CoolFace
Modelpublic

taesiri/whisper-small-fa-7

sourceHugging Faceapache-2.0updated 4y agoView on Hugging Face
8likes42downloads
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 Small Fa - BuzzyBuzzy

This model is a fine-tuned version of openai/whisper-small on the Common Voice 11.0 dataset. It achieves the following results on the evaluation set:

  • —Loss: 0.6904
  • —Wer: 34.2597

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: 16
  • —seed: 42
  • —optimizer: Adam with betas=(0.9,0.999) and epsilon=1e-08
  • —lrschedulertype: linear
  • —lrschedulerwarmup_steps: 500
  • —training_steps: 20000
  • —mixedprecisiontraining: Native AMP

Training results

Training LossEpochStepValidation LossWer
0.00990.8620000.611637.0499
0.00791.7240000.634037.3746
0.00692.5860000.615935.9455
0.00153.4480000.618735.3281
0.00094.3100000.644935.0783
0.00075.15120000.644034.8452
0.00026.01140000.666434.3263
0.00016.87160000.681734.3471
0.07.73180000.680934.1737
0.08.59200000.690434.2597

Framework versions

  • —Transformers 4.26.0.dev0
  • —Pytorch 1.13.1+cu117
  • —Datasets 2.8.0
  • —Tokenizers 0.13.2