def test_int_neg_():
data = np.array([-1, -2, 3, 4, 5, -6])
expected = np.array([1, 2, -3, -4, -5, 6])
a = IntTensor(data)
a.neg_()
np.testing.assert_almost_equal(a.to_numpy(), expected,
decimal=decimal_accuracy, verbose=verbosity)\
def test_int_lt_():
data = np.array([1,2,3,4])
compare_data = np.array([2,2,5,1])
tensor = IntTensor(data)
compare_to = IntTensor(compare_data)
expected = np.array([1,0,1,0])
tensor.lt_(compare_to)
np.testing.assert_almost_equal(tensor.to_numpy(), expected,
decimal=decimal_accuracy, verbose=verbosity)
def test_int_neg():
data = np.array([-1, -2, 3, 4, 5, -6])
expected = np.array([1, 2, -3, -4, -5, 6])
a = IntTensor(data)
b = a.neg()
np.testing.assert_almost_equal(b.to_numpy(), expected,
decimal=decimal_accuracy, verbose=verbosity)
# a doesn't change (non-inline)
np.testing.assert_almost_equal(a.to_numpy(), data,
decimal=decimal_accuracy, verbose=verbosity)
def test_int_acos():
data = np.array([-1, 0, 1, 1, 2])
expected = np.array([3.14159265, 1.57079633, 0, 0, np.nan])
a = IntTensor(data)
b = a.acos()
np.testing.assert_almost_equal(b.to_numpy(), expected,
decimal=decimal_accuracy, verbose=verbosity)
# a doesn't change
np.testing.assert_almost_equal(a.to_numpy(), data,
decimal=decimal_accuracy, verbose=verbosity)
def test_int_cos():
data = np.array([30, 60, 90, 180])
expected = np.array([.1542515, -0.952413, -0.4480736, -0.5984601])
a = IntTensor(data)
b = a.cos()
np.testing.assert_almost_equal(b.to_numpy(), expected,
decimal=decimal_accuracy, verbose=verbosity)
# a doesn't change (non-inline)
np.testing.assert_almost_equal(a.to_numpy(), data,
decimal=decimal_accuracy, verbose=verbosity)
def test_int_reciprocal():
data = np.random.rand(3,2)
expected = np.reciprocal(data)
a = IntTensor(data)
b = a.reciprocal()
np.testing.assert_almost_equal(b.to_numpy(), expected,
decimal=decimal_accuracy, verbose=verbosity)
# a doesn't change (non-inline)
np.testing.assert_almost_equal(a.to_numpy(), data,
decimal=decimal_accuracy, verbose=verbosity)
def test_int_sin():
data = np.array([15, 60, 90, 180])
expected = np.array([0.65028784, -0.30481062, 0.89399666, -0.80115264])
a = IntTensor(data)
b = a.sin()
np.testing.assert_almost_equal(b.to_numpy(), expected,
decimal=decimal_accuracy, verbose=verbosity)
# a doesn't change
np.testing.assert_almost_equal(a.to_numpy(), data,
decimal=decimal_accuracy, verbose=verbosity)
def test_int_trace():
data = np.random.rand(5,5)
expected = data.trace()
a = IntTensor(data)
b = a.trace()
np.testing.assert_almost_equal(b, expected,
decimal=decimal_accuracy, verbose=verbosity)
# a doesn't change (non-inline)
np.testing.assert_almost_equal(a.to_numpy(), data,
decimal=decimal_accuracy, verbose=verbosity)
def test_int_topk():
data = np.array([[[3, 2], [12, 4]], [[6, 44], [43, 2]]])
a = IntTensor(data)
result = a.topk(1)
expected = np.array([[[3], [12]], [[43], [44]]])
np.testing.assert_almost_equal(result.to_numpy(),
expected,
decimal=decimal_accuracy,
verbose=verbosity)
# a doesn't change (non-inline)
np.testing.assert_almost_equal(a.to_numpy(),
data,
decimal=decimal_accuracy,
verbose=verbosity)
