whisper over vllm. 30 seconds transcriptions
#232
by jxadro - opened
Hi.
Is there any way to transcribe audios longer than 30 seconds when using whisper over vllm?
More than manually split the audio in chunks, the issue with this approach is that you must to manually handle the overlap and later join of the transcriptions.
I'm looking if there is something out of the box to transcribe more than 30 seconds using the model or vllm.
Thank you.