llama cpp python no longer uses GPU abetlen/llama-cpp-python#912

h2oai · Nov 15, 2023 · 77a4457 · 77a4457
1 parent a295ce8
commit 77a4457
Showing 1 changed file with 2 additions and 1 deletion.
diff --git a/docs/README_LINUX.md b/docs/README_LINUX.md
@@ -175,7 +175,8 @@ These instructions are for Ubuntu x86_64 (other linux would be similar with diff
     * GGUF ONLY for CUDA GPU (keeping CPU package in place to support CPU + GPU at same time):
       ```bash
       pip uninstall -y llama-cpp-python-cuda
-      pip install https://github.com/jllllll/llama-cpp-python-cuBLAS-wheels/releases/download/textgen-webui/llama_cpp_python_cuda-0.2.18+cu118-cp310-cp310-manylinux_2_31_x86_64.whl
+      python -m pip install llama-cpp-python --prefer-binary --upgrade --extra-index-url=https://jllllll.github.io/llama-cpp-python-cuBLAS-wheels/AVX2/cu118
+      # pip install https://github.com/jllllll/llama-cpp-python-cuBLAS-wheels/releases/download/textgen-webui/llama_cpp_python_cuda-0.2.18+cu118-cp310-cp310-manylinux_2_31_x86_64.whl
       ```
     * GGUF ONLY for CPU-AVX (can be used with -cuda one above)
       ```bash