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 }}
Uh oh!
There was an error while loading.
Please reload this page
.
NVIDIA
/
TensorRT-LLM
Public
Notifications
You must be signed in to change notification settings
Fork
2.7k
Star
14.5k
Code
Issues
567
Pull requests
861
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 Gemma4 MTP assistant support
- #15833
#15833
Merged
lfr-0531
merged 27 commits into
NVIDIA:main
NVIDIA/TensorRT-LLM:main
from
lfr-0531:user/fanrongl/gemma4-mtp-support
lfr-0531/TensorRT-LLM:user/fanrongl/gemma4-mtp-support
Copy head branch name to clipboard
Aug 4, 2026
Conversation
Commits
27
(27)
Checks
Files changed
Merged
[None][feat] Add Gemma4 MTP assistant support
#15833
lfr-0531
merged 27 commits into
NVIDIA:main
NVIDIA/TensorRT-LLM:main
from
lfr-0531:user/fanrongl/gemma4-mtp-support
lfr-0531/TensorRT-LLM:user/fanrongl/gemma4-mtp-support
Copy head branch name to clipboard
Commits
Commits on Aug 1, 2026
[None][feat] Add Gemma4 MTP assistant support
Show description for 28c0a43
lfr-0531
committed
28c0a43
View commit details
Copy full SHA for 28c0a43
Browse repository at this point
[None][docs] Document Gemma4 MTP support
Show description for 80522a4
lfr-0531
committed
80522a4
View commit details
Copy full SHA for 80522a4
Browse repository at this point
[None][docs] Remove Gemma4 MTP footnote
Show description for 66faefd
lfr-0531
committed
66faefd
View commit details
Copy full SHA for 66faefd
Browse repository at this point
[None][fix] Avoid Gemma4 Bandit false positive
Show description for b075c1c
lfr-0531
committed
b075c1c
View commit details
Copy full SHA for b075c1c
Browse repository at this point
[None][fix] Fix Gemma4 MTP decoding correctness
Show description for 2276ccd
lfr-0531
committed
2276ccd
View commit details
Copy full SHA for 2276ccd
Browse repository at this point
[None][perf] Restore Gemma4 assistant CUDA graph replay
Show description for ff036a4
lfr-0531
committed
ff036a4
View commit details
Copy full SHA for ff036a4
Browse repository at this point
[None][fix] Harden Gemma4 MTP runtime integration
Show description for 2c91e64
lfr-0531
committed
2c91e64
View commit details
Copy full SHA for 2c91e64
Browse repository at this point
[None][fix] Avoid redundant Gemma4 assistant KV cache
Show description for a0470c2
lfr-0531
committed
a0470c2
View commit details
Copy full SHA for a0470c2
Browse repository at this point
[None][test] Clarify Gemma4 MTP graph pool coverage
Show description for 02fd45e
lfr-0531
committed
02fd45e
View commit details
Copy full SHA for 02fd45e
Browse repository at this point
feat: move Gemma4 MTP to one-model shared KV
Show description for 8ac60c4
lfr-0531
committed
8ac60c4
View commit details
Copy full SHA for 8ac60c4
Browse repository at this point
[None][refactor] minimize Gemma4 MTP integration
Show description for e567d43
lfr-0531
committed
e567d43
View commit details
Copy full SHA for e567d43
Browse repository at this point
[None][refactor] minimize Gemma4 MTP VLM proxies
Show description for 02f6283
lfr-0531
committed
02f6283
View commit details
Copy full SHA for 02f6283
Browse repository at this point
[None][refactor] use native FlashInfer decode for Gemma4 MTP
Show description for 988d31d
lfr-0531
committed
988d31d
View commit details
Copy full SHA for 988d31d
Browse repository at this point
[None][refactor] generalize shared-KV MTP integration
Show description for 0d2bb3c
lfr-0531
committed
0d2bb3c
View commit details
Copy full SHA for 0d2bb3c
Browse repository at this point
[None][refactor] generalize shared-KV draft capability
Show description for da665e4
lfr-0531
committed
da665e4
View commit details
Copy full SHA for da665e4
Browse repository at this point
[None][refactor] refine shared-KV FlashInfer integration
Show description for f58eac0
lfr-0531
committed
f58eac0
View commit details
Copy full SHA for f58eac0
Browse repository at this point
[None][refactor] scope shared-KV config to Gemma4
Show description for 192d76b
lfr-0531
committed
192d76b
View commit details
Copy full SHA for 192d76b
Browse repository at this point
[None][fix] stabilize Gemma4 shared-KV MTP
Show description for 9529bbb
lfr-0531
committed
9529bbb
View commit details
Copy full SHA for 9529bbb
Browse repository at this point
[None][feat] support separate-KV FlashInfer drafting
Show description for b89c10b
lfr-0531
committed
b89c10b
View commit details
Copy full SHA for b89c10b
Browse repository at this point
[None][refactor] simplify Gemma4 shared-KV MTP integration
Show description for 2e8c4de
lfr-0531
committed
2e8c4de
View commit details
Copy full SHA for 2e8c4de
Browse repository at this point
[None][refactor] simplify Gemma4 MTP helpers and tests
Show description for 5ade312
lfr-0531
committed
5ade312
View commit details
Copy full SHA for 5ade312
Browse repository at this point
[None][refactor] simplify Gemma4 config and embedding sharing
Show description for a9881dc
lfr-0531
committed
a9881dc
View commit details
Copy full SHA for a9881dc
Browse repository at this point
[None][refactor] address Gemma4 review feedback
Show description for 1937164
lfr-0531
committed
1937164
View commit details
Copy full SHA for 1937164
Browse repository at this point
[None][fix] fix speculative decoding CI regressions
Show description for 660735f
lfr-0531
committed
660735f
View commit details
Copy full SHA for 660735f
Browse repository at this point
[None][fix] harden Gemma4 speculative draft setup
Show description for 5dcaf06
lfr-0531
committed
5dcaf06
View commit details
Copy full SHA for 5dcaf06
Browse repository at this point
Commits on Aug 3, 2026
[None][fix] scope FlashInfer one-engine guard
Show description for 8df01d7
lfr-0531
committed
8df01d7
View commit details
Copy full SHA for 8df01d7
Browse repository at this point
Merge branch 'main' into user/fanrongl/gemma4-mtp-support
lfr-0531
authored
4207c44
View commit details
Copy full SHA for 4207c44
Browse repository at this point
You can’t perform that action at this time.
