I have tried with some github implementation on YOLOv3 in tensorflow. I trained yolov3 for faces with WIDER face dataset, I haven't changed the original configuration of YOLOv3. After training the loss didn't improve after 14.34 and after few epochs it becomes NaN.Currently, for WIDER faces the AP is 71.16%, and detection performance for small faces is not good. I want to change the hyperparameters of YOLOv3 to improve the loss and better detection accuracy. I am new to Deep Learning and CNN. Any kind of suggestions is appreciated.