README.md

Faster R-CNN: Towards Real-Time Object Detection with Region Proposal Networks

Introduction

@article{Ren_2017,
   title={Faster R-CNN: Towards Real-Time Object Detection with Region Proposal Networks},
   journal={IEEE Transactions on Pattern Analysis and Machine Intelligence},
   publisher={Institute of Electrical and Electronics Engineers (IEEE)},
   author={Ren, Shaoqing and He, Kaiming and Girshick, Ross and Sun, Jian},
   year={2017},
   month={Jun},
}

Results and models

Backbone	Style	Lr schd	Mem (GB)	Inf time (fps)	box AP	Download
R-50-FPN	caffe	1x	3.8		37.8	model | log
R-50-FPN	pytorch	1x	4.0	21.4	37.4	model | log
R-50-FPN	pytorch	2x	-	-	38.4	model | log
R-101-FPN	caffe	1x	5.7		39.8	model | log
R-101-FPN	pytorch	1x	6.0	15.6	39.4	model | log
R-101-FPN	pytorch	2x	-	-	39.8	model | log
X-101-32x4d-FPN	pytorch	1x	7.2	13.8	41.2	model | log
X-101-32x4d-FPN	pytorch	2x	-	-	41.2	model | log
X-101-64x4d-FPN	pytorch	1x	10.3	9.4	42.1	model | log
X-101-64x4d-FPN	pytorch	2x	-	-	41.6	model | log

Different regression loss

We trained with R-50-FPN pytorch style backbone for 1x schedule.

Backbone	Loss type	Mem (GB)	Inf time (fps)	box AP	Download
R-50-FPN	L1Loss	4.0	21.4	37.4	model | log
R-50-FPN	IoULoss			37.9	model | log
R-50-FPN	GIoULoss			37.6	model | log
R-50-FPN	BoundedIoULoss			37.4	model | log

Pre-trained Models

We also train some models with longer schedules and multi-scale training. The users could finetune them for downstream tasks.

Backbone	Style	Lr schd	Mem (GB)	Inf time (fps)	box AP	Download
R-50-FPN	caffe	2x	4.3		39.7	model | log
R-50-FPN	caffe	3x	4.3		40.2	model | log