search

shirayu / whispering

Streaming transcriber with whisper
MIT License
682 stars 53 forks source link

Return partial analysis result #8

Open shirayu opened 2 years ago

shirayu commented 2 years ago

Whisper assumes a 30-second interval as an input. So, whispering does not request analysis from whisper until 30 seconds have elapsed without --allow-padding. However, it is useful to show temporary transcriptions for short intervals.

KTRosenberg commented 2 years ago

Might there be any progress on this? Also, how could this be made to work with more than 30 seconds and with arbitrary splits? This would be great for continuous recognition.

shirayu commented 1 year ago

Hi, I haven't worked on this yet, while I added --frame argument on v0.6.4. Thank you for your interest.