NASFPN

NAS-FPN

This repository implements NAS-FPN in the SimpleDet framework.

Qucik Start

# train baseline retinanet following the setting of NAS-FPN
python3 detection_train.py --config config/NASFPN/retina_r50v1b_fpn_640_1@256_25epoch.py

# train NAS-FPN
python3 detection_train.py --config config/NASFPN/retina_r50v1b_nasfpn_640_7@256_25epoch.py
python3 detection_train.py --config config/NASFPN/retina_r50v1b_nasfpn_1024_7@256_25epoch.py
python3 detection_train.py --config config/NASFPN/retina_r50v1b_nasfpn_1280_7@384_25epoch.py

# train hand-crafted neck
python3 detection_train.py --config config/NASFPN/retina_r50v1b_tdbu_1280_3@384_25epoch.py

Results and Models

All AP results are reported on test-dev of the COCO dataset.

Model	InputSize	Backbone	Neck	Train Schedule	GPU	Image/GPU	FP16	Train MEM	Train Speed	Box AP(Mask AP)	Link
RetinaNet	640	R50v1b-FPN	1@256	25 epoch	8X 1080Ti	8	yes	6.6G	85 img/s	37.4	model
NAS-FPN	640	R50v1b-FPN	7@256	25 epoch	8X 1080Ti	8	yes	7.8G	66 img/s	40.1	model
NAS-FPN	1024	R50v1b-FPN	7@256	25 epoch	8X 1080Ti	4	yes	9.1G	17 img/s	44.2	model
NAS-FPN	1280	R50v1b-FPN	7@384	25 epoch	8X 1080Ti	2	yes	8.9G	10 img/s	45.3	model
TD-BU*	1280	R50v1b-FPN	3@384	25 epoch	8X 1080Ti	3	yes	10.5G	12 img/s	44.7	model

* Short for TopDown-BottomUp neck which is highly symmetric proposed by Zehao.

Reference

@inproceedings{ghiasi2019fpn,
  title={NAS-FPN: Learning Scalable Feature Pyramid Architecture for Object Detection},
  author={Ghiasi, Golnaz and Lin, Tsung-Yi and Pang, Ruoming and Le, Quoc V},
  booktitle={CVPR},
  year={2019}
}