git commit -m "Fix sample_rate calculation with WeightedRandomSampler

[MN-NR]

Types of changes

• Bug fix (non-breaking change which fixes an issue)
• New feature (non-breaking change which adds functionality)
• Breaking change (fix or feature that would cause existing functionality to change)
• Docs change / refactoring / dependency upgrade

Motivation and Context / Related issue

Fixes #813

When DataLoader uses WeightedRandomSampler, privacy accounting was silently broken. The sample_rate was computed from len(data_loader) (number of batches) instead of the true dataset size, causing privacy budget to burn 100x-1000x faster than expected.

Root Cause

• sample_rate = 1 / len(data_loader) was computed AFTER _prepare_data_loader() replaced the sampler
• For WeightedRandomSampler with num_samples=128 and batch_size=16 on a 100k dataset:
  • Buggy: sample_rate = 1/8 = 0.125 (8 batches)
  • Correct: sample_rate = 16/100000 = 0.00016
  • Result: 781x overestimate → epsilon burns 781x faster than user expects

Fix

Capture true_dataset_size and original_batch_size BEFORE _prepare_data_loader() modifies the sampler, ensuring privacy accounting uses the actual dataset size.

How Has This Been Tested

• Verified with the reproduction case from issue WeightedRandomSampler causes silent epsilon miscalculation) #813
• Sample rate now correctly computes to 0.00016 instead of 0.125
• Expected batch size now correctly set to 16 instead of 12500
• Ran existing test suite to ensure no regressions

Checklist

• The documentation is up-to-date with the changes I made.
• I have read the CONTRIBUTING document and completed the CLA (see CONTRIBUTING).
• All tests passed, and additional code has been covered with new tests.

[meta-cla]

Hi @MN-Noor!

Thank you for your pull request and welcome to our community.

Action Required

In order to merge any pull request (code, docs, etc.), we require contributors to sign our Contributor License Agreement, and we don't seem to have one on file for you.

Process

In order for us to review and merge your suggested changes, please sign at https://code.facebook.com/cla. If you are contributing on behalf of someone else (eg your employer), the individual CLA may not be sufficient and your employer may need to sign the corporate CLA.

Once the CLA is signed, our tooling will perform checks and validations. Afterwards, the pull request will be tagged with CLA signed. The tagging process may take up to 1 hour after signing. Please give it that time before contacting us about it.

If you have received this in error or have any questions, please contact us at cla@meta.com. Thanks!

[meta-cla]

Thank you for signing our Contributor License Agreement. We can now accept your code for this (and any) Meta Open Source project. Thanks!

[meta-codesync]

@facebook-github-bot has imported this pull request. If you are a Meta employee, you can view this in D100263104. (Because this pull request was imported automatically, there will not be any future comments.)