Allen-lz / pedestrian_detection

Run

1. Requirements:

   • python 3.6.8, pytorch 1.5.0, torchvision 0.6.0, cuda 10.1

2. Steps to run:

   • Step1: put the model weightrcnn_emd_simple_mge.pth on ./model/rcnn_emd_simple/outputs. then, put the test images in ./test_imgs/ or put the test video in ./test_video/

   cd tools
   python inference.py
   python video_inference.py
   

   • Note: the result will be saved in ./tools/outputs