Skip to content
llama.cpp releases · Infrastructure

b9879

ggml-cuda: optimize conv_transpose_1d indexing (#25310)