Search results

5090 black is not support？

Status: Open.

#109 In thu-ml/SpargeAttn;

· secret-deus opened on Jan 12, 2026

First, I created a sparse attention module for using Flux 1.D which I tested. Some thoughts, questions...

Status: Open.

#107 In thu-ml/SpargeAttn;

· ukaprch opened on Dec 29, 2025

Using the same selection algorithm, but employing non-quantized Flashinfer for computation, why are the generated results all garbled?

Status: Open.

#106 In thu-ml/SpargeAttn;

· hjdjs12 opened on Dec 27, 2025

Possible logical issues in some code within the <code>setup.py</code> file.

Status: Open.

#104 In thu-ml/SpargeAttn;

· junyii37 opened on Dec 25, 2025

windows 下 源码安装的时候 执行link 报错 如何解决？

Status: Open.

#103 In thu-ml/SpargeAttn;

· tailuchen opened on Dec 24, 2025

Does SparseAttn support NVIDIA GB10 architecture?

Status: Open.

#102 In thu-ml/SpargeAttn;

· GarfieldHuang opened on Dec 22, 2025

Please use <code>spas_sage2_attn_meansim_topk_cuda</code> and <code>block_sparse_sage2_attn_cuda</code> APIs

enhancementNew feature or requestNew feature or request

Status: Open.

Feature

#101 In thu-ml/SpargeAttn;

· jt-zhang opened on Dec 16, 2025

Question about TopK

Status: Open.

#99 In thu-ml/SpargeAttn;

· jhl13 opened on Nov 24, 2025

Calling SpargeAttn Twice Corrupts CUDA Context (unspecified launch failure)

Status: Open.

#97 In thu-ml/SpargeAttn;

· sajalmaheshwari624 opened on Nov 19, 2025

Release of FLUX model dict

Status: Open.

#96 In thu-ml/SpargeAttn;

· SeanAlpaca818 opened on Nov 13, 2025

[BUG] SPARSE ATTRS OF TUNED-MODELS WILL BE LOADED INCORRECTLY!

Status: Open.

#95 In thu-ml/SpargeAttn;

· kissint8 opened on Nov 13, 2025

Different Block Sizes for Query and Key

Status: Open.

#93 In thu-ml/SpargeAttn;

· mikeallen39 opened on Nov 4, 2025