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
Developer articles
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 }}
ModelTC
LightLLM
Repository navigation
Code
Issues
101
(101)
Pull requests
87
(87)
Discussions
Actions
Projects
Security and quality
Insights
More
items
fix(glm5): align FP8 precision and preserve reasoning tool calls
- #1657
#1657
Open
shihaobai
wants to merge 14 commits into
main
ModelTC/LightLLM:main
from
bsh/glm53_fix
ModelTC/LightLLM:bsh/glm53_fix
Copy head branch name to clipboard
Conversation
Commits
14
(14)
Checks
Files changed
Open
fix(glm5): align FP8 precision and preserve reasoning tool calls
#1657
shihaobai
wants to merge 14 commits into
main
ModelTC/LightLLM:main
from
bsh/glm53_fix
ModelTC/LightLLM:bsh/glm53_fix
Copy head branch name to clipboard
Commits
Commits on Oct 21, 2026
fix: parse GLM tool calls inside reasoning
shihaobai
committed
ea2b4db
View commit details
Copy full SHA for ea2b4db
Browse repository at this point
fix: harden GLM reasoning and XML tool parsing
shihaobai
committed
e8c8bc8
View commit details
Copy full SHA for e8c8bc8
Browse repository at this point
refactor: keep GLM tool parsing within existing detectors
shihaobai
committed
637c079
View commit details
Copy full SHA for 637c079
Browse repository at this point
Commits on Oct 22, 2026
refactor: separate thinking policy from reasoning state
shihaobai
committed
6c902a0
View commit details
Copy full SHA for 6c902a0
Browse repository at this point
refactor: simplify GLM parser state and argument handling
shihaobai
committed
bdda980
View commit details
Copy full SHA for bdda980
Browse repository at this point
fix(glm5): align indexer, routing and mHC numerics
Show description for 710118e
shihaobai
committed
710118e
View commit details
Copy full SHA for 710118e
Browse repository at this point
fix(glm5): match training post rounding and close precision tests
Show description for 92d580e
shihaobai
committed
92d580e
View commit details
Copy full SHA for 92d580e
Browse repository at this point
Commits on Oct 23, 2026
fix(moe): honor UE8M0 scales in Triton expert activations
shihaobai
committed
aee915e
View commit details
Copy full SHA for aee915e
Browse repository at this point
fix(glm5): align eager SwiGLU, W2 routing and mHC post with training
shihaobai
committed
3e7a521
View commit details
Copy full SHA for 3e7a521
Browse repository at this point
fix(glm5): revert FFN numerical alignment for isolated indexer ablation
shihaobai
committed
35f71e9
View commit details
Copy full SHA for 35f71e9
Browse repository at this point
feat(glm5): add opt-in BF16 indexer precision ablation
shihaobai
committed
6d219d3
View commit details
Copy full SHA for 6d219d3
Browse repository at this point
feat(fp8): decouple activation scale format from checkpoint weights
shihaobai
committed
bb6833d
View commit details
Copy full SHA for bb6833d
Browse repository at this point
Commits on Oct 27, 2026
refactor(glm5): trim precision fixes to FP8 and prefill combine
shihaobai
committed
a139d3f
View commit details
Copy full SHA for a139d3f
Browse repository at this point
Merge PR #1651: preserve GLM tool calls inside reasoning
shihaobai
committed
2e042f4
View commit details
Copy full SHA for 2e042f4
Browse repository at this point
You can’t perform that action at this time.