Implements dual-chunk-flash-attn backend for dual chunk attention with sparse attention support (#11844)

This commit is contained in:
Tao He
2025-05-12 19:52:47 -07:00
committed by GitHub
parent f6518b2b48
commit 60f7624334
17 changed files with 2444 additions and 32 deletions
File diff suppressed because it is too large Load Diff