All issues

Search results

Is the main branch SLA2?

Status: Open.

#24 In thu-ml/SLA;

· SwayStar123 opened on Mar 2, 2026

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

Status: Open.

#23 In thu-ml/SLA;

· limou102 opened on Mar 2, 2026

Question Regarding FLOPs Calculation for LightningDiT-1p0B/1

Status: Open.

#22 In thu-ml/SLA;

· hu-jy-0406 opened on Feb 19, 2026

VarLen Support

Status: Open.

#21 In thu-ml/SLA;

· ValentinBonhomme opened on Jan 14, 2026

All values input SLA and the resulting tensor become NaN

Status: Open.

#20 In thu-ml/SLA;

· ZouYa99 opened on Jan 14, 2026

SLA and SageSLA (SLA based on SageAttention) Code Update

enhancementNew feature or request New feature or request

Status: Open.

Feature

#9 In thu-ml/SLA;

· jt-zhang opened on Dec 16, 2025

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

Status: Open.

#5 In thu-ml/SLA;

· dz1iang opened on Oct 30, 2025

WAN 2.1 Performance vs FA3

Status: Open.

#3 In thu-ml/SLA;

· OutofAi opened on Oct 21, 2025

Release SLA code and artifacts on Hugging Face

Status: Open.

#2 In thu-ml/SLA;

· NielsRogge opened on Oct 1, 2025