Adding tensors of different sizes

1

torch can add tensors of different sizes. But it is not clear what is the rule when two tensors of different sizes are added.

import torch
import numpy
x = torch.from_numpy(numpy.array([[1,2,3], [4, 5, 6]]))
y = torch.from_numpy(numpy.array([1,2,3]))
print x
print y
print torch.add(x, y)

2

It looks like it tries to broadcast the tensors so they’re the same size and then adds them together. I agree the documentation is a little unclear about this.

1 Like