Does torch-directml support Conv3D?

[AdvancedHe]

As shown in the tittle, does torch-directml support support Conv3D?
When I use Conv3D, it appears:

File "C:\Software\PythonIDE\Anaconda\envs\torch_directml\lib\site-packages\torch\nn\modules\conv.py", line 604, in _conv_forward
    return F.conv3d(
RuntimeError: input must be 4-dimensional

The input shape is: torch.Size([1, 320, 26, 40, 64]).

The code can run normally on the CUDA. But when I use torch-directml and set device = torch_directml.device(torch_directml.default_device())

It appears above error. I suspect this due to torch-directml can't support Conv3D operator.

How should I solve this problem?

Thanks a lot!