Flash-MSA: Accelerating Million-Token Training with Sparse Attention Kernels
20 points
4 hours ago
| 0 comments
| nanduruganesh.github.io
| HN
No one has commented on this post.