def test_vector_float64(self):
a = np.array([1.5, 2.5, 3.5]).astype(np.float64)
res = tensor(a)
np.testing.assert_array_equal(res.to_cpu(), a)
assert res.dtype == np.float64
assert res.shape == a.shape
assert res._shape_gpu == (len(a), )
def test_2d_matrix(self):
a = np.array([[1, 2, 3], [1, 2, 3]])
res = tensor(a)
np.testing.assert_array_equal(res.to_cpu(), a)
assert res.dtype == np.int64
assert res.shape == a.shape
assert res._shape_gpu == (np.prod(a.shape), )
def test_vector_int32(self):
a = np.array([1, 2, 3]).astype(np.int32)
res = tensor(a)
np.testing.assert_array_equal(res.to_cpu(), a)
assert res.dtype == np.int32
assert res.shape == a.shape
assert res._shape_gpu == (len(a), )
class TestMagicMethods:
@pytest.mark.parametrize(
'tens, other, gold',
[(tensor([1, 2, 3]), 1, np.array([2, 3, 4])),
(tensor([1, 2, 3]), tensor([1, 2, 3]), np.array([2, 4, 6]))])
def test_add(self, tens, other, gold):
np.testing.assert_array_equal((tens + other).to_cpu(), gold)
@pytest.mark.parametrize(
'tens, other, gold',
[(tensor([1, 2, 3]), 2, np.array([2, 4, 6])),
(tensor([1, 2, 3]), tensor([1, 2, 3]), np.array([1, 4, 9]))])
def test_mul(self, tens, other, gold):
np.testing.assert_array_equal((tens * other).to_cpu(), gold)
@pytest.mark.parametrize(
'tens1, tens2, gold',
[(tensor([1, 2, 3]), tensor(
[1, 2, 3]), np.array([[1, 2, 3], [2, 4, 6], [3, 6, 9]])),
(tensor([[1, 2, 3], [1, 2, 3], [1, 2, 3]
]), tensor([[1, 2, 3], [1, 2, 3], [1, 2, 3]]),
np.array([[6, 12, 18], [6, 12, 18], [6, 12, 18]]))])
def test_pow(self, tens1, tens2, gold):
np.testing.assert_array_equal((tens1**tens2).to_cpu(), gold)
class TestTensorSumOperations:
@pytest.mark.parametrize(
'tens, add, gold',
[(tensor([1, 2, 3]), 2, np.array([3, 4, 5])),
(tensor([1, 2, 3], dtype=np.float64), 2.5, np.array([3.5, 4.5, 5.5])),
(tensor([[[1, 2, 3], [1, 2, 3]], [[1, 2, 3], [1, 2, 3]]]), 2,
np.array([[[3, 4, 5], [3, 4, 5]], [[3, 4, 5], [3, 4, 5]]]))])
def test_add_scalar(self, tens, add, gold):
np.testing.assert_array_equal(tens.scalar_sum(add).to_cpu(), gold)
@pytest.mark.parametrize(
'tens1, tens2, gold',
[(tensor([1, 2, 3]), tensor([1, 2, 3]), np.array([2, 4, 6])),
(tensor([[[1, 2, 3], [1, 2, 3]], [[1, 2, 3], [1, 2, 3]]
]), tensor([[[1, 2, 3], [1, 2, 3]], [[1, 2, 3], [1, 2, 3]]]),
np.array([[[2, 4, 6], [2, 4, 6]], [[2, 4, 6], [2, 4, 6]]]))])
def test_tensor_sum(self, tens1, tens2, gold):
np.testing.assert_array_equal(tens1.tensor_sum(tens2).to_cpu(), gold)
class TestTensorProductOperations:
@pytest.mark.parametrize(
'tens, mult, gold',
[(tensor([1, 2, 3]), 2, np.array([2, 4, 6])),
(tensor([1, 2, 3], dtype=np.float64), 2.5, np.array([2.5, 5, 7.5])),
(tensor([[[1, 2, 3], [1, 2, 3]], [[1, 2, 3], [1, 2, 3]]]), 2,
np.array([[[2, 4, 6], [2, 4, 6]], [[2, 4, 6], [2, 4, 6]]]))])
def test_mult_scalar(self, tens, mult, gold):
np.testing.assert_array_equal(tens.scalar_mult(mult).to_cpu(), gold)
@pytest.mark.parametrize(
'tens1, tens2, gold',
[(tensor([1, 2, 3]), tensor([1, 2, 3]), np.array([1, 4, 9])),
(tensor([[[1, 2, 3], [1, 2, 3]], [[1, 2, 3], [1, 2, 3]]
]), tensor([[[1, 2, 3], [1, 2, 3]], [[1, 2, 3], [1, 2, 3]]]),
np.array([[[1, 4, 9], [1, 4, 9]], [[1, 4, 9], [1, 4, 9]]]))])
def test_inner_product(self, tens1, tens2, gold):
np.testing.assert_array_equal(
tens1.inner_product(tens2).to_cpu(), gold)
@pytest.mark.parametrize(
'tens1, tens2, gold',
[(tensor([1, 2, 3]), tensor(
[1, 2, 3]), np.array([[1, 2, 3], [2, 4, 6], [3, 6, 9]])),
(tensor([1, 2, 3, 4]), tensor([1, 2, 3]),
np.array([[1, 2, 3], [2, 4, 6], [3, 6, 9], [4, 8, 12]])),
(tensor([1, 2, 3]), tensor([1, 2, 3, 4]),
np.array([[1, 2, 3, 4], [2, 4, 6, 8], [3, 6, 9, 12]]))])
def test_outer_product(self, tens1, tens2, gold):
np.testing.assert_array_equal(
tens1.outer_product(tens2).to_cpu(), gold)
@pytest.mark.parametrize(
'tens1, tens2, gold',
[(tensor([[1, 2, 3], [1, 2, 3], [1, 2, 3]
]), tensor([[1, 2, 3], [1, 2, 3], [1, 2, 3]]),
np.array([[6, 12, 18], [6, 12, 18], [6, 12, 18]])),
(tensor([[1], [2], [3]]), tensor([[1, 2, 3, 4]]),
np.array([[1, 2, 3, 4], [2, 4, 6, 8], [3, 6, 9, 12]])),
(tensor([[1, 2], [1, 2], [1, 2]]), tensor([[1, 2, 3, 4], [
1, 2, 3, 4
]]), np.array([[3, 6, 9, 12], [3, 6, 9, 12], [3, 6, 9, 12]]))])
def test_matmul(self, tens1, tens2, gold):
np.testing.assert_array_equal(tens1.matmul(tens2).to_cpu(), gold)