accento-v2.0 / README.md
DelosLogic's picture
Upload folder using huggingface_hub
99dfa31 verified
metadata
language:
  - en
tags:
  - whisper
  - speech-recognition
  - trinidadian-creole
  - accent
  - asr
license: mit
model-index:
  - name: Accento V2.0
    results:
      - task:
          type: automatic-speech-recognition
        metrics:
          - name: WER
            type: wer
            value: 19.94
          - name: CER
            type: cer
            value: 10

Accento V2.0 - Trinidadian Creole English ASR

Accento V2.0 is a fine-tuned Whisper Large V3 Turbo model optimized for Trinidadian Creole English.

Performance

  • WER: 19.94% (with beam_size=3)
  • CER: 10.00%
  • 54% better than base Whisper
  • 27% better than Accento V1.0

Usage

from accento import AccentoTranscriber

# Auto-downloads from Hugging Face if not found locally
transcriber = AccentoTranscriber(model_path="models/accento-v2.0")
result = transcriber.transcribe("audio.wav")
print(result.text)

Technical Details

  • Base: Whisper Large V3 Turbo (809M params)
  • Method: LoRA (rank=32, alpha=64)
  • Adapters: ~106M parameters
  • Training: 179 labeled samples + iterative training + model soups

License

MIT License