AttributeError: module 'torch.cuda' has no attribute '_UntypedStorage'

[domattioli]

I haven’t found this issue anywhere else yet – I’m running pytorch3D (0.3.0), which requires pytorch (1.12.1). No issues running the same script for a different dataset. With the more extensive dataset, I receive the AttributeError in the subject header and RuntimeError: Pin memory threat exited unexpectedly after 8 iterations. The latter error is associated with the pytorch dataloader, but all suggested solutions say to update to pytorch >= 1.7, which I have. I’m wondering if my cuda setup is problematic?

Edit: running the same script with the less extensive dataset also produces the AttributeError in the subject. However, the error is not fatal.

Exception in thread Thread-2:
Traceback (most recent call last):
  File "/home/domattioli/miniconda3/envs/anerf/lib/python3.8/threading.py", line 932, in _bootstrap_inner
    self.run()
  File "/home/domattioli/miniconda3/envs/anerf/lib/python3.8/threading.py", line 870, in run
    self._target(*self._args, **self._kwargs)
  File "/home/domattioli/miniconda3/envs/anerf/lib/python3.8/site-packages/torch/utils/data/_utils/pin_memory.py", line 28, in _pin_memory_loop
    r = in_queue.get(timeout=MP_STATUS_CHECK_INTERVAL)
  File "/home/domattioli/miniconda3/envs/anerf/lib/python3.8/multiprocessing/queues.py", line 116, in get
    return _ForkingPickler.loads(res)
  File "/home/domattioli/miniconda3/envs/anerf/lib/python3.8/site-packages/torch/multiprocessing/reductions.py", line 303, in rebuild_storage_fd
    shared_cache[fd_id(fd)] = StorageWeakRef(storage)
  File "/home/domattioli/miniconda3/envs/anerf/lib/python3.8/site-packages/torch/multiprocessing/reductions.py", line 65, in __setitem__
    self.free_dead_references()
  File "/home/domattioli/miniconda3/envs/anerf/lib/python3.8/site-packages/torch/multiprocessing/reductions.py", line 70, in free_dead_references
    if storage_ref.expired():
  File "/home/domattioli/miniconda3/envs/anerf/lib/python3.8/site-packages/torch/multiprocessing/reductions.py", line 35, in expired
    return torch.Storage._expired(self.cdata)  # type: ignore[attr-defined]
  File "/home/domattioli/miniconda3/envs/anerf/lib/python3.8/site-packages/torch/storage.py", line 757, in _expired
    return eval(cls.__module__)._UntypedStorage._expired(*args, **kwargs)
AttributeError: module 'torch.cuda' has no attribute '_UntypedStorage'

[Robert_Crovella]

NVIDIA doesn’t develop, maintain, or support pytorch. You might want to ask pytorch questions on a pytorch forum.

If you are wondering whether you have a proper CUDA setup, that question belongs on the CUDA setup forum, and the verification steps are provided in the CUDA linux install guide.

[domattioli]

Makes sense, thanks

[csarofeen]

NVIDIA most definitely does have a PyTorch team, but the PyTorch forums are still a great place to ask questions. As the PyTorch forum member with the most posts manages the PyTorch Core team @ NVIDIA.

[mohitgaikwad2804]

On date 27/ 08/ 2022:

I got this error when working with Pytorch 1.12, but the error eliminated with Pytorch 1.10