def test_delegation_transpose() -> None:
# this delegates transpose down to the lowest child,
# and wraps the result using the same class tree as the original object
t = SyftTensor.FloatTensor([[4, 5, 6]])
t_t = t.transpose(0, 1)
assert get_children_types(t) == [SyftTensor, FloatTensor, DataTensor, torch.Tensor]
assert get_children_types(t_t) == [
SyftTensor,
FloatTensor,
DataTensor,
torch.Tensor,
]
assert isinstance(t_t, SyftTensor)
assert all(t_t.data == torch.tensor([[4.0], [5.0], [6.0]]))
def test_matmul_floattensor() -> None:
t1 = SyftTensor.FloatTensor([[1.0, 2.0, 3.0], [4.0, 5.0, 6.0]])
t2 = SyftTensor.FloatTensor([[1.0, 2.0], [3.0, 4.0], [5.0, 6.0]])
t3 = t1 @ t2
assert torch.equal(t3.data, torch.tensor([[22.0, 28.0], [49.0, 64.0]]))
def test_matmul_inttensor() -> None:
t1 = SyftTensor.IntegerTensor([[1.0, 2.0, 3.0], [4.0, 5.0, 6.0]])
t2 = SyftTensor.IntegerTensor([[1.0, 2.0], [3.0, 4.0], [5.0, 6.0]])
t3 = t1 @ t2
assert torch.equal(t3.data, torch.IntTensor([[22, 28], [49, 64]]))
def test_division_floattensor() -> None:
t1 = SyftTensor.FloatTensor([6, 9, 12])
integer = 3
t3 = t1 / integer
assert torch.equal(t3.data, torch.tensor([2.0, 3.0, 4.0]))
def test_multiplication_floattensor() -> None:
t1 = SyftTensor.FloatTensor([1, 2, 3])
t2 = SyftTensor.FloatTensor([4, 5, 6])
t3 = t1 * t2
assert torch.equal(t3.data, torch.tensor([4.0, 10.0, 18.0]))
def test_subtraction_floattensor() -> None:
t1 = SyftTensor.FloatTensor([1, 2, 3])
t2 = SyftTensor.FloatTensor([4, 5, 6])
t3 = t1 - t2
assert isinstance(t3, SyftTensor)
assert torch.equal(t3.data, torch.tensor([-3.0, -3.0, -3.0]))
def test_addition_floattensor() -> None:
t1 = SyftTensor.FloatTensor([1, 2, 3])
t2 = SyftTensor.FloatTensor([4, 5, 6])
t3 = t1 + t2
assert isinstance(t3, SyftTensor)
assert torch.equal(t3.data, torch.tensor([5.0, 7.0, 9.0]))
def test_children() -> None:
t = SyftTensor.FloatTensor([1, 2, 3])
assert get_children_types(t) == [SyftTensor, FloatTensor, DataTensor, torch.Tensor]