Popular repositories Loading

1. ferrite ferrite Public

   Run Qwen3.5 on any CPU with a from-scratch Rust inference engine that beats llama.cpp and hits 89 tok/s on a four-core laptop with no GPU.

   TypeScript 3

2. corestudio2.github.io corestudio2.github.io Public

   Run Qwen3.5 LLMs on any CPU at llama.cpp-beating speed with this from-scratch Rust inference engine.

   TypeScript