fix no_shard training convergency and add unittest for no_shard by wplf · Pull Request #3754 · NVIDIA/Megatron-LM

wplf · 2026-03-09T15:54:54Z

What does this PR do ?

Fixes no_shard convergence by correcting grad_norm and all_reduce logic.

fix no_shard grad_norm calculation
fix no_shard all_reduce grad continue bug
Make no_shard to be compatiable without overlap_param_gather and overlap_grad_reduce
add unittest for no_shard
add assert False for no_shard and init_model_with_meta_device follow fully_shard.py:326

Grad_norm and loss convergency and unittest result is below.

⚠️ For major changes (either in lines of code or in its impact), please make sure to first share a design doc with the team. If you're unsure what's the best way to do so, contact the @mcore-oncall.

Contribution process

Pre-checks

I have added relevant unit tests
I have added relevant functional tests
I have added proper typing to my code Typing guidelines
I have added relevant documentation
I have run the autoformatter.sh on my PR

Code review

Feel free to message or comment the @mcore-oncall to help accelerate your merge into main. The less complex your PR is, the faster it will be approved and merged!

All PRs start as draft. If you open a non-draft PR, it will be automatically converted to draft.

Step 1: Mark PR as "Ready for Review"

When your PR is ready, click Ready for Review.
An oncall reviewer is auto-assigned and expert reviewers are notified based on your changes.
- Some PRs may jump straight to step 2. This is determined by .github/CODEOWNERS.

⚠️ Only mark as ready once merge-conflicts are resolved and the CI is passing.
Final Review might get declined if these requirements are not fulfilled.

Step 2: Final Review

For PRs that change megatron/core, once all expert reviewers have approved, the Final Review label is applied automatically and final reviewers are assigned.

For PRs outside megatron/core, this step is skipped.

Step 3: Approved

Once all required reviewers have approved, the Approved label is applied automatically.

Merge

Any member of mcore-engineers will be able to merge your PR.

For MRs into `dev` branch

The proposed review process for `dev` branch is under active discussion.

MRs are mergable after one approval by either eharper@nvidia.com or zijiey@nvidia.com.

copy-pr-bot · 2026-03-09T15:54:59Z

This pull request requires additional validation before any workflows can run on NVIDIA's runners.

Pull request vetters can view their responsibilities here.

Contributors can view more details about this message here.

github-actions · 2026-03-09T15:55:10Z

This PR has been automatically converted to draft because all PRs must start as drafts.

When you are ready for review, click Ready for Review to begin the review process. This will:

Add the oncall reviewer (optional reviewer)
Add required review teams based on your changes

See the contribution guide for more details.

cspades

LGTM!

yaox12 · 2026-04-20T04:34:01Z

/ok to test e6a1239

add unittest for no_shard and add empty cache to avoid OOM add meta_device_check for no_shard following fully_shard.py:326 Signed-off-by: jinliangl <jinliangl@nvidia.com>

wplf · 2026-05-21T07:21:49Z

/ok to test 153b1f8

svcnvidia-nemo-ci · 2026-05-21T09:38:35Z

🔄 Merge queue validation started!

You can track the progress here: https://github.com/NVIDIA/Megatron-LM/actions/runs/26218104171

…IA#3754) Signed-off-by: jinliangl <jinliangl@nvidia.com>

…IA#3754) Signed-off-by: jinliangl <jinliangl@nvidia.com> Signed-off-by: yhgalaxy <yhgalaxy@outlook.com>

…IA#3754) Signed-off-by: jinliangl <jinliangl@nvidia.com> Signed-off-by: Jon Barker <jbarker@aws-cmh-slurm-1-vscode-02.cm.cluster>

# Description NVIDIA#4024 NVIDIA#3746 NVIDIA#2799 NVIDIA#3676 NVIDIA#4133 NVIDIA#3295 NVIDIA#3831 NVIDIA#3754 ## Type of change - [ ] New feature (non-breaking change which adds functionality) - [ ] Infra/Build change (changes to CI/CD workflows or build scripts) - [ ] Code refactoring - [ ] Documentation change - [x] Bug fix - [ ] Breaking change ## Changes - Content 1 - Content 2 - Content 3 - Content 4 ## Checklist - [ ] I have read and followed the contributing guidelines - [ ] The functionality is complete - [ ] I have commented my code, particularly in coverage report uploading steps - [ ] I have made corresponding changes to the documentation - [ ] My changes generate no new warnings - [ ] I have added/updated tests that prove my feature works - [ ] New and existing unit tests pass locally --------- Signed-off-by: Cory Ye <cye@nvidia.com> Signed-off-by: yexiaochuan <yexiaochuan@baidu.com> Signed-off-by: Akshat Kumar <akshat230405@gmail.com> Signed-off-by: Youngeun Kwon <youngeunk@nvidia.com> Signed-off-by: jinliangl <jinliangl@nvidia.com> Co-authored-by: Ritesh Patel <ripatel@nvidia.com> Co-authored-by: Cory Ye <cye@nvidia.com> Co-authored-by: Cory Ye <44509866+cspades@users.noreply.github.com> Co-authored-by: Xiaochuan Ye <tap91624@gmail.com> Co-authored-by: yexiaochuan <yexiaochuan@baidu.com> Co-authored-by: Akshat8510 <akshat230405@gmail.com> Co-authored-by: Kunlun Li <94586211+kunlunl@users.noreply.github.com> Co-authored-by: eternally-z <zzywzj@gmail.com> Co-authored-by: Philip Petrakian <ppetrakian@nvidia.com> Co-authored-by: Youngeun Kwon <youngeunk@nvidia.com> Co-authored-by: xuwchen <xuwenc@nvidia.com> Co-authored-by: Li Jinliang <jinliangl@nvidia.com>

…IA#3754) Signed-off-by: jinliangl <jinliangl@nvidia.com>

…IA#3754) Signed-off-by: jinliangl <jinliangl@nvidia.com> Signed-off-by: Dmytro Pykhtar <dpykhtar@nvidia.com>

wplf requested review from a team as code owners March 9, 2026 15:54

svcnvidia-nemo-ci marked this pull request as draft March 9, 2026 15:55

wplf marked this pull request as ready for review March 9, 2026 15:55

svcnvidia-nemo-ci requested a review from a team March 9, 2026 15:55

wplf force-pushed the fix-mfsdp-no-shard branch from 920e8d0 to d2fde24 Compare March 9, 2026 16:09

cspades approved these changes Mar 9, 2026

View reviewed changes

Comment thread megatron/core/distributed/fsdp/src/megatron_fsdp/param_and_grad_buffer.py

svcnvidia-nemo-ci added the Final Review PR is in the "final review" stage label Mar 9, 2026

cspades added the module: megatron-fsdp label Mar 9, 2026

cspades mentioned this pull request Mar 9, 2026

Modify mfsdp default data-parallel-sharding-strategy #3691

Merged

6 tasks

wplf mentioned this pull request Mar 12, 2026

[dev] fix no_shard training convergency and add unittest for no_shard #3835

Merged

wplf closed this Mar 22, 2026

wplf force-pushed the fix-mfsdp-no-shard branch from 54d7959 to a811ac1 Compare March 22, 2026 14:30

svcnvidia-nemo-ci added this to the Core 0.16 milestone Mar 22, 2026

copy-pr-bot Bot temporarily deployed to test March 22, 2026 14:31 Inactive

ko3n1g temporarily deployed to main March 22, 2026 14:32 — with GitHub Actions Inactive

wplf reopened this Mar 22, 2026

shjwudp approved these changes Mar 23, 2026

View reviewed changes

kvareddy approved these changes Mar 23, 2026

View reviewed changes

wplf enabled auto-merge March 31, 2026 07:13

wplf disabled auto-merge March 31, 2026 07:22

wplf enabled auto-merge March 31, 2026 07:22

svcnvidia-nemo-ci added Approved All necessary approvals have been made and removed Final Review PR is in the "final review" stage labels Apr 14, 2026

