whisper.cpp
d052e64d
- CUDA: batched+noncont MMQ, refactor bs>1 MoE code (llama/13199)
Go
Login via GitHub
Home
Pricing
FAQ
Install
Login
via GitHub
Commit
View On
GitHub
Commit
1 year ago
CUDA: batched+noncont MMQ, refactor bs>1 MoE code (llama/13199)
References
#3108 - sync : ggml
Author
JohannesGaessler
Committer
ggerganov
Parents
780750a1
Loading