Skip to content
Navigation Menu
Sign in
Products
Platform
GitHub
Repository hosting, code review, and collaboration
Copilot
AI coding agent for tasks, projects and automated work
Build & Ship
Actions
CI/CD and automation
Code Quality
Quality analysis, fixes, and enforcement
Codespaces
Cloud development environments
Secure
Code Security
Vulnerability detection and remediation
Secret Protection
Secret scanning and leak prevention
All Features
Developers
Learn
Documentation
Training
Certifications
Build
Build with Copilot
Customize Copilot
Integrations
Explore
Trending repos
GitHub Sponsors
Events
Changelog
Changelog
Your source for the latest features, improvements, and fixes across GitHub products, all in one place.
Businesses
By Stage
Enterprises
Team
Startups
Nonprofits
By Industry
Healthcare
Financial services
Manufacturing
Government
See all industries
Evaluate GitHub
Why GitHub
Customer Stories
Business insights
Trust Center
Add-ons
GitHub Advanced Security
Enterprise-grade code security
Copilot for business
Enterprise-grade AI features
Premium Support
Enterprise-grade 24/7 support
Resources
Learn & Connect
Documentation
Resource center
Blog
Events & webinars
Community Forum
Customer Stories
Integrations
Support
Customer support
Training
GitHub Status
Company
About
Partners
Changelog
GitHub Universe
GitHub Universe
The event for developers, by developers — talks, workshops, and what’s next for GitHub.
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
spark-local-inference
Repository navigation
Code
Issues
31
(31)
Pull requests
23
(23)
Actions
Projects
Security and quality
Insights
More
items
All pull requests
Labels
9
(9)
Milestones
0
(0)
New pull request
Search pull requests
is
:
pr
state
:
open
is:pr state:open
Clear filter
Search
Pull requests
Open
23
(23)
Closed
61
(61)
Author
Label
Projects
Milestones
Reviews
Assignee
Sort by
Newest
descending
Sort…
Sort by Newest descending
More items
Comfortable display density
Compact display density
SM103 prefill: Triton block-FP8 and route quantization, serial MoE reduction, and fewer GLM copies
#134
·
original-el8
opened
Oct 28, 2026
Collaborator
·
·
Declare benchmark calls and operand alignment for raced b12x tactics
#132
·
original-el8
opened
Oct 28, 2026
Collaborator
·
·
Fix nondeterministic SM103 mHC prefill projection
#131
·
original-el8
opened
Oct 28, 2026
Collaborator
·
·
Fix deferred-copy views, CuTe tensor pickling and GB300 package naming
#129
·
original-el8
opened
Oct 28, 2026
Collaborator
·
·
Bound export shutdown by peer exit; recognize phased PCIe joint modes
#128
·
voipmonitor
opened
Oct 28, 2026
Collaborator
·
·
Reject unsupported Qwen dense storage during checkpoint validation
#127
·
radimsem
opened
Oct 28, 2026
·
·
Link received payload objects into the package cache instead of copying
#125
·
javieratomic
opened
Oct 28, 2026
Contributor
·
·
Hash and publish package payloads on parallel workers
#124
·
javieratomic
opened
Oct 28, 2026
Contributor
·
·
Make the compiler memory headroom a preparation setting
#122
·
javieratomic
opened
Oct 28, 2026
Contributor
·
·
Append DeepSeek V4.1 window snapshots during cache writes
#108
·
voipmonitor
opened
Oct 27, 2026
Collaborator
·
·
Pin FlashInfer with caller-owned attention output support
#106
·
voipmonitor
opened
Oct 27, 2026
Collaborator
·
·
Reuse compiled forwards across compatible kernel plans
#105
·
voipmonitor
opened
Oct 27, 2026
Collaborator
·
·
Consult ignore_eos when classifying a finish reason at the output limit
#102
·
mark-yong
opened
Oct 26, 2026
Contributor
·
·
Run DeepSeek V4.1 mHC through a TileLang kernel pair on SM120/121
#82
·
original-el8
opened
Oct 25, 2026
Collaborator
·
·
fix(heliopause): fall back when no-replace rename is unsupported
#79
·
efegokdemir
opened
Oct 25, 2026
·
·
1
1
Absorb a short final message segment into the prefill chunk
#69
·
original-el8
opened
Oct 24, 2026
Collaborator
·
·
Fix GLM tool-call argument tag parsing
#59
·
aarongiuffre
opened
Oct 23, 2026
·
·
Keep feedback work visible while ordinary decode is in flight
#43
·
Maxrio9
opened
Oct 23, 2026
Contributor
·
·
1
preparation: tolerate optional restored/rejected candidate fields in the tuning report
#33
·
msuiche
opened
Oct 22, 2026
·
·
fix(qualify): render native JSON schema instructions in frozen prompts
#19
·
Maxrio9
opened
Oct 22, 2026
Contributor
·
·
fix(export): read CUDA payloads into pinned host memory
#18
·
Maxrio9
opened
Oct 22, 2026
Contributor
·
·
fix: preserve both private draft pages in shifted prefix checkpoints
#16
·
Maxrio9
opened
Oct 22, 2026
Contributor
·
·
runtime: share bounded image windows between requests
#8
·
acronjob
opened
Oct 21, 2026
·
·
You can’t perform that action at this time.