whisper-small-dv / README.md
Adbhut's picture
Create README.md
fdf5650
metadata
license: apache-2.0
datasets:
  - mozilla-foundation/common_voice_13_0
language:
  - dv
metrics:
  - wer
pipeline_tag: automatic-speech-recognition

This is whisper-small finetuned for 500 steps on Common Voice 13, lang=Divehi. Achieves normalized wer=12.66 on test set.