CascadeRCNN implementation using PyTorch

Install

• pip install mmcv-full==1.5.1
• pip install mmdet==2.24.1

Train

• bash ./main.sh ./nets/exp01.py $ --train for training, $ is number of GPUs

Results

Detector	Backbone	Neck	LR Schedule	Box mAP	Mask mAP	Config
Cascade R-CNN	Swin-T	FPN	1x	-	-	exp01
Cascade R-CNN	Swin-T	FPN	3x	-	-	exp02
Cascade R-CNN	Swin-T	FPN	3x	-	-	exp03
Cascade R-CNN	Swin-T	FPN	3x	-	-	exp04
Cascade R-CNN	Swin-T	PAFPN	3x	-	-	exp05
Cascade R-CNN	Swin-T	PAFPN + DyHead	3x	-	-	exp06

TODO

• exp01, default Cascade R-CNN detector
• exp02, added MOSAIC and MixUp
• exp03, added GN and WS
• exp04, added PolyLoss
• exp05, added PAFPN
• exp06, added DyHead

Reference

• https://github.com/ultralytics/yolov5
• https://github.com/open-mmlab/mmdetection