datawhalechina/self-llm ? reverse-engineered prompt

I want a Chinese friendly learning repo that teaches beginners how to use open source large models on Linux from start to finish.

Please build it as a tutorial style project with clear notebooks and markdown guides for setup, local deployment, daily use, and fine tuning. It should cover common models like Qwen, Llama, ChatGLM, InternLM, MiniCPM, and a few multimodal ones too. I’d like examples for command line chatting, simple demo deployment, and fine tuning with LoRA and full parameter training. If useful, include separate guidance for different hardware environments like NVIDIA, AMD, and Ascend, and make it easy for students to follow step by step.

Make the README welcoming and easy to scan, with quick start instructions, model support lists, and a few inspiring example projects that show what people can do with these models. Use plain language, keep the learning path beginner friendly, and organize it so someone can go from setup to running and then to fine tuning without getting lost.