DetNet

pytorch-1.1 torchvision-0.3

References

• https://github.com/guoruoqian/DetNet_pytorch
• https://github.com/jwyang/fpn.pytorch
• https://github.com/jwyang/faster-rcnn.pytorch/tree/pytorch-1.0

Prerequisites

• Python 2.7 or 3.6
• Pytorch 1.1 or higher
• CUDA 9.0 or higher
• tensorboardX

Preparation

First of all, clone the code

git clone https://github.com/leowangzi/DetNet.git
cd DetNet

Then, create a folder:

mkdir data

or

ln -s [source_data] data

What we are doing and going to do

• Support pytorch-1.1 (master).
• Support torchvision-0.3 (master).
• soft-nms

Benchmarking

I benchmark this code thoroughly on pascal voc2007 and 07+12. Below are the results:

1). PASCAL VOC 2007 (Train/Test: 07trainval/07test, scale=600, ROI Align)

model（FPN）	GPUs	Batch Size	lr	lr_decay	max_epoch	Speed/epoch	Memory/GPU	mAP
DetNet59	1 GTX 1080 (Ti)	2	1e-3	10	12	1.07hr	5412MB	75.5

2). PASCAL VOC 07+12 (Train/Test: 07+12trainval/07test, scale=600, ROI Align)

model（FPN）	GPUs	Batch Size	lr	lr_decay	max_epoch	Speed/epoch	Memory/GPU	mAP
DetNet59	1 GTX 1080 (Ti)	1	1e-3	10	12	2.33hr	8015MB	81.5