ales's picture
update model card README.md
9c1d071
|
raw
history blame
2.42 kB
metadata
language:
  - be
license: apache-2.0
tags:
  - whisper-event
  - generated_from_trainer
datasets:
  - mozilla-foundation/common_voice_11_0
metrics:
  - wer
model-index:
  - name: Whisper Tiny Belarusian
    results:
      - task:
          name: Automatic Speech Recognition
          type: automatic-speech-recognition
        dataset:
          name: mozilla-foundation/common_voice_11_0 be
          type: mozilla-foundation/common_voice_11_0
          config: be
          split: validation
          args: be
        metrics:
          - name: Wer
            type: wer
            value: 60.07326007326007

Whisper Tiny Belarusian

This model is a fine-tuned version of openai/whisper-tiny on the mozilla-foundation/common_voice_11_0 be dataset. It achieves the following results on the evaluation set:

  • Loss: 0.6389
  • Wer: 60.0733

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: 0.0001
  • train_batch_size: 32
  • eval_batch_size: 32
  • seed: 42
  • optimizer: Adam with betas=(0.9,0.999) and epsilon=1e-08
  • lr_scheduler_type: linear
  • lr_scheduler_warmup_steps: 10
  • training_steps: 100
  • mixed_precision_training: Native AMP

Training results

Training Loss Epoch Step Validation Loss Wer
2.5622 0.1 10 1.5402 94.5055
1.3719 0.2 20 1.0012 75.2747
0.9898 0.3 30 0.8217 72.7106
0.9742 0.4 40 0.7924 72.5275
0.6951 0.5 50 0.7628 76.1905
0.7824 0.6 60 0.6738 65.3846
0.6818 0.7 70 0.6389 60.0733
0.7823 0.8 80 0.6208 65.7509
0.5994 0.9 90 0.5901 61.9048
0.6647 1.0 100 0.5790 61.7216

Framework versions

  • Transformers 4.26.0.dev0
  • Pytorch 1.13.0+cu116
  • Datasets 2.7.1.dev0
  • Tokenizers 0.13.2