Why normalize image like this? The normalization parameter in example turorial

Recently, I am learning about image classification.
I have read several tutorial and examples.
But one thing bother me is that why everyone using
Normalize([0.485, 0.456, 0.406], [0.229, 0.224, 0.225]

And no one give explanation, so i just wondering why.
I tried other normalization, but it also works fine.

1 Like

[0.485, 0.456, 0.406], [0.229, 0.224, 0.225] are means and stds computed on the whole training datasets.

4 Likes