master #67

Closed
yqxtqymn wants to merge 5 commits from yqxtqymn/ai-voice-cloning:master into master
First-time contributor

whisper->whisperx

whisper->whisperx
yqxtqymn added 4 commits 2023-03-06 02:00:21 +00:00
yqxtqymn added 1 commit 2023-03-06 02:04:22 +00:00
removed some comments
First-time contributor

This is good idea, whispherx is much better at precision and diazrization.

Feature request: Take an audio file with multiple speakers (for ex: a long podcast with 2 speakers), diazrize it using whishperx and then then seperate the audio tracks by speaker1 or speaker2 and then use one of them for training. Is it possible?

This is good idea, whispherx is much better at precision and diazrization. Feature request: Take an audio file with multiple speakers (for ex: a long podcast with 2 speakers), diazrize it using whishperx and then then seperate the audio tracks by speaker1 or speaker2 and then use one of them for training. Is it possible?
Owner

Implemented (and credited) in 0e3bbc55f8, since:

  • tab conversion to spaces (yuck)
  • a lot of odd formatting changes that would be a pain to clean up
  • it was based on an older commit, and rebasing by hand would be a chore
  • PR doesn't retain whispercpp
Implemented (and credited) in 0e3bbc55f82dff79674fd7f6075b4331e00aa296, since: * tab conversion to spaces (yuck) * a lot of odd formatting changes that would be a pain to clean up * it was based on an older commit, and rebasing by hand would be a chore * PR doesn't retain whispercpp
mrq closed this pull request 2023-03-06 05:22:26 +00:00

Pull request closed

Sign in to join this conversation.
No description provided.