FlashAttention-2: Faster Attention with Better Parallelism and Work Partitioning paper: https://
tridao.me/publications/f
lash2/flash2.pdf
…
github: https://
github.com/Dao-AILab/flas
h-attention
… Scaling Transformers to longer sequence lengths has been a major problem in the last several years, promising to improve performance

→ View original post on X — @_akhaliq