trunk/4687d3465afbe4b55b85296aecdec528d710c3af: [dynamo] Keep the subclass type when resyncing tensor metadata (#193969)
- PyTorch: 176 events in the last 90 days
- PyTorch: 174th Release in the last 90 days
- Previous: earlier the same day · trunk/09d585b8f4e7a450edfda867c0ce24e1034076d0
What happened
Issue Fixes #193933 Summary Dynamo models a tensor subclass without __torch_dispatch__ as a TensorWithTFOverrideVariable holding the Python type in class_type , while the fake tensor behind the proxy stays a plain FakeTensor . When an inplace op bumps the fake version, synchronize_attributes re-derives every specialized property from that fake tensor, and get_specialized_props resolves class_type back to torch.Tenso…
Summary assembled by rule from the sources below