[Nemo1] Generate sharded optimizer state dicts only if needed for saving

[ananthsub]

What does this PR do ?

If not saving optimizers in the checkpoint, don't generate the sharded optimizer state dict, as it will ultimately be discarded. The ckpt contents are the same, and this is strictly a perf improvement

Collection: NLP

Changelog

Usage

• You can potentially add a usage example below

# Add a code snippet demonstrating how to use this 

GitHub Actions CI

The Jenkins CI system has been replaced by GitHub Actions self-hosted runners.

The GitHub Actions CI will run automatically when the "Run CICD" label is added to the PR.
To re-run CI remove and add the label again.
To run CI on an untrusted fork, a NeMo user with write access must first click "Approve and run".

Before your PR is "Ready for review"

Pre checks:

• Make sure you read and followed Contributor guidelines
• Did you write any new necessary tests?
• Did you add or update any necessary documentation?
• Does the PR affect components that are optional to install? (Ex: Numba, Pynini, Apex etc)
  • Reviewer: Does the PR have correct import guards for all optional libraries?

PR Type:

• New Feature
• Bugfix
• Documentation

If you haven't finished some of the above items you can still open "Draft" PR.

Who can review?

Anyone in the community is free to review the PR once the checks have passed.
Contributor guidelines contains specific people who can review PRs to various areas.

Additional Information

• Related to # (issue)

[github-actions]

beep boop 🤖: 🙏 The following files have warnings. In case you are familiar with these, please try helping us to improve the code base.

---

Your code was analyzed with PyLint. The following annotations have been identified:

************* Module nemo.collections.nlp.parts.nlp_overrides
nemo/collections/nlp/parts/nlp_overrides.py:212:0: C0301: Line too long (140/119) (line-too-long)
nemo/collections/nlp/parts/nlp_overrides.py:217:0: C0301: Line too long (149/119) (line-too-long)
nemo/collections/nlp/parts/nlp_overrides.py:233:0: C0301: Line too long (123/119) (line-too-long)
nemo/collections/nlp/parts/nlp_overrides.py:424:0: C0301: Line too long (136/119) (line-too-long)
nemo/collections/nlp/parts/nlp_overrides.py:510:0: C0301: Line too long (152/119) (line-too-long)
nemo/collections/nlp/parts/nlp_overrides.py:730:0: C0301: Line too long (140/119) (line-too-long)
nemo/collections/nlp/parts/nlp_overrides.py:735:0: C0301: Line too long (149/119) (line-too-long)
nemo/collections/nlp/parts/nlp_overrides.py:1001:0: C0301: Line too long (128/119) (line-too-long)
nemo/collections/nlp/parts/nlp_overrides.py:1005:0: C0301: Line too long (141/119) (line-too-long)
nemo/collections/nlp/parts/nlp_overrides.py:1009:0: C0301: Line too long (149/119) (line-too-long)
nemo/collections/nlp/parts/nlp_overrides.py:1066:0: C0301: Line too long (135/119) (line-too-long)
nemo/collections/nlp/parts/nlp_overrides.py:1193:0: C0301: Line too long (121/119) (line-too-long)
nemo/collections/nlp/parts/nlp_overrides.py:1775:0: C0301: Line too long (152/119) (line-too-long)
nemo/collections/nlp/parts/nlp_overrides.py:248:4: C0116: Missing function or method docstring (missing-function-docstring)
nemo/collections/nlp/parts/nlp_overrides.py:381:4: C0116: Missing function or method docstring (missing-function-docstring)
nemo/collections/nlp/parts/nlp_overrides.py:425:4: C0116: Missing function or method docstring (missing-function-docstring)
nemo/collections/nlp/parts/nlp_overrides.py:602:4: C0116: Missing function or method docstring (missing-function-docstring)
nemo/collections/nlp/parts/nlp_overrides.py:618:4: C0116: Missing function or method docstring (missing-function-docstring)
nemo/collections/nlp/parts/nlp_overrides.py:637:4: C0116: Missing function or method docstring (missing-function-docstring)
nemo/collections/nlp/parts/nlp_overrides.py:878:4: C0116: Missing function or method docstring (missing-function-docstring)
nemo/collections/nlp/parts/nlp_overrides.py:997:0: C0115: Missing class docstring (missing-class-docstring)
nemo/collections/nlp/parts/nlp_overrides.py:1686:4: C0116: Missing function or method docstring (missing-function-docstring)
nemo/collections/nlp/parts/nlp_overrides.py:1774:4: C0116: Missing function or method docstring (missing-function-docstring)
nemo/collections/nlp/parts/nlp_overrides.py:18:0: W0611: Unused import re (unused-import)
nemo/collections/nlp/parts/nlp_overrides.py:106:4: W0611: Unused tensorstore imported from megatron.core.dist_checkpointing.strategies (unused-import)

-----------------------------------
Your code has been rated at 9.69/10

---

Thank you for improving NeMo's documentation!

[maanug-nv]

LGTM. should not functionally change ckpt content, just perf improvement correct?

[ananthsub]

LGTM. should not functionally change ckpt content, just perf improvement correct?

That's correct. The ckpt contents are the same, and this is strictly a perf improvement