You signed in with another tab or window. Reload to refresh your session.You signed out in another tab or window. Reload to refresh your session.You switched accounts on another tab or window. Reload to refresh your session.Dismiss alert
We have a few torch.Tensor.storage() calls across the codebase, most notably in the communication module. PyTorch raises the following warning (see below).
Update code wherever needed (call Tensor.untyped_storage() instead of Tensor.storage(), check for backward compatibility).
Code snippet triggering the error
No response
Error message or erroneous outcome
UserWarning: TypedStorage is deprecated. It will be removed in the future and UntypedStorage will be the only storage class. This should only matter to you if you are using storages directly. To access UntypedStorage directly, use tensor.untyped_storage() instead of tensor.storage()
Version
1.3.x
Python version
None
PyTorch version
None
MPI version
No response
The text was updated successfully, but these errors were encountered:
Note: since this will only become an actual bug when deprecation actually happens in a future PyTorch release, we handeled our adaptations as "feature" instead of "bug", i.e. #1230 has been merged into main (development version) instead of a bug fix version of the current release.
What happened?
We have a few
torch.Tensor.storage()
calls across the codebase, most notably in thecommunication
module. PyTorch raises the following warning (see below).Update code wherever needed (call
Tensor.untyped_storage()
instead ofTensor.storage()
, check for backward compatibility).Code snippet triggering the error
No response
Error message or erroneous outcome
Version
1.3.x
Python version
None
PyTorch version
None
MPI version
No response
The text was updated successfully, but these errors were encountered: