Cloned from experiencor/keras-yolo3. See original repo for full doc. Also recommended is the tutorial How to Perform Object Detection With YOLOv3 in Keras.
All I did was rewriting yolo3_one_file_to_detect_them_all.py into y3detect.py , making uses of utils/*.py as much as possible. The model conversion code is removed. Use convert.py from qqwweee/keras-yolo3 to do the conversion just once.
wget https://pjreddie.com/media/files/yolov3.weights.
python3 qqwweee-keras-yolo3/convert.py qqwweee-keras-yolo3/yolov3.cfg yolov3.weights yolo.h5
python3 y3detect.py -w yolo.h5 -o /tmp/result/ *.jpg