[resolved] How to normalize images to use in pertained networks, e.g. alexnet

vision
1

I wanted to use pertained alexnet from “torchvision.models.alexnet”, I was wondering how should I normalize my images. The images are numpy arrays with size 3,224,224 and with values between 0 to 255 (inclusive). Thanks.

2

Never mind, I found the answer here “https://github.com/pytorch/vision#models”.