Wrapper for Distil-Whisper https://github.com/huggingface/distil-whisper
The wrapper takes miff file which refers to either an AudioDocument or a VideoDocument. You can choose the specific model you want to use by add parameter. Four models are available: 'distil-large-v3', 'distil-large-v2', 'distil-medium.en', 'distil-small.en'. The default model is 'distil-small.en'.
The output miff file will contain three objects: Uri.SENTENCE, which is the smallest text unit recognized by the distil-whisper; AnnotationTypes.TimeFrame, which represents the timeframe that each sentences take place; DocumentTypes.TextDocument, which contains all the text recognized by distil-whisper in the whole audio/video; AnnotationTypes.BoundingBox, which show the alignments between Timeframe <-> SENTENCE, and auodio/video <-> TextDocument.
General user instructions for CLAMS apps are available at CLAMS Apps documentation: https://apps.clams.ai/clamsapp/
- Requires numpy<2
- Requires python library transformers and accelerate
- Requires ffmpeg-python for the VideoDocument