Whisper to Praat TextGrid

[jhdeov]

Howdy!

I made a colab file that can take as do long-form transcription from speech to Praat TextGrids. The script takes as input:

• a folder of untranscribed audio files,
• the name of a Whisper model from Hugging Face
• name of a language

It'll generate Praat TextGrids. I've managed to use it so far to help people do long-form transcriptions of Turkish, Portuguese, and Armenian corpora.

https://github.com/jhdeov/whisper-to-textgrid-batch/