refactor: copy ensemble_boxes_nms from ensemble_boxes package

[RanbirAulakh]

Issue #, if available: n/a

Notes

This file contains code for performing non-maximum suppression (NMS) on bounding boxes obtained from multiple object detection models. The code was originally taken from the following GitHub repository:

https://github.com/ZFTurbo/Weighted-Boxes-Fusion/blob/master/ensemble_boxes/ensemble_boxes_nms.py

The original author is 'ZFTurbo' (https://kaggle.com/zfturbo).

This implementation provides functions for standard NMS, linear soft-NMS, and gaussian soft-NMS. It also includes a method for preparing the boxes, scores, and labels before applying NMS.

Along with some motification to match the conventional OSML styles such as typings and docstrings.

Checklist

Before you submit a pull request, please make sure you have the following:

• Code changes are compact and well-structured to facilitate easy review
• Changes are documented in the README.md and other relevant documentation pages
• PR title and description accurately reflect the changes and are detailed enough for historical tracking
• PR contains tests that cover all new code and the code has been manual tested
• All new dependencies are declared (if any), and no unnecessary libraries are added
• Performance impacts (if any) of the changes are evaluated and documented
• Security implications of the changes (if any) are reviewed and addressed
• I have read the Contributing Guidelines and agree to follow the Code of Conduct

By submitting this pull request, I confirm that you can use, modify, copy, and redistribute this contribution, under the terms of your choice.

[drduhe]

LGTM