Replies: 2 comments 2 replies
-
Re-downloading the "large" model should get the new "large" model via whisper.cpp in SE. Yes, the new large model works really well in the one test I did :) |
Beta Was this translation helpful? Give feedback.
0 replies
-
Greetings. Delivered version SE 3.6.10 Beta. Loaded the large Whisper model. I understand correctly that the GPU is not used in this case, but the CPU is used? |
Beta Was this translation helpful? Give feedback.
2 replies
Sign up for free
to join this conversation on GitHub.
Already have an account?
Sign in to comment
-
openai/whisper#397
reading this thread new model for whisper seems to improves the error rate.
Beta Was this translation helpful? Give feedback.
All reactions