Spaces:

sanchit-gandhi

/

whisper-jax

like 2.2k

Running

Would be cool to bump the model version but there are reports of regression going from large-v2 to large-v3 (e.g. as described here: https://github.com/openai/whisper/discussions/1806 and https://github.com/openai/whisper/discussions/1783)

Overall I think these cases are few and far between, and that using large-v3 would be worthwhile

Latest version of the Space is using large-v3!