Rescale image data [0, 255] to [-1, 1]

Hello, as the title says, when I use torchvision.transforms.Compose, is there a way to do the tensor rescaling or redistribution method to complete it after the Compose fuction ingredient transforms.ToTensor()?

Given that torchvision.transform.ToTensor already standardize the data to [0, 1], you could just multiply that by 2 and subtract 1 on your input layer.

Ok, thank you for reminding what ToTensor have done.