* upload vityolo model

* add DINO

add msdeformable attention cuda op

fix export bug

* fix decoder norm

* add python example for face detection part

为了方便二次开发，这里添加了直接使用Python脚本的形式进行人脸检测的示例。

* change the demo image

* add python example for face detection part in English readme file

sync the same face detection python example from the Chinese readme file

* fix a link bug

* add python example for face detection part

为了方便二次开发，这里添加了直接使用Python脚本的形式进行人脸检测的示例。

* change the demo image

* add yolov3_darknet_voc model and docs

* add yolov3 darknet53 voc 416 320 mAP, test=document_fix

* fix config, test=document_fix

* finish_cotuning

* fix_bug

* add_readme

* finish contrast

* add yml for single npu and polish call to paddle api

* add yml for single npu and polish call to paddle api

* add CWD distill code

* fix type

* fix bugs

* add retinanet teacher yml

* fix comments

* add sku110k models

* add link in ppyoloe doc

* modify directory of sku110k

* fix data config problem

* add semi cocoo txt file and fix doc, test=document_fix

* add ssld vd in fcos doc, test=document_fix

* fix save_model epoch_id and rename folder

* fix use_simple_ema and docs

* fix docs, test=document_fix

* add ssod denseteacher fcos codes and refine baseline doc

* fix docs

* fix deploy and ColorJitter float32

* fix docs

* fix docs, test=document_fix

* fix dead links, test=document_fix

* fix docs, test=document_fix

* add iou-fcos and ssod baseline

* add gen_semi_coco, fix doc

* fix doc links

* remove duplicate backbone variant, test=document_fix

* remove smalldet cfgs and fix doc

* add speed test doc and deploy model links

* fix doc

* add smalldet demo jpg

* remove duplicate backbone variant, test=document_fix

* add ppyoloe+ voc

* fix doc

* remove duplicate backbone variant, test=document_fix

* fix slice_image deploy infer index

* fix sm_use for ppyoloe_head and output for box_dis

* fix pos_embed of training

* remove duplicate backbone variant, test=document_fix

* add retinanet_r50_2x

* [smdet] add tal_cr and custompan with transfomer, add box_distribution to pred reg_max, test=document_fix

* add tal_cr