This project aimming to run gemma using rust, (削除) which can provide high performance to infer. (削除ここまで)

I apologize for the suboptimal performance of this code. It doesn't fully leverage Rust's capabilities. If you're looking for a more efficient implementation of Gemme2 that runs well on a computer, please visit lmrs. This code is well structured and is primarily intended as a reference and learning tool for the Rust equivalent of gemma_pytorch now.

• Reference implement

pytorch implementation

lmrs

chat tempelate

chat tempelate in cpp

• tokenizer
• model