llm

llm was an early attempt to port llama.cpp to Rust to make it available for use as a library, as well as to improve its general maintainability. I took over ownership of the original library from setzer22, and redesigned it to facilitate my library vision. Unfortunately, this was well before the advent of coding agents, so it soon became impossible for me and my team of contributors to keep up with the pace of advancements in llama.cpp.

Its lasting impact was providing me with the motivation to design GGUF.