llama.cpp b9907 Release: Common Header Fix & Updated Binaries

llama.cpp released version b9907, fixing a missing include in `common.h`. This update provides new pre-compiled binaries for macOS, iOS, Linux, Android, and Windows, supporting various CPU and GPU backends including CUDA, Vulkan, and OpenVINO.