Conversion keeps failing when trying to sub-sample a tensor using a list

[aneeshbhattacharya]

Description of the bug:

Bad StatusOr access: FAILED_PRECONDITION: DNN library initialization failed. Look at the errors above for more details.

While executing %nonzero : [num_users=2] = call_function[target=torch.ops.aten.nonzero.default](args = (%write_mlir_debuginfo_10317,), kwargs = {})
Original traceback:
File "/home/rvt/Desktop/XFeat/xfeat-port-dlc/src/accelerated_features/modules/xfeat_full_with_matcher.py", line 90, in forward
idx0, idx1 = self.match(desc1[0], desc2[0]) # Since we are only matching 1 image at a time, simply match
File "/home/rvt/Desktop/XFeat/xfeat-port-dlc/src/accelerated_features/modules/xfeat_full_with_matcher.py", line 215, in match
valid_indices = torch.nonzero(combined_mask)[:, 0] # [K], where K is number of valid matches

/home/rvt/.conda/envs/latest_ai_edge/lib/python3.11/site-packages/torch/utils/pytree.py:803: UserWarning: To copy construct from a tensor, it is recommended to use sourceTensor.clone().detach() or sourceTensor.clone().detach().requires_grad(True), rather than torch.tensor(sourceTensor).
leaves = list(leaves)
E1026 01:21:50.875000 87860 site-packages/torch/fx/experimental/recording.py:298] failed while running _rename_unbacked_to(*(u0, u6), **{})
Traceback (most recent call last):

NOTE: Same issue also occurs with an operation torch.mask_select() and tensor[list]

Actual vs expected behavior:

No response

Any other information you'd like to share?

No response

[pkgoogle]

Hi @aneeshbhattacharya, can you please share more context/code which produces the error (preferably the more code the better). Also please share code snippets with triple backticks for better formatting:

Example how it should look when editing:
```py
# Example python code
```

produced comment:

# Example python code

Thanks for your help.