Hacker Newsnew | past | comments | ask | show | jobs | submitlogin

The transcript is really bad for a podcast I tried, I'm almost certain that's OpenAIs whisper, which when I tested wasn't very accurate when people don't speak extremely clearly or over each other.


Is it the small or large model? Whisper is leagues beyond anything previously available, while being runnable on consumer grade hardware.


Do you have the source? Would like to try it through my tools




Guidelines | FAQ | Lists | API | Security | Legal | Apply to YC | Contact

Search: