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

Handle f([]) vs. f() in fake tensor caching (#162284)

Browse Source 
Fixes https://github.com/pytorch/pytorch/issues/162279
Pull Request resolved: https://github.com/pytorch/pytorch/pull/162284
Approved by: https://github.com/manuelcandales, https://github.com/aorenste
This commit is contained in:
angelayi
2025-09-08 18:28:01 +00:00
committed by PyTorch MergeBot
parent 314d47a210
commit fbcabb4fbd
2 changed files with 14 additions and 0 deletions
Download Patch File Download Diff File Expand all files Collapse all files

4
torch/_subclasses/fake_tensor.py
View File

@ -1677,6 +1677,10 @@ class FakeTensorMode(TorchDispatchMode):
)
from torch._higher_order_ops.utils import FunctionalizeCtxWrapper
if isinstance(args, (list, tuple, dict)):
result.append(type(args))
result.append(f"length_{len(args)}")
if isinstance(args, dict):
self._prep_args_for_hash(result, args.keys(), state, id_hashed_objects)
self._prep_args_for_hash(result, args.values(), state, id_hashed_objects)