App
github-com-thu-ml-spargeattn
x
Search
Install Pinokio
Log in
Register
Log in
Register
Light mode
SpargeAttn
https://github.com/thu-ml/spargeattn
updated 2/25/2026, 12:07:21 AM
indexed 7/16/2026, 7:05:54 AM
[ICML2025] SpargeAttention: A training-free sparse attention that accelerates any model inference.
Follow
1
Loading community details…
Community
Search this app
Sort by
Recent activity
Newest
Top: Past day
Top: Past week
Top: Past month
Top: Past year
Top: All
Post about SpargeAttn...
Post
Loading...
SpargeAttn on Pinokio
Pinokio Apps Using This Repo
No Pinokio apps using this repo yet.
Community tags
None yet.
Check-ins
(0)
View all
Platforms (0)
No reports yet.
Arch (0)
No reports yet.
GPU (0)
No reports yet.
RAM (0)
No reports yet.
VRAM (0)
No reports yet.
Recent commits
reduce repo size
Jintao Zhang
7 months ago
ae5b629
Add new article reference for SpargeAttention2
Jintao Zhang
7 months ago
6fb72d6
fix cuda launch error on devices < sm89
whx1003
9 months ago
bfd980b
Merge pull request #62 from t0saki/main
Jintao Zhang
9 months ago
d097a13
Modify CogVideoX model path in inference example
jt-zhang
9 months ago
c01e14e
Please use `spas_sage2_attn_meansim_topk_cuda` and `block_sparse_sage2_attn_cuda` APIs
Jintao Zhang
9 months ago
9906f67
fix sm90 pad_size
whx1003
9 months ago
01c4b48
Please use `spas_sage2_attn_meansim_topk_cuda` and `block_sparse_sage2_attn_cuda` APIs
Jintao Zhang
9 months ago
2e6e89e
Please use `spas_sage2_attn_meansim_topk_cuda` and `block_sparse_sage2_attn_cuda` APIs
Jintao Zhang
9 months ago
62532ec
Please use `spas_sage2_attn_meansim_topk_cuda` and `block_sparse_sage2_attn_cuda` API
Jintao Zhang
9 months ago
baf8739