Issues

• Is the main branch SLA2?

  thu-ml/SLA#24 · Filter by author SwayStar123 opened on Mar 2, 2026
  2

• How to finetune after replacing the original full attention with the SLA operator?

  thu-ml/SLA#23 · Filter by author limou102 opened on Mar 2, 2026

• Question Regarding FLOPs Calculation for LightningDiT-1p0B/1

  thu-ml/SLA#22 · Filter by author hu-jy-0406 opened on Feb 19, 2026

• VarLen Support

  thu-ml/SLA#21 · Filter by author ValentinBonhomme opened on Jan 14, 2026

• All values input SLA and the resulting tensor become NaN

  thu-ml/SLA#20 · Filter by author ZouYa99 opened on Jan 14, 2026

• SLA and SageSLA (SLA based on SageAttention) Code Update

  enhancement
  Feature

  thu-ml/SLA#9 · Filter by author jt-zhang opened on Dec 16, 2025
  1

• wan2.2 is faster than FA2 but consumes more VRAM.

  thu-ml/SLA#5 · Filter by author dz1iang opened on Oct 30, 2025
  5

• WAN 2.1 Performance vs FA3

  thu-ml/SLA#3 · Filter by author OutofAi opened on Oct 21, 2025
  5

• Release SLA code and artifacts on Hugging Face

  thu-ml/SLA#2 · Filter by author NielsRogge opened on Oct 1, 2025