3 ms·
Whisper small/tiny/base are almost four years old (they were not updated for Whisper v2 or v3). Is there really nothing better to benchmark against by now?
by modeless 3mo ago
Whisper small/tiny/base are almost four years old (they were not updated for Whisper v2 or v3). Is there really nothing better to benchmark against by now?
- Chu4eeno 3mo agoThere's tons, Parakeet was the last I remember seeing which seemed to gain traction (independent lightweight implementations etc).
- behnamoh 3mo agoParakeet isn't as good as whisper large.
- daemonologist 3mo agoParakeet is way faster (on Nvidia hardware) but not quite as accurate in my experience.
- meatmanek 3mo agoIt's also super fast on CPU.
- garblegarble 3mo agoI have tried everything (that will run on a 12GB RTX 4070) and I have yet to find anything with better accuracy than Whisper V2 Large for my dataset (discord audio from TTRPG sessions, isolated per-speaker, mostly non-American accents)
- xd1936 3mo agoSame, for my English-only podcast
- modeless 3mo agoNot v3?
- Ey7NFZ3P0nzAe 3mo agoVoxtral to me what better
- satvikpendem 3mo agoNvidia's Nemotron subsumes their older Parakeet model now even for real time streaming.
- satvikpendem 3mo agoThere are many [0], you can search and filter by streaming and open weight only as well. Looks like Voxtral and Nvidia's Nemotron are best. [0] https://artificialanalysis.ai/speech-to-text/non-streaming https://artificialanalysis.ai/speech-to-text/non-streaming
- aditya1821 3mo ago[dead]