Pytorch mobile object detection example

I tried to quantize the model but fail(details write in this post), could you quantize it?