Thanks to #ramalamma detecting my #vulkan capable #integratedgpu I can now run a lot of models without the CPU cores melting. I still need to work out the right runes for #ellama to work properly with the #mistral model though.