llama.cpp
57c03b78
- metal : improve perf via smaller int registers
Go
Login via GitHub
Home
Pricing
FAQ
Install
Login
via GitHub
Commit
View On
GitHub
Commit
1 year ago
metal : improve perf via smaller int registers
References
#5021 - ggml : add Flash Attention
Author
ggerganov
Committer
ggerganov
Parents
6be02b59
Loading