added 'weights_only' param in torch.load examples

[Viditagarwal7479]

Fixes #111876

torch.load without setting weights_only=True is unsafe. So updating examples of torch.load to use weights_only=True where possible and weights_only=False elsewhere with a warning of being unsafety.

cc @svekars @carljparker

[pytorch-bot]

🔗 Helpful Links

🧪 See artifacts and rendered test results at hud.pytorch.org/pr/112860

• 📄 Preview Python docs built from this PR
• 📄 Preview C++ docs built from this PR
• ❓ Need help or want to give feedback on the CI? Visit the bot commands wiki or our office hours

Note: Links to docs will display an error until the docs builds have been completed.

✅ No Failures

As of commit 809c2f4 with merge base b10fa8a ():
💚 Looks good so far! There are no failures yet. 💚

This comment was automatically generated by Dr. CI and updates every 15 minutes.

[linux-foundation-easycla]

The committers listed above are authorized under a signed CLA.

• ✅ login: Viditagarwal7479 / name: Vidit Agarwal (809c2f4)

[Viditagarwal7479]

@pytorchbot label "docathon-h2-2023"

[vadimkantorov]

Also, very recently on a nightly torch.load(, weights_only=True) threw as strange deprecation warning, repro in #52181 (comment)

Maybe would be nice to assert in test that it produces no warnings

[Viditagarwal7479]

Do you mean to add context manager, with warnings.catch_warnings(action="error") in test/test_serialization.py. So the depreciation warning doesn't comes into the stable PyTorch?

[vadimkantorov]

I think, this UserWarning needs to be fixed by not using the deprecated UntypedStorage... It's just strange that hopefully-soon-to-be-default-safe-method raises deprecation warnings :) which sort of deceiptfully signals that the user's not doing the right thing by using weights_only=True

[vadimkantorov]

Deprecation UserWarning was on 2.1.0.dev20230802+cpu, not sure how it is on more recent nightlies, hope it is fixed now :)

[Viditagarwal7479]

That depreciation warning is also coming in 2.1.0 stable with cuda 11.8

[Viditagarwal7479]

weights_only=True let's us only load torch.tensors specifically speaking torch.TypedStorage and python default datatypes (int, float, list, tuples,...). Loading python default datatypes with this argument doesn't raise any warning but loading tensors raises warning.
Internally also when using Storage class a depreciation warning is set which can be seen at:
https://pytorch.org/docs/stable/storage.html

"All storage classes except for torch.UntypedStorage will be removed in the future, and torch.UntypedStorage will be used in all cases."

[Viditagarwal7479]

@kit1980 a gentle reminder to review the PR.

[kit1980]

@pytorchbot merge

[pytorchmergebot]

Merge failed

Reason: This PR needs a release notes: label
If your changes are user facing and intended to be a part of release notes, please use a label starting with release notes:.

If not, please add the topic: not user facing label.

To add a label, you can comment to pytorchbot, for example
@pytorchbot label "topic: not user facing"

For more information, see
https://github.com/pytorch/pytorch/wiki/PyTorch-AutoLabel-Bot#why-categorize-for-release-notes-and-how-does-it-work.

Details for Dev Infra team

Raised by workflow job

[svekars]

@pytorchbot merge

[pytorchmergebot]

Merge started

Your change will be merged once all checks pass (ETA 0-4 Hours).

Learn more about merging in the wiki.

Questions? Feedback? Please reach out to the PyTorch DevX Team

Advanced Debugging

Check the merge workflow status
here