch: 14 | lr: 0.800000 | lrcriterion: 0.008000 | runtime: 03:01:58 | bch(ms): 1242.27 | smp(ms): 0.63 | fwd(ms): 659.35 | crit-fwd(ms): 31.51 | bwd(ms): 513.93 | optim(ms): 45.46 | loss: 1.71560 | train-LER: 4.90 | train-WER: 13.95 | dev-clean.lst-loss: 0.91734 | dev-clean.lst-LER: 3.35 | dev-clean.lst-WER: 9.64 | dev-other.lst-loss: 2.98421 | dev-other.lst-LER: 10.96 | dev-other.lst-WER: 24.83 | avg-isz: 1228 | avg-tsz: 206 | max-tsz: 430 | hrs: 959.74 | thrpt(sec/sec): 316.45
epoch: 15 | lr: 0.800000 | lrcriterion: 0.008000 | runtime: 03:02:37 | bch(ms): 1246.71 | smp(ms): 0.62 | fwd(ms): 663.88 | crit-fwd(ms): 32.52 | bwd(ms): 513.11 | optim(ms): 46.09 | loss: 1.64549 | train-LER: 4.72 | train-WER: 13.51 | dev-clean.lst-loss: 0.90495 | dev-clean.lst-LER: 3.33 | dev-clean.lst-WER: 9.73 | dev-other.lst-loss: 2.95551 | dev-other.lst-LER: 11.11 | dev-other.lst-WER: 25.32 | avg-isz: 1228 | avg-tsz: 206 | max-tsz: 430 | hrs: 959.74 | thrpt(sec/sec): 315.32
epoch: 16 | lr: 0.800000 | lrcriterion: 0.008000 | runtime: 03:02:40 | bch(ms): 1247.02 | smp(ms): 0.63 | fwd(ms): 665.41 | crit-fwd(ms): 32.28 | bwd(ms): 512.02 | optim(ms): 45.88 | loss: 1.58524 | train-LER: 4.57 | train-WER: 13.12 | dev-clean.lst-loss: 0.88581 | dev-clean.lst-LER: 3.23 | dev-clean.lst-WER: 9.41 | dev-other.lst-loss: 2.97461 | dev-other.lst-LER: 11.08 | dev-other.lst-WER: 25.10 | avg-isz: 1228 | avg-tsz: 206 | max-tsz: 430 | hrs: 959.74 | thrpt(sec/sec): 315.24
epoch: 17 | lr: 0.800000 | lrcriterion: 0.008000 | runtime: 03:02:39 | bch(ms): 1247.00 | smp(ms): 0.63 | fwd(ms): 662.46 | crit-fwd(ms): 33.10 | bwd(ms): 515.31 | optim(ms): 45.43 | loss: 1.53279 | train-LER: 4.43 | train-WER: 12.77 | dev-clean.lst-loss: 0.87535 | dev-clean.lst-LER: 3.16 | dev-clean.lst-WER: 9.17 | dev-other.lst-loss: 2.92725 | dev-other.lst-LER: 10.71 | dev-other.lst-WER: 24.27 | avg-isz: 1228 | avg-tsz: 206 | max-tsz: 430 | hrs: 959.74 | thrpt(sec/sec): 315.25
epoch: 18 | lr: 0.800000 | lrcriterion: 0.008000 | runtime: 03:02:39 | bch(ms): 1247.01 | smp(ms): 0.63 | fwd(ms): 663.09 | crit-fwd(ms): 31.64 | bwd(ms): 514.55 | optim(ms): 45.55 | loss: 1.48162 | train-LER: 4.30 | train-WER: 12.46 | dev-clean.lst-loss: 0.87968 | dev-clean.lst-LER: 3.07 | dev-clean.lst-WER: 8.73 | dev-other.lst-loss: 2.98276 | dev-other.lst-LER: 10.45 | dev-other.lst-WER: 23.58 | avg-isz: 1228 | avg-tsz: 206 | max-tsz: 430 | hrs: 959.74 | thrpt(sec/sec): 315.24
