After installing llama-cpp-python using https://github.com/dougeeai/llama-cpp-python-wheels, the GPU starts to work

CMAKE_ARGS="-DGGML_CUDA=on" pip install llama-cpp-python --upgrade --force-reinstall --no-cache-dir

After executing this command and restarting ComfyUI, the problem was solved.

After installing llama-cpp-python using https://github.com/dougeeai/llama-cpp-python-wheels, the GPU starts to work

Thank you very much, but this page does not have wheels available for Linux.

After installing llama-cpp-python using https://github.com/dougeeai/llama-cpp-python-wheels, the GPU starts to work

Thanks for the wheels, very helpful