About DataLoader

[kingsleyluoxin]

Hello, I noticed that in your code data_loader.py, you loaded the source and target labels as

lab_source = Image.open(lab_source_path) lab_source = lab_source.resize( [self.opt.loadSize[0], self.opt.loadSize[1]], Image.BICUBIC) img_source, lab_source, scale = paired_transform( self.opt, img_source, lab_source) lab_source = self.transform_no_augment(lab_source)

When transforming lab_source, transforms.Normalize((0.5, 0.5, 0.5), (0.5, 0.5, 0.5)) is applied.

However, isn't this transforming expected to be applied to a 3-channel image? Should we use transforms.Normalize(0.5, 0.5) here?

[lyndonzheng]

Hi @kingsleyluoxin , the transforms.Normalize(mean, std) is applied to normalize the input as (x-mean)/std. For the image translation, we normalize the input image from [0,1] to [-1,1].

[kingsleyluoxin]

Thanks for the explanation. I do understand the function here :)

However, I think the parameters here should be (0.5, 0.,5) instead of ((0.5, 0.5, 0.5), (0.5, 0.5, 0.5)), since the depth image is single channel. :)