Skip to content
Navigation Menu
Sign in
Appearance settings
Platform
AI CODE CREATION
GitHub Copilot
Write better code with AI
GitHub Copilot app
Direct agents from issue to merge
MCP Registry
Integrate external tools
DEVELOPER WORKFLOWS
Actions
Automate any workflow
Codespaces
Instant dev environments
Issues
Plan and track work
Code Review
Manage code changes
Code Quality
Enforce quality at merge
APPLICATION SECURITY
GitHub Advanced Security
Find and fix vulnerabilities
Code security
Secure your code as you build
Secret protection
Stop leaks before they start
EXPLORE
Why GitHub
Documentation
Blog
Changelog
Marketplace
View all features
Solutions
BY COMPANY SIZE
Enterprises
Small and medium teams
Startups
Nonprofits
BY USE CASE
App Modernization
DevSecOps
DevOps
CI/CD
View all use cases
BY INDUSTRY
Healthcare
Financial services
Manufacturing
Government
View all industries
View all solutions
Resources
EXPLORE BY TOPIC
AI
Software Development
DevOps
Security
View all topics
EXPLORE BY TYPE
Customer stories
Events & webinars
Ebooks & reports
Business insights
GitHub Skills
SUPPORT & SERVICES
Documentation
Customer support
Community forum
Trust center
Partners
View all resources
Open Source
COMMUNITY
GitHub Sponsors
Fund open source developers
PROGRAMS
Security Lab
Maintainer Community
GitHub Stars
Archive Program
REPOSITORIES
Topics
Trending
Collections
Enterprise
ENTERPRISE SOLUTIONS
Enterprise platform
AI-powered developer platform
AVAILABLE ADD-ONS
GitHub Advanced Security
Enterprise-grade security features
Copilot for Business
Enterprise-grade AI features
Premium Support
Enterprise-grade 24/7 support
Pricing
Search
/
Sign in
Sign up
Appearance settings
You signed in with another tab or window.
Reload
to refresh your session.
You signed out in another tab or window.
Reload
to refresh your session.
You switched accounts on another tab or window.
Reload
to refresh your session.
Dismiss alert
{{ message }}
NVIDIA
/
TensorRT-LLM
Public
Notifications
You must be signed in to change notification settings
Fork
2.8k
Star
14.7k
Code
Issues
606
Pull requests
894
Discussions
Actions
Projects
Security and quality
0
Insights
Additional navigation options
Code
Issues
Pull requests
Discussions
Actions
Projects
Security and quality
Insights
[None][feat] add VisualGen VSA and SOL sparse attention on PrimTS block-sparse FMHA
- #18079
#18079
Open
heyuhhh
wants to merge 19 commits into
NVIDIA:main
NVIDIA/TensorRT-LLM:main
from
heyuhhh:user/yuhangh/update_vsa
heyuhhh/TensorRT-LLM:user/yuhangh/update_vsa
Copy head branch name to clipboard
Conversation
Commits
19
(19)
Checks
Files changed
Open
[None][feat] add VisualGen VSA and SOL sparse attention on PrimTS block-sparse FMHA
#18079
heyuhhh
wants to merge 19 commits into
NVIDIA:main
NVIDIA/TensorRT-LLM:main
from
heyuhhh:user/yuhangh/update_vsa
heyuhhh/TensorRT-LLM:user/yuhangh/update_vsa
Copy head branch name to clipboard
Commits
Commits on Sep 19, 2026
refactor: align VisualGen sparse attention workflow
Show description for 889b3b0
heyuhhh
committed
889b3b0
View commit details
Copy full SHA for 889b3b0
Browse repository at this point
perf: fuse VSA predictor preprocessing into Triton kernels
Show description for 325b263
heyuhhh
committed
325b263
View commit details
Copy full SHA for 325b263
Browse repository at this point
fix: tolerate backend-specific kwargs in VisualGen TRTLLM attention
Show description for 76cd2f3
heyuhhh
committed
76cd2f3
View commit details
Copy full SHA for 76cd2f3
Browse repository at this point
refactor: share the sparse timestep schedule across VisualGen algorithms
Show description for 49be991
heyuhhh
committed
49be991
View commit details
Copy full SHA for 49be991
Browse repository at this point
feat: add the TRTLLM SOL backend and unify VisualGen SOL under one config
Show description for 34ef975
heyuhhh
committed
34ef975
View commit details
Copy full SHA for 34ef975
Browse repository at this point
Commits on Sep 20, 2026
refactor: inject shared FMHA state through the TRTLLM attention constructor
Show description for 4fe3be5
heyuhhh
committed
4fe3be5
View commit details
Copy full SHA for 4fe3be5
Browse repository at this point
refactor: keep the prepared sparse timestep in the component attention state
Show description for 0c00f38
heyuhhh
committed
0c00f38
View commit details
Copy full SHA for 0c00f38
Browse repository at this point
refactor: make the SOL predictor a functional operator and add the exact threshold
Show description for 9c666fb
heyuhhh
committed
9c666fb
View commit details
Copy full SHA for 9c666fb
Browse repository at this point
refactor: unify sparse algorithm naming in the VisualGen attention dispatch
Show description for c63a310
heyuhhh
committed
c63a310
View commit details
Copy full SHA for c63a310
Browse repository at this point
fix: resolve review feedback on VisualGen VSA and SOL sparse attention
Show description for 85b6b4d
heyuhhh
committed
85b6b4d
View commit details
Copy full SHA for 85b6b4d
Browse repository at this point
Merge branch 'main' into user/yuhangh/update_vsa
heyuhhh
authored
4fe7022
View commit details
Copy full SHA for 4fe7022
Browse repository at this point
Commits on Sep 21, 2026
Merge branch 'main' into user/yuhangh/update_vsa
heyuhhh
authored
0f0ac14
View commit details
Copy full SHA for 0f0ac14
Browse repository at this point
Commits on Sep 22, 2026
test: group VisualGen sparse attention tests per algorithm
Show description for 8a48808
heyuhhh
committed
8a48808
View commit details
Copy full SHA for 8a48808
Browse repository at this point
docs: describe VisualGen sparse attention by algorithm and list SOL in the model matrix
Show description for f5046d3
heyuhhh
committed
f5046d3
View commit details
Copy full SHA for f5046d3
Browse repository at this point
Merge branch 'main' into user/yuhangh/update_vsa
heyuhhh
authored
451dc02
View commit details
Copy full SHA for 451dc02
Browse repository at this point
fix: fuse QKV for the dense calls of the TRTLLM SOL backend
Show description for 11f096b
heyuhhh
committed
11f096b
View commit details
Copy full SHA for 11f096b
Browse repository at this point
test: point the B300 list at the reorganized SOL module
Show description for 5583786
heyuhhh
committed
5583786
View commit details
Copy full SHA for 5583786
Browse repository at this point
Merge branch 'main' into user/yuhangh/update_vsa
heyuhhh
authored
72d194a
View commit details
Copy full SHA for 72d194a
Browse repository at this point
Commits on Sep 23, 2026
Merge branch 'main' into user/yuhangh/update_vsa
heyuhhh
authored
5fe496a
View commit details
Copy full SHA for 5fe496a
Browse repository at this point
You can’t perform that action at this time.
