This is new yolo annotation tool which is added new features. I have posted three blogs for how to train yolov2 and v3 using our custom images.

You can follow three steps for annotate the image to yolo.
Steps:

run main.py
run convert.py
run process.py

Now i have implemented the convert.py process in main.py. So you don't run the convert.py for yolo annotating.
Do steps:

run main.py
run process.py

NOTE: If you use new annotation tool, please create classes.txt file and write all classes what you train the objects. Because i read the all classes from classes.txt.

The dataset is ready for yolo training.