Fix: 30s issue during inference
#1
by
abar-uwc
- opened
Use return_timestamps argument and remove chunk_length_s this will work for longer duration of audio.
i am running on 1 hour hindi audio, still i get some missing transcriptions,