How can I add custom attributes like device_local or host_local to a PyTorch tensor without affecting TensorImpl or StorageImpl? I have a use case where I need to convert an external tensor into a PyTorch tensor while preserving such properties.

How to add custom attributes to pytorch tensor

bailuan May 28, 2025, 2:34am 4

Thanks, I have added the additional attr to myTensorImpl based on tensorImpl, it works.

Topic		Replies	Views
Is there a place for storing custom data within PyTorch Tensor hardware-backends	8	1540	March 23, 2023
Possible to use custom backend to create TensorImpl that allows custom datatype? hardware-backends	2	400	April 11, 2025
Tensor subclass now support deepcopy! frontend API	0	846	September 28, 2021
Memory operations on a custom backend hardware-backends	4	1157	July 5, 2022
How to support heterogeneous memories hardware-backends	0	535	June 29, 2023