wav2vec2-large-mms-1b-hindi_2-colab

作者 rashmi035 · 下载 0 · 点赞 0 · 更新 2023-11-03

generated_from_trainer

wav2vec2-large-mms-1b-hindi_2-colab

This model is a fine-tuned version of facebook/mms-1b-fl102 on an unknown dataset. It achieves the following results on the evaluation set:

Model description

More information needed

Intended uses & limitations

More information needed

Training and evaluation data

More information needed

Training procedure

Training hyperparameters

The following hyperparameters were used during training:

Training results

Training Loss Epoch Step Validation Loss Wer
7.5635 0.02 20 7.4070 0.9985
13.6122 0.04 40 14.5202 1.0
10.4272 0.06 60 8.7994 1.5440
8.1195 0.08 80 10.3713 1.0
9.9347 0.1 100 7.1064 1.0
5.752 0.12 120 5.6953 1.0
5.1715 0.14 140 5.0103 1.0
6.3111 0.15 160 4.6935 1.0
4.4929 0.17 180 5.4670 1.0263
6.038 0.19 200 5.6732 1.3148
4.1732 0.21 220 4.2880 1.0015
3.8954 0.23 240 4.3895 1.0
3.9351 0.25 260 3.7766 1.0
3.6591 0.27 280 3.7521 1.0
3.6009 0.29 300 3.8260 1.0
3.5822 0.31 320 3.5655 1.0
3.5705 0.33 340 3.6623 1.0
3.6825 0.35 360 3.5988 1.0
3.5239 0.37 380 3.5307 1.0
3.558 0.39 400 3.5847 1.0
3.4658 0.41 420 3.4300 1.0
3.4045 0.43 440 3.5261 1.0
3.4564 0.44 460 3.4799 1.0
3.4403 0.46 480 3.4126 1.0
3.4733 0.48 500 3.5358 1.0
3.445 0.5 520 3.3526 1.0
3.4155 0.52 540 3.3508 1.0
3.412 0.54 560 3.3205 1.0
3.2547 0.56 580 3.3143 1.0
3.2652 0.58 600 3.3057 1.0
3.1801 0.6 620 3.2361 1.0
3.2835 0.62 640 3.3567 1.0
3.3545 0.64 660 3.2300 1.0
3.1898 0.66 680 3.1771 1.0
3.1109 0.68 700 3.3033 1.0
3.1631 0.7 720 3.0177 0.9997
3.0386 0.71 740 3.0339 0.9997
3.074 0.73 760 3.0702 1.0
2.8598 0.75 780 2.8458 1.0
2.8116 0.77 800 2.9836 0.9995
2.8086 0.79 820 2.5641 1.0
2.6645 0.81 840 2.6182 1.0
2.7035 0.83 860 2.5176 0.9995
2.4736 0.85 880 2.3965 0.9995
2.6259 0.87 900 2.5697 1.0
2.44 0.89 920 2.3085 1.0
2.22 0.91 940 2.1551 0.9997
2.5394 0.93 960 2.1955 1.0
2.1734 0.95 980 2.1015 1.0
2.407 0.97 1000 2.3892 1.0
2.1967 0.99 1020 1.9439 0.9943
2.1704 1.0 1040 1.9236 0.9827
1.9929 1.02 1060 1.9353 0.9964
2.1652 1.04 1080 2.1551 0.9899
2.003 1.06 1100 1.9230 0.9820
2.0048 1.08 1120 1.9293 0.9869
2.1665 1.1 1140 1.8845 0.9990
1.8297 1.12 1160 1.7173 0.9866
1.8388 1.14 1180 1.8550 0.9871
1.8399 1.16 1200 1.7772 0.9789
1.7256 1.18 1220 1.7840 0.9863
2.0516 1.2 1240 1.7693 0.9520
1.8014 1.22 1260 1.6744 0.9814
1.8244 1.24 1280 1.6614 0.9907
1.8233 1.26 1300 1.5975 0.9948
1.6977 1.28 1320 1.5738 0.9874
1.9592 1.29 1340 1.5922 0.9897
1.6181 1.31 1360 1.4764 0.9626
1.6739 1.33 1380 1.5381 0.9928
1.6855 1.35 1400 1.4613 0.9410
1.5535 1.37 1420 1.4878 0.9348
1.7467 1.39 1440 1.6077 0.9618
1.6744 1.41 1460 1.4419 0.9727
1.6115 1.43 1480 1.6700 0.9379
1.7357 1.45 1500 1.5228 0.9964
1.7096 1.47 1520 1.4350 0.9611
1.7402 1.49 1540 1.4351 0.9567
1.4819 1.51 1560 1.4062 0.9727
1.6863 1.53 1580 1.4908 0.9889
1.5539 1.55 1600 1.4099 0.9827
1.5733 1.57 1620 1.4508 0.9209
1.7331 1.58 1640 1.3913 0.9755
1.4361 1.6 1660 1.3525 0.9237
1.4806 1.62 1680 1.3748 0.9557
1.5834 1.64 1700 1.3428 0.9386
1.4226 1.66 1720 1.2990 0.9523
1.6159 1.68 1740 1.3351 0.9428
1.4486 1.7 1760 1.2982 0.9276
1.3682 1.72 1780 1.3810 0.9312
1.3828 1.74 1800 1.2621 0.9242
1.4604 1.76 1820 1.2883 0.9051
1.4368 1.78 1840 1.2462 0.9191
1.3652 1.8 1860 1.2544 0.8935
1.4347 1.82 1880 1.2682 0.9185
1.4109 1.84 1900 1.2385 0.8966
1.251 1.86 1920 1.2293 0.9015
1.4793 1.87 1940 1.2410 0.9075
1.2481 1.89 1960 1.1916 0.9134
1.2951 1.91 1980 1.2061 0.8891
1.3724 1.93 2000 1.1730 0.9381
1.3093 1.95 2020 1.1763 0.8951
1.3305 1.97 2040 1.1709 0.9028
1.3152 1.99 2060 1.1619 0.9015

Framework versions