CoolFace
Modelpublic

akera/whisper-medium-sb-lug-eng_archive

sourceHugging Faceapache-2.0updated 2y agoView on Hugging Face
0likes4downloads
Model Card

<!-- This model card has been generated automatically according to the information the Trainer had access to. You should probably proofread and complete it, then remove this comment. -->

whisper-medium-sb-lug-eng

This model is a fine-tuned version of openai/whisper-medium on the generator dataset. It achieves the following results on the evaluation set:

  • Loss: 0.1064
  • Wer Lug: 0.239
  • Wer Eng: 0.147
  • Wer Mean: 0.193
  • Cer Lug: 0.075
  • Cer Eng: 0.075
  • Cer Mean: 0.075

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:

  • learning_rate: 1e-05
  • trainbatchsize: 16
  • evalbatchsize: 8
  • seed: 42
  • optimizer: Adam with betas=(0.9,0.999) and epsilon=1e-08
  • lrschedulertype: linear
  • lrschedulerwarmup_steps: 500
  • training_steps: 12000
  • mixedprecisiontraining: Native AMP

Training results

Training LossEpochStepValidation LossWer LugWer EngWer MeanCer LugCer EngCer Mean
0.68020.04175000.24810.4970.1510.3240.1240.0780.101
0.5260.083310000.18140.3620.3740.3680.1030.2360.17
0.45510.12515000.16490.3050.1350.220.0870.0660.076
0.39820.166720000.15220.290.1360.2130.0840.0720.078
0.37810.208325000.14360.2610.1510.2060.0820.0790.081
0.37490.2530000.13740.2670.2560.2620.0820.1510.117
0.33970.291735000.13580.2720.1610.2170.0850.0940.09
0.3210.333340000.12860.2590.1350.1970.080.0650.073
0.30140.37545000.12710.2520.1780.2150.0780.1390.108
0.31240.416750000.12410.2520.1320.1920.0770.0630.07
0.26681.038355000.11900.2490.1280.1880.0760.0620.069
0.23681.0860000.12350.2480.360.3040.0780.240.159
0.21881.121765000.11740.2460.1310.1890.0770.0650.071
0.22261.163370000.11390.2420.130.1860.0750.0620.069
0.22681.20575000.11580.2410.1360.1880.0770.0660.072
0.2411.246780000.11440.2460.2060.2260.0770.1690.123
0.23151.288385000.10910.2280.1420.1850.0740.0740.074
0.19321.3390000.11060.2350.1280.1810.0750.0610.068
0.23341.371795000.10910.2390.1380.1890.0740.0720.073
0.19681.4133100000.10720.2370.1290.1830.0740.0630.069
0.15732.035105000.10690.2370.1480.1920.0750.0760.076
0.15272.0767110000.10680.2350.1460.1910.0730.0760.075
0.1682.1183115000.10610.2350.1450.190.0740.0730.074
0.14922.16120000.10640.2390.1470.1930.0750.0750.075

Framework versions

  • Transformers 4.42.3
  • Pytorch 2.2.0
  • Datasets 2.20.0
  • Tokenizers 0.19.1