Fix missing attention_mask in calibration dataloader

[cjluo-nv]

Summary

• When include_labels=False (the default for PTQ calibration), get_dataset_dataloader was discarding the attention_mask produced by the tokenizer and only returning input_ids.
• Without attention_mask, HuggingFace models create a full causal mask, causing padding tokens to participate in attention during calibration and skewing quantization statistics.
• This fix includes attention_mask alongside input_ids so the model correctly ignores padding tokens during calibration forward passes.

Details

In modelopt/torch/utils/dataset_utils.py, the tokenizer call at line 387 with padding=True produces both input_ids and attention_mask. The include_labels=True path (line 406) already preserves the full batch_encoded dict including attention_mask. However, the include_labels=False path was only keeping input_ids "for backward compatibility."

During the calibration forward loop (_forward_loop → _process_batch), the batch dict is unpacked as **kwargs into model.forward(). Without attention_mask, HF models default to attending to all positions including padding, which pollutes calibration statistics.

Practical impact: With batch_size=1 there is no padding so the bug is invisible. With larger batch sizes and variable-length samples, shorter sequences get padded and the effect grows.

Test plan

• Existing unit tests pass (tests/unit/torch/utils/test_dataset_utils.py)
• Pre-commit hooks pass
• [ ] Verify PTQ accuracy with batch_size > 1 on a padded calibration dataset (GPU required)

🤖 Generated with Claude Code

[github-actions]

PR Preview Action v1.8.1
🚀 View preview at https://NVIDIA.github.io/Model-Optimizer/pr-preview/pr-1261/
Built to branch gh-pages at 2026-04-14 19:50 UTC. Preview will be ready when the GitHub Pages deployment is complete.

[codecov]

Codecov Report

✅ All modified and coverable lines are covered by tests.
✅ Project coverage is 76.90%. Comparing base (5ff1d7b) to head (c40a374).
⚠️ Report is 6 commits behind head on main.

Additional details and impacted files

@@            Coverage Diff             @@
##             main    #1261      +/-   ##
==========================================
- Coverage   76.91%   76.90%   -0.02%     
==========================================
  Files         350      350              
  Lines       40481    40859     +378     
==========================================
+ Hits        31137    31423     +286     
- Misses       9344     9436      +92     

Flag	Coverage Δ
examples	44.14% <100.00%> (+1.18%)	⬆️
gpu	57.39% <100.00%> (-0.14%)	⬇️
unit	55.61% <0.00%> (+0.09%)	⬆️

Flags with carried forward coverage won't be shown. Click here to find out more.

☔ View full report in Codecov by Sentry.
📢 Have feedback on the report? Share it here.

🚀 New features to boost your workflow:

• ❄️ Test Analytics: Detect flaky tests, report on failures, and find test suite problems.