llama.cpp
ac26f270
- cuda : increase C to 128 for better performance
Go
Login via GitHub
Home
Pricing
FAQ
Install
Login
via GitHub
Commit
View On
GitHub
Commit
1 year ago
cuda : increase C to 128 for better performance
References
flash-attn-cuda
Author
ggerganov
Committer
ggerganov
Parents
9a5c2a16
Loading