jla-org/jax-linear-attention

Fast linear-attention, SSM and attention kernels (Pallas Mosaic GPU) and a training stack for JAX on NVIDIA Hopper

1 best points · 0 comments · 1 Hacker News post · 7 GitHub stars · View on GitHub

Hacker News discussions

• Show HN: We ported FLA's Triton kernels to Jax Pallas; they're faster on H100 (1 points, 0 comments, by KarLexAI, 2026-10-22)