frozenleaves/pytorch
1
0
Fork 0
mirror of https://github.com/pytorch/pytorch.git synced 2025-10-20 21:14:14 +08:00
Code Issues Packages Projects Releases Wiki Activity

Make sure requires_grad is propagated for all backend

Browse Source 
The if statement is not strictly necessary but that avoid having to call this function if we don't need it.
Pull Request resolved: https://github.com/pytorch/pytorch/pull/76256
Approved by: https://github.com/ezyang, https://github.com/soulitzer
This commit is contained in:
Alban Desmaison
2022-04-25 19:31:24 +00:00
committed by PyTorch MergeBot
parent 920876d693
commit e4d5801e36
2 changed files with 15 additions and 3 deletions
Download Patch File Download Diff File Expand all files Collapse all files

14
test/test_serialization.py
View File

@ -20,8 +20,8 @@ from torch._utils_internal import get_file_path_2
from torch._utils import _rebuild_tensor from torch._utils import _rebuild_tensor
from torch.serialization import check_module_version_greater_or_equal from torch.serialization import check_module_version_greater_or_equal
from torch.testing._internal.common_utils import TestCase, IS_WINDOWS, \ from torch.testing._internal.common_utils import TestCase, IS_WINDOWS, TEST_DILL, \
TEST_DILL, run_tests, download_file, BytesIOContext, TemporaryFileName run_tests, download_file, BytesIOContext, TemporaryFileName, parametrize, instantiate_parametrized_tests
from torch.testing._internal.common_device_type import instantiate_device_type_tests from torch.testing._internal.common_device_type import instantiate_device_type_tests
from torch.testing._internal.common_dtype import all_types_and_complex_and from torch.testing._internal.common_dtype import all_types_and_complex_and
@ -948,8 +948,18 @@ class TestSubclassSerialization(TestCase):
self.assertEqual(new_tensor.elem, my_tensor.elem) self.assertEqual(new_tensor.elem, my_tensor.elem)
self.assertEqual(new_tensor.foo, foo_val) self.assertEqual(new_tensor.foo, foo_val)
@parametrize('requires_grad', (True, False))
def test_cloned_deepcopy(self, requires_grad):
my_tensor = torch.rand(2, requires_grad=requires_grad, device='meta')
new_tensor = deepcopy(my_tensor)
self.assertEqual(new_tensor.requires_grad, my_tensor.requires_grad)
instantiate_device_type_tests(TestBothSerialization, globals()) instantiate_device_type_tests(TestBothSerialization, globals())
instantiate_parametrized_tests(TestSubclassSerialization)
if __name__ == '__main__': if __name__ == '__main__':
run_tests() run_tests()

4
torch/_tensor.py
View File

@ -95,6 +95,7 @@ class Tensor(torch._C._TensorBase):
# does accurate alias tracking; however, the code below # does accurate alias tracking; however, the code below
# doesn't work because of # doesn't work because of
# https://github.com/pytorch/pytorch/issues/47442 # https://github.com/pytorch/pytorch/issues/47442
# Update the test in test_serialization if you remove 'meta' from here
if self.is_sparse or self.device.type in ['lazy', 'xla', 'mlc', 'ort', 'meta', 'hpu'] or \ if self.is_sparse or self.device.type in ['lazy', 'xla', 'mlc', 'ort', 'meta', 'hpu'] or \
(type(self) is not Tensor and self.data_ptr() == 0): (type(self) is not Tensor and self.data_ptr() == 0):
new_tensor = self.clone() new_tensor = self.clone()
@ -150,7 +151,8 @@ class Tensor(torch._C._TensorBase):
new_tensor = new_tensor.conj_physical() new_tensor = new_tensor.conj_physical()
if self.is_neg(): if self.is_neg():
new_tensor = new_tensor.neg() new_tensor = new_tensor.neg()
new_tensor.requires_grad = self.requires_grad if self.requires_grad:
new_tensor.requires_grad_()
if self.grad is not None: if self.grad is not None:
new_tensor.grad = self.grad.__deepcopy__(memo) new_tensor.grad = self.grad.__deepcopy__(memo)