Update ModelPT to use torch.load with weights only

[chtruong814]

What does this PR do ?

Update ModelPT to use torch.load with weights only

Collection: [Note which collection this PR will affect]

Changelog

• Add specific line by line info of high level changes in this PR.

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The Jenkins CI system has been replaced by GitHub Actions self-hosted runners.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

[chtruong814]

@nithinraok I'm a bit surprised at the failure in this test because weights only should have been the default anyway.
https://github.com/NVIDIA-NeMo/NeMo/actions/runs/20732262103/job/59615859508?pr=15255

Any idea what might be going on?

[nithinraok]

@nithinraok I'm a bit surprised at the failure in this test because weights only should have been the default anyway. https://github.com/NVIDIA-NeMo/NeMo/actions/runs/20732262103/job/59615859508?pr=15255

Any idea what might be going on?

@blisc could you help fix above issue. I think maybe related to OmegaConf serialization while unpickling

[blisc]

Can we add weights_only as an option to maybe_init_from_pretrained_checkpoint() as opposed to strictly requiring it?

[chtruong814]

@blisc is it possible to restrict the allowed objects needed? If maybe_init_from_pretrained_checkpoint will allow anything still, we'd run into the same issue we need to resolve.