• add data augmentation
• add Multi-scale training
• add onnx (doing)

• train code in : https://github.com/ouyanghuiyu/Snet

• pytorch 1.2.0
• torchvision 0.4

git clone https://github.com/ouyanghuiyu/Thundernet_Pytorch.git

cd lib && python setup.py  build_ext --inplace
cd psroialign/PSROIAlign && sh build.sh 

Download VOC0712 datasets ln -s "YOUR PATH" data

cd script
sh  train_49.sh
sh  train_146.sh
sh  train_535.sh

cd script
sh  pre.sh

• add coco train and test
• add NCNN inference

Please cite the paper in your publications if it helps your research:

@article{zheng2019thundernet,
  title={ThunderNet: Towards Real-time Generic Object Detection},
  author={Zheng Qin, Zeming Li,Zhaoning Zhang,Yiping Bao,Gang Yu, Yuxing Peng, Jian Sun},
  journal={arXiv preprint arXiv:1903.11752},
  year={2019}
}