Phlip79 removed the request for review from a team April 15, 2026 16:18

yaox12 disabled auto-merge April 20, 2026 04:33

yaox12 enabled auto-merge April 20, 2026 04:33

wplf added 2 commits May 21, 2026 00:17

fix no_shard training convergency

efd3cba

add unittest for no_shard and add empty cache to avoid OOM add meta_device_check for no_shard following fully_shard.py:326 Signed-off-by: jinliangl <jinliangl@nvidia.com>

use more elegant way to indicate nosharding strategy

153b1f8

wplf force-pushed the fix-mfsdp-no-shard branch from e6a1239 to 153b1f8 Compare May 21, 2026 07:19

copy-pr-bot Bot temporarily deployed to public May 21, 2026 07:22 Inactive

copy-pr-bot Bot temporarily deployed to test May 21, 2026 07:22 Inactive

copy-pr-bot Bot temporarily deployed to public May 21, 2026 07:25 Inactive

copy-pr-bot Bot temporarily deployed to public May 21, 2026 07:26 Inactive

copy-pr-bot Bot temporarily deployed to public May 21, 2026 07:34 Inactive

yaox12 added this pull request to the merge queue May 21, 2026

Merged via the queue into NVIDIA:main with commit 80a2d39 May 21, 2026
73 checks passed

janEbert pushed a commit to janEbert/Megatron-LM that referenced this pull request Jun 2, 2026

fix no_shard training convergency and add unittest for no_shard (NVID…

271c79b

…IA#3754) Signed-off-by: jinliangl <jinliangl@nvidia.com>

mathemakitten pushed a commit to mathemakitten/Megatron-LM that referenced this pull request Jun 12, 2026

fix no_shard training convergency and add unittest for no_shard (NVID…

d9eab37

…IA#3754) Signed-off-by: jinliangl <jinliangl@nvidia.com>

yhgalaxy pushed a commit to yhgalaxy/Megatron-LM that referenced this pull request Jun 17, 2026

fix no_shard training convergency and add unittest for no_shard (NVID…

5c2945f

…IA#3754) Signed-off-by: jinliangl <jinliangl@nvidia.com> Signed-off-by: yhgalaxy <yhgalaxy@outlook.com>

zhaoyinglia mentioned this pull request Jul 16, 2026

cherry-pick fix pr for megatron_fsdp flagos-ai/Megatron-LM-FL#79

Merged

13 tasks

terminator123 pushed a commit to 021ai/Megatron-LM that referenced this pull request Aug 3, 2026

fix no_shard training convergency and add unittest for no_shard (NVID…

46ce5a0

…IA#3754) Signed-off-by: jinliangl <jinliangl@nvidia.com>

Navigation Menu

Uh oh!

fix no_shard training convergency and add unittest for no_shard - #3754

fix no_shard training convergency and add unittest for no_shard#3754
yaox12 merged 2 commits into
NVIDIA:mainfrom
wplf:fix-mfsdp-no-shard

wplf commented Mar 9, 2026 •

edited

Loading

Uh oh!

copy-pr-bot Bot commented Mar 9, 2026

Uh oh!

github-actions Bot commented Mar 9, 2026

Uh oh!

cspades left a comment

Uh oh!

Uh oh!

yaox12 commented Apr 20, 2026

Uh oh!

wplf commented May 21, 2026

Uh oh!

svcnvidia-nemo-ci commented May 21, 2026

Uh oh!

Uh oh!

Reviewers

Assignees

Labels

Projects

Milestone

Development

Uh oh!

8 participants

Uh oh!

Conversation

wplf commented Mar 9, 2026 • edited Loading Uh oh! There was an error while loading. Please reload this page.

Uh oh!

What does this PR do ?

Contribution process

Pre-checks

Code review

Step 1: Mark PR as "Ready for Review"

Step 2: Final Review

Step 3: Approved

Merge

Uh oh!

copy-pr-bot Bot commented Mar 9, 2026

Uh oh!

github-actions Bot commented Mar 9, 2026

Uh oh!

cspades left a comment

Choose a reason for hiding this comment

Uh oh!

Uh oh!

yaox12 commented Apr 20, 2026

Uh oh!

wplf commented May 21, 2026

Uh oh!

svcnvidia-nemo-ci commented May 21, 2026

Uh oh!

Uh oh!

Reviewers

Assignees

Labels

Projects

Milestone

Development

Uh oh!

8 participants

wplf commented Mar 9, 2026 •

edited

Loading