def test_input_layer_forward_shape_inner_check():
input_layer = Input(name='input_layer', shape=(2, ))
input_data = np.array([[1, 2, 3]])
with pytest.raises(
ValueError,
match=f'Dimension 3 does not match expected dimension of 2'):
input_layer.forward(input_data)
def test_input_layer_forward_shape_outer_check():
# Test that number of dimensions of `Input` layer matches number of dimensions passed to `forward`
input_layer = Input(name='input_layer', shape=(2, ))
input_data = np.array([0, 1])
with pytest.raises(
ValueError,
match=re.escape(
'Inputs must be of at least dimension 2 (including batch)')):
input_layer.forward(input_data)
input_data = np.array([[[0, 1]]])
with pytest.raises(
ValueError,
match=
'Number of dimensions must be consistent with input layer shape'):
input_layer.forward(input_data)
input_layer = Input(name='input_layer', shape=(2, 3))
input_data = np.array([[1, 2]])
with pytest.raises(
ValueError,
match=
'Number of dimensions must be consistent with input layer shape'):
input_layer.forward(input_data)
def test_input_layer_forward_numpy_type_check():
# Test that input argument to `forward` is a NumPy array
input_layer = Input(name='input_layer', shape=(2, ))
input_data = [0, 1]
with pytest.raises(ValueError,
match='Inputs must be of type numpy.ndarray'):
input_layer.forward(input_data)
input_data = 5
with pytest.raises(ValueError,
match='Inputs must be of type numpy.ndarray'):
input_layer.forward(input_data)
def test_input_layer_forward_numpy_inner_type_check():
# Test that NumPy input argument to `forward` contains int32, int64, float32, float64 values
input_layer = Input(name='input_layer', shape=(2, ))
input_data = np.array(['a', 'b'])
with pytest.raises(
ValueError,
match='Inputs must be of type int32, int64, float32, float64'):
input_layer.forward(input_data)
input_data = np.array([['e']])
with pytest.raises(
ValueError,
match='Inputs must be of type int32, int64, float32, float64'):
input_layer.forward(input_data)
input_data = np.array([[5]], dtype=np.int16)
with pytest.raises(
ValueError,
match='Inputs must be of type int32, int64, float32, float64'):
input_layer.forward(input_data)
def test_input_layer_forward_proper():
input_layer = Input(name='input_layer', shape=(2, ))
input_data = np.array([[1, 2]])
expected = np.array([[1, 2]])
res = input_layer.forward(input_data)
assert np.array_equal(res, expected)
def test_input_layer_name():
# Test that `name` parameter gets constructed properly from base `Layer` class
input_layer = Input(name='input_layer', shape=(256, ))
assert input_layer.name == 'input_layer'
def test_input_layer_shape():
# Test that `shape` parameter gets constructed properly
input_layer = Input(name='input_layer', shape=(256, ))
assert input_layer.output_shape == (256, )