HKUSTDial/flash-sparse-attention

Trainable fast and memory-efficient sparse attention

by Data Intelligence and Analytics Lab @ HKUST(GZ) Python BSD-3-Clause licence 54 forks Created May 2025 View on GitHub

Stars
748
4% of HKUSTDial's stars
Rank
#87,220
of 851,979 repos, top 11%
In Python
#16,014
of 173,293 repos, top 10%

Among 851,979 repos

Compare with

630–1k stars: 35,602 repos, including flash-sparse-attention

Stars along the bottom, and repos in each band on a log scale

Quick stats

Next rank
+1 star to reach #87,114
Next milestone
+19 stars to reach the top 10%
Popular
One of 490,661 repos with 100+ stars
Yearly average
496 stars a year since May 2025

Stars counted Recounted daily Next count around