epoch: 19 | lr: 0.800000 | lrcriterion: 0.008000 | runtime: 03:02:06 | bch(ms): 1243.25 | smp(ms): 0.63 | fwd(ms): 661.69 | crit-fwd(ms): 32.40 | bwd(ms): 512.30 | optim(ms): 45.73 | loss: 1.43785 | train-LER: 4.18 | train-WER: 12.15 | dev-clean.lst-loss: 0.85120 | dev-clean.lst-LER: 3.00 | dev-clean.lst-WER: 8.65 | dev-other.lst-loss: 2.91884 | dev-other.lst-LER: 10.46 | dev-other.lst-WER: 23.70 | avg-isz: 1228 | avg-tsz: 206 | max-tsz: 430 | hrs: 959.74 | thrpt(sec/sec): 316.20
epoch: 20 | lr: 0.800000 | lrcriterion: 0.008000 | runtime: 03:02:47 | bch(ms): 1247.82 | smp(ms): 0.62 | fwd(ms): 660.06 | crit-fwd(ms): 31.83 | bwd(ms): 518.12 | optim(ms): 45.82 | loss: 1.39559 | train-LER: 4.07 | train-WER: 11.88 | dev-clean.lst-loss: 0.86887 | dev-clean.lst-LER: 3.03 | dev-clean.lst-WER: 8.71 | dev-other.lst-loss: 2.92101 | dev-other.lst-LER: 10.43 | dev-other.lst-WER: 23.56 | avg-isz: 1228 | avg-tsz: 206 | max-tsz: 430 | hrs: 959.74 | thrpt(sec/sec): 315.04
epoch: 21 | lr: 0.800000 | lrcriterion: 0.008000 | runtime: 03:02:46 | bch(ms): 1247.78 | smp(ms): 0.63 | fwd(ms): 666.16 | crit-fwd(ms): 31.71 | bwd(ms): 511.81 | optim(ms): 46.02 | loss: 1.35781 | train-LER: 3.97 | train-WER: 11.62 | dev-clean.lst-loss: 0.83098 | dev-clean.lst-LER: 3.05 | dev-clean.lst-WER: 8.83 | dev-other.lst-loss: 2.85621 | dev-other.lst-LER: 10.50 | dev-other.lst-WER: 23.90 | avg-isz: 1228 | avg-tsz: 206 | max-tsz: 430 | hrs: 959.74 | thrpt(sec/sec): 315.05
epoch: 22 | lr: 0.800000 | lrcriterion: 0.008000 | runtime: 03:17:17 | bch(ms): 1346.83 | smp(ms): 0.63 | fwd(ms): 746.63 | crit-fwd(ms): 37.09 | bwd(ms): 525.09 | optim(ms): 50.35 | loss: 1.32315 | train-LER: 3.88 | train-WER: 11.39 | dev-clean.lst-loss: 0.83229 | dev-clean.lst-LER: 2.99 | dev-clean.lst-WER: 8.57 | dev-other.lst-loss: 2.86042 | dev-other.lst-LER: 10.28 | dev-other.lst-WER: 23.32 | avg-isz: 1228 | avg-tsz: 206 | max-tsz: 430 | hrs: 959.74 | thrpt(sec/sec): 291.88
epoch: 23 | lr: 0.800000 | lrcriterion: 0.008000 | runtime: 03:33:27 | bch(ms): 1457.27 | smp(ms): 0.63 | fwd(ms): 838.07 | crit-fwd(ms): 43.67 | bwd(ms): 537.11 | optim(ms): 56.20 | loss: 1.28979 | train-LER: 3.80 | train-WER: 11.19 | dev-clean.lst-loss: 0.81627 | dev-clean.lst-LER: 2.93 | dev-clean.lst-WER: 8.55 | dev-other.lst-loss: 2.79074 | dev-other.lst-LER: 10.07 | dev-other.lst-WER: 22.93 | avg-isz: 1228 | avg-tsz: 206 | max-tsz: 430 | hrs: 959.74 | thrpt(sec/sec): 269.76
Sign up for free to join this conversation on GitHub. Already have an account? Sign in to comment