Comment by J_Shelby_J

Hey, mixlayer is really cool.

I also have a Rust LLM inference project. The overlap is very high between what mixlayer is doing and what my project is doing. It's actually crazy how we basically have the same features. [1] Right now I'm still using llama.cpp on the backend, but eventually want to move to candle via mistral.rs.

[1] https://github.com/ShelbyJenkins/llm_client