concatenating a list of 2d-tensor along different two axis respectively, leads to tensor A,B where A.T==B, but their mean values along the same axis is slightly different (A.T.mean(axis=0) != B.mean(axis=0)), why? Theoretically they are the same.

concatenating a list of 2d-tensor along different two axis respectively, leads to tensor A,B where A.T==B, but their mean values along the same axis is slightly different (A.T.mean(axis=0) != B.mean(axis=0)), why? Theoretically they are the same.

How come pytorch makes such a weird mistake in higher digits in division?

I have a concatenated tensor that represents a batch of point clouds with pointwise features.
Shape: (batch_size * num_points, feature_dim)

I have a source tensor of size (3, 2) and an index tensor of size (3, 3) containing integer values 0, 1, or 2. In pytorch I can do tensor indexing source[index] to get a tensor of size (3, 3, 2). Example: