[PeterStuer]

For image generation the article's setup might be viable, but do not expect to run LLM's with satisfactory quality and speed on 12GB vram.