I have already tried the sklearn.preprocessing.scale(), but it doesn’t work. The error information is ValueError: Found array with dim 4. the scale function expected <= 2.
You could maybe use
torchvision.transforms.Normalize or write the normalization manually.
E.g. if you want to standardize the input you could subtract the mean and divide by the stddev.
Feel free to add more information about your use case, if you get stuck.
Thanks！ I will try it by your method