Lm.rs: Minimal CPU LLM inference in Rust with no dependency

The GitHub project lm.rs, created by Samuel Vitorino, demonstrates minimal LLM inference using Rust and is aimed at running language models on the CPU without relying on machine learning libraries. The project includes benchmarks and download links for models and tokenizers, as well as instructions to convert models to the LMRS format.

• Project includes minimal Rust code for LLM CPU inference
• Supports Llama3.2 1B and 3B models
• Models available on Hugging Face
• Provides quantized model download options
• WebUI available for backend interaction