| Linux, CUDA, AVX512. | |
| ```sh | |
| $ git clone "https://github.com/abetlen/llama-cpp-python" --depth=1 | |
| $ cd llama-cpp-python | |
| $ git submodule update --init --recursive | |
| $ CMAKE_ARGS="-DLLAMA_CUBLAS=on" python -m pip install . | |
| ``` |
| Linux, CUDA, AVX512. | |
| ```sh | |
| $ git clone "https://github.com/abetlen/llama-cpp-python" --depth=1 | |
| $ cd llama-cpp-python | |
| $ git submodule update --init --recursive | |
| $ CMAKE_ARGS="-DLLAMA_CUBLAS=on" python -m pip install . | |
| ``` |