llama.cpp
CUDA: quantized KV support for FA vec
#7527
Merged

Loading