orthrus

★ 0 updated 2mo ago ⑂ fork

Fast, lossless LLM inference via dual-view diffusion decoding.

No plain-English explanation yet — one is being written right now. Check back in a minute.