[DEV] fix layerwise torch_dist checkpointing fails due to empty rank

[FDecaYed]

What does this PR do ?

Fix torch_dist checkpointing format fail when there are empty groups due to layerwise sharding.
cherry-picked into master muon PR here #2241 for main

Contribution process

flowchart LR
    A[Pre-checks] --> B[PR Tests]
    subgraph Code Review/Approval
        C1[Expert Review] --> C2[Final Review]
    end
    B --> C1
    C2 --> D[Merge]

Loading

Pre-checks

• I want this PR in a versioned release and have added the appropriate Milestone (e.g., Core 0.8)
• 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

The following process is enforced via the CODEOWNERS file for changes into megatron/core. For changes outside of megatron/core, it is up to the PR author whether or not to tag the Final Reviewer team.

For MRs into `main` branch

(Step 1): Add PR label Expert Review

(Step 2): Collect the expert reviewers reviews

1. Attach the Expert Review label when your PR is ready for review.
2. GitHub auto-assigns expert reviewers based on your changes. They will get notified and pick up your PR soon.

⚠️ Only proceed to the next step once all reviewers have approved, merge-conflict are resolved and the CI is passing.
Final Review might get declined if these requirements are not fulfilled.

(Step 3): Final Review

1. Add Final Review label
2. GitHub auto-assigns final reviewers based on your changes. They will get notified and pick up your PR soon.

(Optional Step 4): Cherry-pick into release branch

If this PR also needs to be merged into core_r* release branches, after this PR has been merged, select Cherry-pick to open a new PR into the release branch.

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.

Merging your PR

Any member of core-adlr and core-nemo will be able to merge your PR.

[copy-pr-bot]

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.

[FDecaYed]

/ok to test 3d1e121

[skyw]

Often line tested more cases and they all seemed to be working.

two warnings remain, I think they are normal?

WARNING:megatron.core.utils:make_tp_sharded_tensor_for_checkpoint received extra kwargs: ['allow_shape_mismatch']

WARNING:megatron.core.dist_checkpointing.optimizer:Sharded parameters mapping is empty. It means tensors in model state dict do not correspond to tensors in optimizer parameters map. Make sure to call state_dict with `keep_vars=True`.

[FDecaYed]

WARNING:megatron.core.utils:make_tp_sharded_tensor_for_checkpoint received extra kwargs: ['allow_shape_mismatch']

this seems to be a issue that's in TOT base branch and not related to this PR

WARNING:megatron.core.dist_checkpointing.optimizer:Sharded parameters mapping is empty. It means tensors in model state dict do not correspond to tensors in optimizer parameters map. Make sure to call state_dict with `keep_vars=True`.

This is expected since we're re-using megatron's method for mapping optim state to param. but in our case, some rank will have empty group. Do you think we should update the warning and mention 'if you're not using layerwise'?

[skyw]

some minor suggestions, otherwise LGTM.

Should have a follow up PR to add test protecting previous failed cases.

[FDecaYed]

/ok to test d582e5e