llama.cpp
f0cbb6dd
- iq1_s: turn off SIMD implementation for QK_K = 64 (it does not work)
Go
Login via GitHub
Home
Pricing
FAQ
Install
Login
via GitHub
Commit
View On
GitHub
Commit
1 year ago
iq1_s: turn off SIMD implementation for QK_K = 64 (it does not work)
References
ik/i-quants-64
#5760 - Make i-quants work with super-blocks of 64 (CPU and Metal)
Author
Iwan Kawrakow
Parents
47d52b2b
Loading