llama.cpp b9882 adds HF_TOKEN support, expands platform coverage

llama.cpp, the popular open-source C++ implementation for running large language models locally, has released version b9882. This signed release (GPG key B5690EEEBB952194) introduces a key change: the scripts now use the HF_TOKEN environment variable when downloading UI assets. The commit by Adrien Gallouët (angt@huggingface.co) means users can pre-set their Hugging Face token to avoid manual authentication during setup. This is particularly useful for users who frequently update or deploy llama.cpp's built-in web UI, as it eliminates permission errors when fetching assets from Hugging Face hubs.

The release also showcases llama.cpp's extensive platform support, with prebuilt binaries for macOS (Apple Silicon with KleidiAI, Intel, iOS XCFramework), Linux (x64 CPU/arm64, s390x, Vulkan, ROCm 7.2, OpenVINO, SYCL FP32/FP16), Android (arm64 CPU), and Windows (x64/arm64 CPU, OpenCL Adreno, CUDA 12/13, Vulkan, OpenVINO, SYCL, HIP). Additionally, openEuler builds are provided for specific AI accelerators. With over 120,000 GitHub stars and 20,300 forks, llama.cpp remains the go-to tool for running quantized models on consumer hardware, and this incremental update further smooths the deployment experience for developers and power users.