argmaxinc / WhisperKit

On-device Speech Recognition for Apple Silicon
https://takeargmax.com/blog/whisperkit
MIT License
3.1k stars 257 forks source link

VAD: First time loading a file it works, second and third time loading the same files it just blanks out #152

Closed iandundas closed 3 months ago

iandundas commented 3 months ago

v0.7.1

With VAD enabled, sometimes it only works the first time.

Machine: M3 Pro

Full video: http://172.104.253.215/CleanShot-2024-05-28-at-13.50.55.mov File used: http://172.104.253.215/atp-7-min-clip.m4a

iandundas commented 3 months ago

This is now fixed 🥳