AirLLM 2.8T LLM inference with single 4GB GPU
About this project
AirLLM 2.8T LLM inference with single 4GB GPU. It is infrastructure rather than a finished product: the layer that serves, quantises, trains or orchestrates models. It is maintained by lyogavin and published as a public repository you can read, fork and run yourself.
On GitHub the repository is tagged chinese-llm, chinese-nlp, finetune, generative-ai, instruct-gpt, instruction-set, llama, llm, lora and open-models, which is a fair summary of what it covers. That generally means an API surface, batching and memory management, and often an OpenAI-compatible endpoint so existing clients keep working unchanged.
It is written mainly in Jupyter Notebook and released under APACHE-2.0, a permissive license — commercial use is fine as long as copyright notices are kept. It is actively maintained: the most recent commit landed on November 18, 2026.
To try it, clone the repository and follow its README; that file is the canonical install guide. Because it runs on your own infrastructure, budget time for environment setup: dependencies, model weights and configuration usually take longer than the first successful run.
Key facts as recorded on GitHub: 42,598 stars; 4,418 forks; primary language Jupyter Notebook; license APACHE-2.0; last commit on November 18, 2026; topics chinese-llm, chinese-nlp, finetune, generative-ai, instruct-gpt, instruction-set.