CAFFE for YOLO

Reference

YOLO9000: Better, Faster, Stronger

http://pjreddie.com/yolo9000/

Usage

Data preparation

  cd data/yolo
  ln -s /your/path/to/VOCdevkit/ .
  python ./get_list.py
  # change related path in script convert.sh
  ./convert.sh 

Train

  cd examples/yolo/darknet_v3
  # change related path in script train.sh
  mkdir models
  ./train_darknet_v3.sh

Test (The first way online)

  # if everything goes well, mAP reach ~56.
  cd examples/yolo/darknet_v3
  ./test_darknet_v3.sh

Test (The second way offline)

  cd examples/eval_detection
  python test_yolo_v2.py