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 }}
NVIDIA
Megatron-LM
Repository navigation
Code
Issues
529
(529)
Pull requests
1.6k
(1.6k)
Discussions
Actions
Projects
Security and quality
Insights
More
items
All pull requests
Labels
105
(105)
Milestones
2
(2)
New pull request
Search pull requests
is
:
pr
state
:
open
is:pr state:open
Clear filter
Search
Pull requests
Open
1,630
(1,630)
Closed
5,838
(5,838)
Author
Label
Projects
Milestones
Reviews
Assignee
Sort by
Newest
descending
Sort…
Sort by Newest descending
More items
Comfortable display density
Compact display density
ci: prefer published merge-group Testmon baselines for PRs
#9677
·
balasaajay
opened
Nov 22, 2026
Contributor
·
3 of 7 tasks
·
·
2
ci: publish merge-group Testmon baselines after merge
#9676
·
balasaajay
opened
Nov 22, 2026
Contributor
·
3 of 7 tasks
·
·
2
ci: record Testmon baselines in main merge groups
#9675
·
balasaajay
opened
Nov 22, 2026
Contributor
·
3 of 7 tasks
·
·
2
Support optimizer-owned master weights in Megatron-FSDP v1
community-request
#9674
·
wokron
opened
Nov 22, 2026
·
5 of 7 tasks
·
·
2
Honour add_extra_token_to_sequence and pad short samples in GPTFIMDataset
community-request
#9673
·
adityaghai07
opened
Nov 22, 2026
Contributor
·
4 of 7 tasks
·
·
1
2
Apply grad_norm_skip_threshold when DSA indexer grads are clipped separately
community-request
#9671
·
adityaghai07
opened
Nov 22, 2026
Contributor
·
4 of 7 tasks
·
·
1
2
Fix token-id prompts and echo=False logprobs in the legacy /completions endpoint
community-request
#9669
·
adityaghai07
opened
Nov 22, 2026
Contributor
·
4 of 7 tasks
·
·
1
2
Count MLA layers in the hybrid-model FLOPs estimate
community-request
#9667
·
adityaghai07
opened
Nov 22, 2026
Contributor
·
4 of 7 tasks
·
·
1
2
Keep --override-ckpt-iteration when loading distributed optimizer state
community-request
#9665
·
adityaghai07
opened
Nov 22, 2026
Contributor
·
4 of 7 tasks
·
·
1
2
Cap the all-to-all drop-and-pad capacity at the local token count
community-request
#9663
·
adityaghai07
opened
Nov 22, 2026
Contributor
·
4 of 7 tasks
·
·
1
2
tests: print device diagnostics during fp8 param test
community-request
#9661
·
syifayuliawan990-svg
opened
Nov 21, 2026
·
·
2
fix(fsdp): disable parameter prefetching during no-grad evaluation
#9660
·
wujingyue
opened
Nov 21, 2026
Contributor
·
·
2
test: refresh MFSDP per-token GB200 golden values
complexity: low
#9659
·
wujingyue
opened
Nov 21, 2026
Contributor
·
1 task
·
·
5
perf(lm-head): run the LM head in sequence chunks to cut peak memory
#9657
·
xrennvidia
opened
Nov 21, 2026
Contributor
·
5 of 7 tasks
·
·
4
Stop TE linear layers from disabling tp_comm_overlap for later layers
community-request
Final Review
#9656
·
Graphi07
opened
Nov 21, 2026
·
4 of 7 tasks
·
·
1
2
Batch invariant All-Gather context parallelism
#9653
·
YigongQin
opened
Nov 21, 2026
Contributor
·
7 tasks
·
·
2
Overlap M-FSDP DP-Outer gradient reduction with backward again, reducing repeated reductions separately
#9652
·
rapatel
opened
Nov 20, 2026
Contributor
·
4 of 7 tasks
·
·
10
Make the paged-stash warmup cache test independent of allocator state
complexity: low
#9651
·
guihong-nv
opened
Nov 20, 2026
Contributor
·
1 of 7 tasks
·
·
5
Skip activation checkpointing when grad is disabled
Final Review
#9650
·
deepujain
opened
Nov 20, 2026
·
2 of 7 tasks
·
·
1
5
fix(inference): stop engine spin and EP hang when waiting requests can't be admitted
complexity: medium
#9646
·
lauradang
opened
Nov 20, 2026
Contributor
·
2 of 7 tasks
·
·
12
Move activation files into new folder
complexity: high
Final Review
skip-bot-approval
#9643
·
jiemingz
opened
Nov 20, 2026
Contributor
·
7 tasks
·
·
12
Release HybridEP routing references after eager combine
#9641
·
wujingyue
opened
Nov 20, 2026
Contributor
·
1 of 7 tasks
·
·
2
Fix the param norm with the precision-aware optimizer
#9639
·
CarlosGomes98
opened
Nov 20, 2026
Contributor
·
2 of 7 tasks
·
·
2
Preserve FIM document boundaries in packed attention metadata
community-request
#9638
·
Hang-get
opened
Nov 20, 2026
Contributor
·
5 of 6 tasks
·
·
2
Load deferred FIM dataset indices before sampling
community-request
#9637
·
Hang-get
opened
Nov 20, 2026
Contributor
·
3 of 5 tasks
·
·
2
Previous
1
2
3
4
5
6
7
…
40
Next
You can’t perform that action at this time.