| | --- |
| | license: mit |
| | tags: |
| | - llama-cpp-python |
| | - prebuilt-wheels |
| | - huggingface-spaces |
| | - cpu-only |
| | - python-3.13 |
| | --- |
| | |
| | # Llama-CPP-Python Pre-built Wheels (CPU Only) |
| |
|
| | This repository provides pre-compiled Python wheels for `llama-cpp-python`, specifically optimized for **Hugging Face Spaces (Free CPU Tier)**. |
| |
|
| | ## ๐ Key Features |
| | - **Zero Compilation:** Skips the 15-minute C++ build process on HF Spaces. |
| | - **Python 3.13 Support:** Built for the latest Python version. |
| | - **Generic CPU:** Compiled with `GGML_NATIVE=OFF` to ensure compatibility with older cloud processors (no "Illegal Instruction" errors). |
| |
|
| | ## ๐ Usage in HF Spaces |
| |
|
| | ### Dockerfile (Recommended) |
| | Add this to your Dockerfile to install the wheel instantly: |
| | ```dockerfile |
| | RUN pip install [https://huggingface.co/Jameson040/llama-cpp-python-wheels/resolve/main/llama_cpp_python-0.3.16-cp313-cp313-linux_x86_64.whl](https://huggingface.co/Jameson040/llama-cpp-python-wheels/resolve/main/llama_cpp_python-0.3.16-cp313-cp313-linux_x86_64.whl) |