Pelee: A Real-Time Object Detection System on Mobile Devices, in PyTorch

A PyTorch implementation of Pelee: A Real-Time Object Detection System on Mobile Devices The official and original Caffe code can be found here.

Description

I train Pelee with pytorch and the result is better than the original paper result,the pretrained model can be downloaded in peleenet.pth.

MAP in VOC2007

Method	07+12	07+12+coco
SSD300	77.2	81.2
SSD+MobileNet	68	72.7
Original Pelee	70.9	76.4
Ours Pelee	71.76	---

Preparation

the supported version is pytorch-0.4.1 or pytorch-1.0

• tqdm
• opencv
• addict
• pytorch>=0.4

• Clone this repository.

git clone https://github.com/yxlijun/Pelee.Pytorch

• Compile the nms and coco tools:

sh make.sh

• Prepare dataset (e.g., VOC, COCO), refer to ssd.pytorch for detailed instructions.

train

you can train different set according to configs/*,First, you should download the pretrained model peleenet.pth,then,move the file to weights/

python train.py --dataset VOC\COCO --config ./configs/Pelee_VOC.py  

if you train with multi gpu

CUDA_VISIBLE_DEVICES=0,1 python train.py --dataset VOC\COCO --config ./configs/Pelee_VOC.py   --ngpu 2

eval

you can evaluate your model in voc and coco

python test.py --dataset VOC\COCO  --config ./configs/Pelee_VOC.py --trained_model ./weights/Pelee_VOC.pth 

demo

you can test your image, First, download the trained model Pelee_VOC.pth file. Then, move the file to weights/.

python demo.py --dataset VOC\COCO  --config ./configs/Pelee_VOC.py --trained_model ./weights/Pelee_VOC.pth --show  

You can see the image with drawed boxes as:

TODO

the code support:

• Support for the MS COCO dataset and VOC PASCAL dataset
• Support for Pelee304_VOC、Pelee304_COCO training and testing
• Support for mulltigpu training
• Support training and and testing in VOC and COCO

References

• Pelee: A Real-Time Object Detection System on Mobile Devices
• M2Det
• Pelee