shirayu/whispering

First try, really interesting, but with some issues (

cibernicola opened this issue · 3 comments

I just tested it, Windows 10 and Spanish language.
I notice phrases that I don't know where they appear in English (?¿) and repetitions of phrases when they are short.

The idea is very very interesting, if I can help, I will continue testing. :)

image

Thank you for trying!
In recent commit 7132db3, I added missing operations.

Furthermore, I am aware that the value of blocksize needs to be corrected.
Perhaps it will work correctly by setting the proper value.

https://github.com/shirayu/whisper_streaming/blob/b2650dbcd8e65c26b108989c27bbc0f891663337/whisper_streaming/cli.py#L36

The value was related to the number of seconds of processing at a time, and that was not the cause of the bug.
(I added -n option to change it.)

I think it will work with a few more modifications around here.

https://github.com/shirayu/whisper_streaming/blob/dcddd0717698a595023be7198e17e8df1e1141d4/whisper_streaming/transcriber.py#L214-L219

I fixed most bugs.
I think there are still a few bugs.
If you find any bugs, please create a new issue.
Thank you.