Nikita Shulga c8889f4e10 cuda._is_in_bad_fork->_C._cuda_isInBadFork (#87317) ... Former is always available, while later is only available if PyTorch compiled with CUDA And if it does, then ``` $ python -c "import torch;print(torch._C._cuda_isInBadFork == torch.cuda._is_in_bad_fork)" True ``` Fixes https://github.com/pytorch/torchdynamo/issues/1709 ( at least the symptom) cc @jansel @lezcano @fdrocha Pull Request resolved: https://github.com/pytorch/pytorch/pull/87317 Approved by: https://github.com/voznesenskym, https://github.com/albanD, https://github.com/soumith, https://github.com/jansel		2022-10-19 22:15:28 +00:00
..
__init__.py
base.py
builder.py	[Dynamo] Support optimizing over any Tensor with requires_grad = True (#87141)	2022-10-19 22:13:07 +00:00
builtin.py	Revert "Dynamo FX graph stack traceback fix (#87136)"	2022-10-19 18:57:24 +00:00
constant.py
dicts.py	[Dynamo] Support optimizing over any Tensor with requires_grad = True (#87141)	2022-10-19 22:13:07 +00:00
functions.py	[dynamo] Introduce get_real_value API to TensorVariable (#87091)	2022-10-17 20:14:43 +00:00
lists.py
misc.py	Revert "Dynamo FX graph stack traceback fix (#87136)"	2022-10-19 18:57:24 +00:00
nn_module.py	Revert "Dynamo FX graph stack traceback fix (#87136)"	2022-10-19 18:57:24 +00:00
tensor.py	cuda._is_in_bad_fork->_C._cuda_isInBadFork (#87317)	2022-10-19 22:15:28 +00:00
torch.py	Revert "Dynamo FX graph stack traceback fix (#87136)"	2022-10-19 18:57:24 +00:00
user_defined.py