llama.cpp releases version b9748, adding a "verbose" field to its schema and providing binaries for macOS, Linux, Android, Windows, and openEuler. The release includes CPU, Vulkan, OpenVINO, SYCL, and ROCm support across multiple architectures, with iOS and Windows CUDA and Vulkan builds available.
llama.cpp releases version b9748 with new binaries and features
from English