/faster_rcnn_pytorch

based on jwyang/faster_rcnn.pytorch

faster_rcnn_pytorch

based on jwyang/faster_rcnn.pytorch

Prerequisities

  • Python 3.6.8
  • Pytorch 1.2.0
  • cuda 10.1
  • cudnn 7.6.2

Preparation

git clone -b pytorch-1.0 https://github.com/jwyang/faster-rcnn.pytorch.git

cd faster-rcnn.pytorch && mkdir data

Dataset

MSCOCO and VOC2012

Pretrained Model

vgg16 and resnet101

Compilation

Install all the python dependencies using pip:

pip3 install -r requirements.txt

Python modules are as follow:

  • cython
  • cffi
  • opencv-python
  • scipy==1.2.0
  • msgpack
  • easydict
  • matplotlib
  • pyyaml
  • tensorboardX

Compile the cuda dependencies using following simple commands:

cd lib
python3 setup.py build develop

All the necessary C modules has been compiled through setup.py, such as NMS, ROI_Pooling, ROI_Align, ROI_Crop