cturan/llama.cpp @ 4064c0e3b6c27440f5d12b7caaf90b4088c28c61

镜像来自 https://github.com/cturan/llama.cpp

0cc4m 4064c0e3b6 Vulkan: Use improved q4_k and q5_k dequant code in dequant shaders (#10798)		1 年之前
..
include	43041d2eb3 ggml: load all backends from a user-provided search path (#10699)	1 年之前
src	4064c0e3b6 Vulkan: Use improved q4_k and q5_k dequant code in dequant shaders (#10798)	1 年之前
.gitignore	17eb6aa8a9 vulkan : cmake integration (#8119)	1 年之前
CMakeLists.txt	59f4db1088 ggml : add predefined list of CPU backend variants to build (#10626)	1 年之前