llama.cpp
c4db5923
- metal : warp-based reduce for rms_norm
Go
Login via GitHub
Home
Pricing
FAQ
Install
Login
via GitHub
Commit
View On
GitHub
Commit
1 year ago
metal : warp-based reduce for rms_norm
References
#4256 - ggml : add ggml_soft_max_ext
Author
ggerganov
Parents
55717c98
Loading