Hi,
I have a 3d tensor of size:
(batch_size, 3, num_points)
each sample represents 3d coordinates (num_points, each point is x,y,z).
I want to use torchvision.transforms.Normalize to normalize each axis
meaning: the mean of x coordinate in each sample will be zero, with variance = 1 (same for y,z).
the normalization should apply to each sample individually, regardless of other samples in batch.
thanks for your help!