llama.cpp b9913 adds Q2_0 quantization and CPU backend

The llama.cpp project released version b9913, introducing Q2_0 quantization with a new type definition and CPU backend support. This update is available across multiple platforms including macOS, Linux, Windows, Android, and openEuler.

• Added Q2_0 quantization type definition and corresponding CPU backend implementation.
• Disabled KleidiAI support for macOS Apple Silicon builds.
• Provided binaries for various architectures and backends, including CUDA 12/13, Vulkan, ROCm, OpenVINO, SYCL, and HIP.
• Released UI binaries alongside the core engine updates.