havaker / ollama-nix Goto Github PK

Hey,
thanks for creating this!

I would love to use this, however I don't have a cuda GPU.

It would be nice if the nixos service had the common package option, to override the package the ollama nixos service will use. Maybe also something like enbaleCudaSupport that makes the path = [ config.boot.kernelPackages.nvidia_x11.bin ]; part optional.

Hi. ollama has changed a lot since 0.1.17. ollama doesn't call the llama-cpp's executable on its own anymore but use their libraries. And very recently rocm support.

Any plans on updating your flake to the latest ollama version?

Flake works well, I have it working with full GPU acceleration on the models that can fit in my Nivdia Vram.

Would you consider committing the flake to the https://github.com/jmorganca/ollama

I'm not sure they would accept it, but it could be fun!

Well done, nice work and great flake.