ParaAttention
★ 1
updated 1y ago
⑂ fork
Context parallel attention that accelerates DiT model inference with dynamic caching
No plain-English explanation yet — one is being written right now. Check back in a minute.