[Feature]: add selfsup performance config

[ZwwWayne]

Thanks for your contribution and we appreciate it a lot. The following instructions would make your pull request more healthy and more easily get feedback. If you do not understand some items, don't worry, just make the pull request and seek help from maintainers.

Motivation

Provide standard benchmark configs for evaluating pre-training methods.

Modification

1. Add configs that have exact same performance as that in detectron2 to ease the benchmark with unsupervisely pre-trained models.
2. Add precise BN to align the performance of caffe models. It has little affect to the performance of PyTorch backbones.

BC-breaking (Optional)

N/A

Use cases (Optional)

If this PR introduces a new feature, it is better to list some use cases here, and update the documentation.

Codebase	Backbone	settings	PreciseBN	Bbox Ap	Mask AP	config
D2	caffe	IN1k pretrain, 90k	Y	39.7	35.9	config
MMDet	caffe	IN1k pretrain, 90k	N	39.6	35.5	config
	pytorch	IN1k pretrain, 90k	N	40.3		config
	caffe	IN1k pretrain, 90k	Y	39.9	35.8	config
	pytorch	IN1k pretrain, 90k	Y	40.3	36.3	config

Checklist

1. Pre-commit or other linting tools are used to fix the potential lint issues.
2. The modification is covered by complete unit tests. If not, please add more unit test to ensure the correctness.
3. If the modification has potential influence on downstream projects, this PR should be tested with downstream projects, like MMDet or MMCls.
4. The documentation has been modified accordingly, like docstring or example tutorials.