* add faster rcnn with swin_transformer tiny

* add LiteHRNet backbone and config .YML

* test lite18-network param

acc is same with ori-model 1. fix default darkpose=ON, 2. += is not inplace

add new keypoint model Lite-HRNet

* add new keypoint model Lite-HRNet

* 1. Add description of network type; 2. use channel_shuffle in ops.py

* use normal to init conv2d

* add network type description

* add gfl model and PicoDet

Besides removing unnecessary imports, this also fixes the missing logger module in test_architectures.py within the static directory.

* modified freeze at in resnet

* freeze_at should less than num_stage

* add fairmot

* rename pretrain_weights

* add comments; reuse focal loss and convnorm

* update comments

* jde and deepsort: model

* fix name of ConvNormLayer for fcos solov2

* fix name in other head convbn

* fix yolo name, fix ttf_head name, test=document_fix

* add ttfnet enhance

* add doc

* fix pafnet training

* add ppyolov2

* fix bugs and modify docs

* modify code and doc according to review

* fix bugs while resolving conflicts

* fix name of pretrain weights

* format, test=document_fix

* add senet

* add annotations according to review

* add Distill model
* add distill+prune

* update mbv3 ssdlite

* fix ssdlite link

* Add ACfpn, `CxAM` and `CnAM` is not included here.

Add `lr_multi_list` param for SSLD pretrained models' usage.

* add yolov4

* code for cornernet

* add doc

* refine custom op

* refine code from review comments

* refine doc

* update doc

* update code

* add cornernet_squeeze_mixup_cosine config

* update code

* update code