Modernize polynomial_custom_function: torch.accelerator and setup_context

[jvz37]

Fixes #3880

Description

Modernizes the custom autograd function tutorial by:

• Replacing hardcoded torch.device("cpu") with torch.accelerator for accelerator-agnostic device selection
• Splitting legacy forward(ctx, input) into forward(input) + setup_context(ctx, inputs, output) per PyTorch 2.0+ recommended pattern

Tested locally: 2000 iterations, loss converged, output identical to original.

Checklist

• The issue that is being fixed is referred in the description (see above "Fixes #ISSUE_NUMBER")
• Only one issue is addressed in this pull request
• [] Labels from the issue that this PR is fixing are added to this pull request
• No unnecessary issues are included into this pull request.

[pytorch-bot]

🔗 Helpful Links

🧪 See artifacts and rendered test results at hud.pytorch.org/pr/pytorch/tutorials/3885

• 📄 Preview Python docs built from this PR

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

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

[jvz37]

Note: the commit author shows jalilAlva (another account I have) due to a local git config mistake. This PR was opened and authored by jvz37, my the account registered for Docathon 2026.