Regrouping Issue/Question

[LexiconGap]

Sorry in advance, I'm probably misunderstanding the regroup feature. I'm testing the regroup feature but it doesn't seem to work how I think it should on even basic strings.

I'm assuming the following code:

model = stable_whisper.load_model('medium')
result = model.transcribe('subs.aac', regroup='sp=, /!/! ')
result.to_srt_vtt('test.srt', word_level=False)

should split up the string "Oh, is something up there?" into "Oh," and "is something up there?". It doesn't. Perhaps I misread something in the documentation but I can't seem to come up with my own regrouping methods.

[jianfch]

The space after , belong to the next word so it should be sp=,* instead of sp=, but just sp=, should work too.

[LexiconGap]

Change made, issue persists

model = stable_whisper.load_model('medium')
result = model.transcribe('subs.aac', regroup='sp=,* /,/!/! ')
result.to_srt_vtt('test.srt', word_level=False)

[jianfch]

Change made, issue persists

Can you save the result as JSON and share it?

[LexiconGap]

Here you go test.json

[jianfch]

It should be fixed in d51edb6ad86b06f4582f4c06fcf8a4b6dc8e0bca (2.13.7).

[LexiconGap]

Awesome! I'll spend tomorrow actually learning how to use github and test to see if the problem is resolved I'm assuming you probably verified it via the json file. thank you!

[jianfch]

The fix should available via PyPI with pip install -U stable-ts.

[LexiconGap]

Amazing! Tested. Issue appears to be resolved.