def test_sanitize_pytorch_types_int8():
_TORCH_BEFORE_1_1 = version.parse(torch.__version__) < version.parse('1.1.0')
dict_to_sanitize = {'a': np.asarray([-1, 1], dtype=np.int8)}
_sanitize_pytorch_types(dict_to_sanitize)
np.testing.assert_array_equal(dict_to_sanitize['a'], [-1, 1])
if _TORCH_BEFORE_1_1:
assert dict_to_sanitize['a'].dtype == np.int16
else:
assert dict_to_sanitize['a'].dtype == np.int8
def test_torch_tensorable_types(numpy_dtype):
"""Make sure that we 'sanitize' only integer types that can not be made into torch tensors natively"""
value = np.zeros((2, 2), dtype=numpy_dtype)
dict_to_sanitize = {'value': value}
_sanitize_pytorch_types(dict_to_sanitize)
torchable = False
try:
torch.Tensor(value)
torchable = True
except TypeError:
pass
tensor = torch.as_tensor(dict_to_sanitize['value'])
tensor_and_back = tensor.numpy()
if tensor_and_back.dtype != value.dtype:
assert tensor_and_back.dtype.itemsize > value.dtype.itemsize
assert not torchable, '_sanitize_pytorch_types modified value of type {}, but it was possible to create a ' \
'Tensor directly from a value with that type'.format(numpy_dtype)
def test_sanitize_pytorch_types_int8():
dict_to_sanitize = {'a': np.asarray([-1, 1], dtype=np.int8)}
_sanitize_pytorch_types(dict_to_sanitize)
np.testing.assert_array_equal(dict_to_sanitize['a'], [-1, 1])
assert dict_to_sanitize['a'].dtype == np.int16
