Commits on Dec 6, 2022

1. Handle Tensor.__deepcopy__ via clone(), on IPU (pytorch#89129) …

   Currently it falls through to a call to `storage()`, which the IPU doesn't support.
   
   I've made the minimal change here for ease of merging (this'd help us if it was in for 1.13.1), however...
   
   **QUESTION**: Is there any reason why `not torch._C._has_storage(self)` needs to *also* be guarded on `self.device.type == privateuseone`? in other words, could the condition for using `clone` not be this?
   
   ```python
   self.is_sparse
   or self.device.type
   in ["lazy", "xla", "mps", "ort", "meta", "hpu", "ipu"]
   or not torch._C._has_storage(self)
   or (type(self) is not Tensor and self.data_ptr() == 0)
   ```
   
   If the condition fails, the very next thing is a call to `self._typed_storage()` which will fail, so it feels to me like *any* case without storage shouldn't fall through to the `storage()` call.
   
   The original PR for adding the 'no storage and device is `PrivateUse1`' condition ([86557](pytorch#86557)) doesn't discuss whether this could be broadened.
   Pull Request resolved: pytorch#89129
   Approved by: https://github.com/albanD

   

   charlie-wt authored and pytorchmergebot committed Dec 6, 2022
   Copy the full SHA

   1e0e36c View commit details

   Browse the repository at this point in the history