Fix: Add null check for num_features parameter in BatchNorm1D

[prjanitor]

Bug Description

The num_features parameter in BatchNorm1D defaults to None, which causes PyTorch's BatchNorm1d to raise an error when instantiated without this required parameter.

Root Cause

The validation logic only checked if num_features was an integer when it was not None:

if num_features is not None and not isinstance(num_features, int):
    raise ValueError("Please provide a valid num_features")

This allowed num_features=None to pass validation, but PyTorch requires this parameter.

Fix

Changed the validation to also check for None:

if num_features is None or not isinstance(num_features, int):
    raise ValueError("Please provide a valid num_features")

Testing

Users will now receive a clear error message when attempting to create a BatchNorm1D layer without specifying num_features.

---

This PR was generated by PRJanitor — an automated tool that finds and fixes small bugs in open-source projects.

We respect your contribution guidelines — if your project doesn't accept bot PRs, we won't send more. You can also add a .github/prjanitor.yml file with enabled: false to opt out explicitly.