I have multiple tensors with different shapes
tensor_1 = torch.tensor([[1.0, 8.0, 31.0], [3.0, 4.0, 0.0]])
tensor_2 = torch.tensor([[1.0, 8.0, 31.0, 212.1], [3.0, 11.2, 4.0, 0.0]])
How can I normalize/shift (not entirely sure what to call it) each of these tensors to be on an N dimensional unit circle/sphere, such that the direction/angle of each of the vectors will be the same (each tensor should have its own unit circle clearly). I found this SO question about how to normalize a single vector, but I'm not entirely sure if mathematically normalizing each one separately will yield the same results
