convolution_overrideable not implemented when executing VAEEncode

### Expected Behavior

Use Inpaint Preprocessor with Load Checkpoint (Realistic Vision V5.1 Inpainting) and ControlNet Model (control_v11p_sd15_inpaint) to inpaint a masked area of a loaded image based on a text prompt.

### Actual Behavior

Whenever the process reaches VAE Encode node, the error message emerges.

### Steps to Reproduce

1. Launch ComfyUI on MacBook Air M3.
2. Select Realistic Vision model v5.1 model in the Load Checkpoint node.
3. Select ControlNet inpaint model in the Load ControlNet Model node.
4. Load an image and create a mask in MaskEditor.
5. Queue prompt.

### Debug Logs

```powershell
"!!! Exception during processing!!! convolution_overrideable not implemented. You are likely triggering this with tensor backend other than CPU/CUDA/MKLDNN, if this is intended, please use TORCH_LIBRARY_IMPL to override this function
Traceback (most recent call last):
File "/Users/yegorartyukh/ComfyUI/execution.py", line 152, in recursive_execute
output_data, output_ui = get_output_data(obj, input_data_all)
^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^
File "/Users/yegorartyukh/ComfyUI/execution.py", line 82, in get_output_data
return_values = map_node_over_list(obj, input_data_all, obj.FUNCTION, allow_interrupt=True)
^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^
File "/Users/yegorartyukh/ComfyUI/execution.py", line 75, in map_node_over_list
results.append(getattr(obj, func)(**slice_dict(input_data_all, i)))
^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^
File "/Users/yegorartyukh/ComfyUI/nodes.py", line 296, in encode
t = vae.encode(pixels[:,:,:,:3])
^^^^^^^^^^^^^^^^^^^^^^^^^^^^
File "/Users/yegorartyukh/ComfyUI/comfy/sd.py", line 346, in encode
samples[x:x+batch_number] = self.first_stage_model.encode(pixels_in).to(self.output_device).float()
^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^
File "/Users/yegorartyukh/ComfyUI/comfy/ldm/models/autoencoder.py", line 179, in encode
z = self.encoder(x)
^^^^^^^^^^^^^^^
File "/opt/homebrew/lib/python3.11/site-packages/torch/nn/modules/module.py", line 1736, in _wrapped_call_impl
return self._call_impl(*args, **kwargs)
^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^
File "/opt/homebrew/lib/python3.11/site-packages/torch/nn/modules/module.py", line 1747, in _call_impl
return forward_call(*args, **kwargs)
^^^^^^^^^^^^^^^^^^^^^^^^^^^^^
File "/Users/yegorartyukh/ComfyUI/comfy/ldm/modules/diffusionmodules/model.py", line 520, in forward
h = self.conv_in(x)
^^^^^^^^^^^^^^^
File "/opt/homebrew/lib/python3.11/site-packages/torch/nn/modules/module.py", line 1736, in _wrapped_call_impl
return self._call_impl(*args, **kwargs)
^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^
File "/opt/homebrew/lib/python3.11/site-packages/torch/nn/modules/module.py", line 1747, in _call_impl
return forward_call(*args, **kwargs)
^^^^^^^^^^^^^^^^^^^^^^^^^^^^^
File "/Users/yegorartyukh/ComfyUI/comfy/ops.py", line 93, in forward
return super().forward(*args, **kwargs)
^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^
File "/opt/homebrew/lib/python3.11/site-packages/torch/nn/modules/conv.py", line 549, in forward
return self._conv_forward(input, self.weight, self.bias)
^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^
File "/opt/homebrew/lib/python3.11/site-packages/torch/nn/modules/conv.py", line 544, in _conv_forward
return F.conv2d(
^^^^^^^^^
NotImplementedError: convolution_overrideable not implemented. You are likely triggering this with tensor backend other than CPU/CUDA/MKLDNN, if this is intended, please use TORCH_LIBRARY_IMPL to override this function"
```

### Other

I was trying different solutions:
1. Python versions 3.10, 3.11, 3.12.
2. Different PyTorch versions.
3. Setting a device explicitly to mps and/or CPU in the main script.

MacBook Air M3 2024, 16GB, macOS Sonoma 14.3, Python 3.11, PyTorch 2.5 Preview (Nightly).