search

NavodPeiris / speechlib

speechlib is a library that can do speaker diarization, transcription and speaker recognition on an audio file to create transcripts with actual speaker names
MIT License
138 stars 12 forks source link

Update core_analysis.py #1

Closed mxzgithub closed 8 months ago

mxzgithub commented 8 months ago

added torch and torchaudio to speed up the process

NavodPeiris commented 8 months ago

thanks again!

in this code, first you should check if cuda is available. then, set the device accordingly. because in some machines only "cpu" is available.

mxzgithub commented 8 months ago

You are totally right. I am no coder, so i hope the second commit fixes the problem.