whisper-small-pl / README.md
Aspik101's picture
Update README.md
1536662
metadata
language:
  - pl
license: apache-2.0
tags:
  - whisper-event
  - generated_from_trainer
datasets:
  - mozilla-foundation/common_voice_11_0
metrics:
  - wer
model-index:
  - name: Whisper Small PL
    results:
      - task:
          name: Automatic Speech Recognition
          type: automatic-speech-recognition
        dataset:
          name: Common Voice 11.0
          type: mozilla-foundation/common_voice_11_0
          config: pl
          split: test
          args: pl
        metrics:
          - name: Wer
            type: wer
            value: 14.49425614099627

Whisper Small PL

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: 0.4444
  • Wer: 14.4943

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
  • train_batch_size: 32
  • eval_batch_size: 16
  • seed: 42
  • optimizer: Adam with betas=(0.9,0.999) and epsilon=1e-08
  • lr_scheduler_type: linear
  • lr_scheduler_warmup_steps: 500
  • training_steps: 5000
  • mixed_precision_training: Native AMP

Training results

Training Loss Epoch Step Validation Loss Wer
0.1639 0.1 500 0.3290 16.6413
0.0674 1.1 1000 0.3224 15.1782
0.0335 2.09 1500 0.3186 14.5394
0.0161 3.09 2000 0.3445 15.0026
0.0101 4.08 2500 0.3777 14.5260
0.0064 5.08 3000 0.3977 14.6264
0.0036 6.08 3500 0.4621 14.6180
0.0025 7.07 4000 0.4639 14.5193
0.0017 8.07 4500 0.4971 14.4725
0.0017 9.07 5000 0.4444 14.4943

Framework versions

  • Transformers 4.26.0.dev0
  • Pytorch 1.13.0+cu117
  • Datasets 2.7.1
  • Tokenizers 0.13.2