My configurations: Ubuntu 16.04, Python 3.5.2, PyTorch 1.3.1, CUDA 10.1, CUDNN 7.6.5, numpy 1.15.4, opencv 3.4.4, matplotlib 3.0.2.
You should also install the library RoiAlign.
Download images and models from Dropbox or Baidu disk, then run
python run.py
All the detection results are in the "results" directory.
The entire document structure:
