alt
Hacker News
Flash-MSA: Accelerating Million-Token Training with Sparse Attention Kernels
10 points
•
by
rawsh
•
today at 8:46 PM
•
0 comments
•
view on HN
Comments