FasterRCNN

Renferences

• https://github.com/jwyang/faster-rcnn.pytorch/tree/pytorch-1.0
• https://github.com/ruotianluo/pytorch-faster-rcnn
• https://github.com/chengsq/pytorch-lighthead

Prerequisites

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

First of all, clone the code

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

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).

Benchmarking

We benchmark our code thoroughly on pascal voc datasets, using resnet101 network architecture. Below are the results:

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

model	#GPUs	batch size	lr	lr_decay	max_epoch	time/epoch	mem/GPU	mAP
[Res-101]	1	1	1e-3	5	10	0.88 hr	5965 MB	75.06

2). PASCAL VOC 2007&2012 (Train/Test: 07+12trainval/07test, scale=600, ROI Align)

model	#GPUs	batch size	lr	lr_decay	max_epoch	time/epoch	mem/GPU	mAP
[Res-101]	1	1	1e-3	5	10	0.88 hr	3200 MB	79.80