llama.cpp
fadde671
- Dequant improvements rebase (#8255)
Go
Login via GitHub
Home
Pricing
FAQ
Install
Login
via GitHub
Commit
View On
GitHub
Commit
1 year ago
Dequant improvements rebase (#8255) * Single load for half2 * Store scales in local mem * Vec load quantized values
References
#8255 - Dequant improvements rebase
Author
AidanBeltonS
Parents
a27152b6
Loading
