Skip to content
#

hardware-native

Here is 1 public repository matching this topic...

GQLSA: Grouped-Query Latent Sparse Attention — A hardware-native attention mechanism combining latent compression, grouped-query sharing, and block-sparse attention for linear O(T) complexity.

  • Updated Sep 8, 2026
  • Python

Add this topic to your repo

To associate your repository with the hardware-native topic, visit your repo's landing page and select "manage topics."

Learn more