This project can automatically detect the local GPU/CPU/RAM configurations and screen out large models suitable for the current hardware from HuggingFace. It is based on comprehensive score rankings like LiveBench and Chatbot Arena ELO, rather than simply sorting by the number of parameters. It also supports functions such as simulating specified GPUs, finding the graphics cards needed to run target models, one-click starting conversations, and generating Python code snippets.