[ThirdParty] [FA] update flash-attention to support use_varlen in flashmask and add several bug fix. (#838)
Core Functional Library for Large Scale Distributed Training
版权所有:中国计算机学会技术支持:开源发展技术委员会 京ICP备13000930号-9 京公网安备 11010802032778号
PaddleFleet
Core Functional Library for Large Scale Distributed Training