output range difference between darknet and keras-yolo3

[choucl]

Hi there,
I tried to convert my tiny yolo to keras, and then convert it to tensorflow lite for edge devices. So I tried to draw the bounding box on my own. However I found out that the output of model is quite different from the output of darknet's. The darknet model output range is between 0 to 1, but the output of keras-yolo3 has some negative numbers inside of it and not 0 to 1. I'm wondering if the output is same as the output of darknet? or I need to do any modification to make it same as the darknet's?