Add custom max iou assigner to prevent CPU OOM in training phase

[jaegukhyun]

Summary

When a training image contain too many gt(over > 10000), current assigner, matching anchor bbox and gt, raises CPU OOM.
This is because assigner makes too large cpu tensor matrix. To fix this problem, this PR add a new custom max iou assigner.
This new custom max iou assigner splits gt bboxes into reasonable size when gt bboxes is too large. This may lead time increase but prevent sudden cpu oom.

How to test

In my desktop which has 64GB RAM cannot train Kiemgetal dataset. However after this fix, my desktop can train Kiemgetal dataset.

Checklist

• I have added unit tests to cover my changes.​
• I have added integration tests to cover my changes.​
• I have added e2e tests for validation.
• I have added the description of my changes into CHANGELOG in my target branch (e.g., CHANGELOG in develop).​
• I have updated the documentation in my target branch accordingly (e.g., documentation in develop).
• I have linked related issues.

License

• I submit my code changes under the same Apache License that covers the project.
  Feel free to contact the maintainers if that's a concern.
• I have updated the license header for each file (see an example below).

# Copyright (C) 2023 Intel Corporation
# SPDX-License-Identifier: Apache-2.0

[sungmanc]

Thanks for the nice work. BTW, how about using our own docstring instead of the original one for CustomMaxIoUAssigner?

[goodsong81]

Thank you for the great work. Minor comment and question from my side: