Spleen example crashes if I modify it to use different input dataset

[project-monai]

Spleen example crashes if I modify it to use different input dataset

[YuanTingHsieh]

After your chain of transforms,
Is the data has same dimension?
Seeing from your code your first image is [44, 55, 83], is the second image the same size as this?

Like in the original spleen example, we use

RandCropByPosNegLabeld(
            keys=["image", "label"],
            label_key="label",
            spatial_size=(96, 96, 96),
            pos=1,
            neg=1,
            num_samples=4,
            image_key="image",
            image_threshold=0,
        ),

This ensures that the data passed in will be in batches of [96, 96, 96] cubes.

[dzenanz]

Thank you! That was the problem. Adding DivisiblePadd with k=32 solves it.

[dzenanz]

The images have the same dimensions after the transforms. Updated notebook is here. Adding

        inputs, labels = (
            batch_data["image"].to(device),
            batch_data["label"].to(device),
        )
        print(f"batch_data image: {batch_data['image'].shape}")
        print(f"batch_data label: {batch_data['label'].shape}")

produces the output:

----------
epoch 1/10
batch_data image: torch.Size([20, 1, 44, 55, 83])
batch_data label: torch.Size([20, 1, 44, 55, 83])
---------------------------------------------------------------------------
RuntimeError                              Traceback (most recent call last)
<ipython-input-12-e43444a90bc8> in <module>
     22 
     23         optimizer.zero_grad()
---> 24         outputs = model(inputs)
     25         loss = loss_function(outputs, labels)
     26         loss.backward()

c:\dev\monai\pyenv\lib\site-packages\torch\nn\modules\module.py in __call__(self, *input, **kwargs)
    530             result = self._slow_forward(*input, **kwargs)
    531         else:
--> 532             result = self.forward(*input, **kwargs)
    533         for hook in self._forward_hooks.values():
    534             hook_result = hook(self, input, result)

c:\dev\monai\pyenv\lib\site-packages\monai\networks\nets\unet.py in forward(self, x)
    190 
    191     def forward(self, x: torch.Tensor) -> torch.Tensor:
--> 192         x = self.model(x)
    193         return x
    194 

c:\dev\monai\pyenv\lib\site-packages\torch\nn\modules\module.py in __call__(self, *input, **kwargs)
    530             result = self._slow_forward(*input, **kwargs)
    531         else:
--> 532             result = self.forward(*input, **kwargs)
    533         for hook in self._forward_hooks.values():
    534             hook_result = hook(self, input, result)

c:\dev\monai\pyenv\lib\site-packages\torch\nn\modules\container.py in forward(self, input)
     98     def forward(self, input):
     99         for module in self:
--> 100             input = module(input)
    101         return input
    102 

c:\dev\monai\pyenv\lib\site-packages\torch\nn\modules\module.py in __call__(self, *input, **kwargs)
    530             result = self._slow_forward(*input, **kwargs)
    531         else:
--> 532             result = self.forward(*input, **kwargs)
    533         for hook in self._forward_hooks.values():
    534             hook_result = hook(self, input, result)

c:\dev\monai\pyenv\lib\site-packages\monai\networks\layers\simplelayers.py in forward(self, x)
     37 
     38     def forward(self, x: torch.Tensor) -> torch.Tensor:
---> 39         return torch.cat([x, self.submodule(x)], self.cat_dim)
     40 
     41 

c:\dev\monai\pyenv\lib\site-packages\torch\nn\modules\module.py in __call__(self, *input, **kwargs)
    530             result = self._slow_forward(*input, **kwargs)
    531         else:
--> 532             result = self.forward(*input, **kwargs)
    533         for hook in self._forward_hooks.values():
    534             hook_result = hook(self, input, result)

c:\dev\monai\pyenv\lib\site-packages\torch\nn\modules\container.py in forward(self, input)
     98     def forward(self, input):
     99         for module in self:
--> 100             input = module(input)
    101         return input
    102 

c:\dev\monai\pyenv\lib\site-packages\torch\nn\modules\module.py in __call__(self, *input, **kwargs)
    530             result = self._slow_forward(*input, **kwargs)
    531         else:
--> 532             result = self.forward(*input, **kwargs)
    533         for hook in self._forward_hooks.values():
    534             hook_result = hook(self, input, result)

c:\dev\monai\pyenv\lib\site-packages\monai\networks\layers\simplelayers.py in forward(self, x)
     37 
     38     def forward(self, x: torch.Tensor) -> torch.Tensor:
---> 39         return torch.cat([x, self.submodule(x)], self.cat_dim)
     40 
     41 

c:\dev\monai\pyenv\lib\site-packages\torch\nn\modules\module.py in __call__(self, *input, **kwargs)
    530             result = self._slow_forward(*input, **kwargs)
    531         else:
--> 532             result = self.forward(*input, **kwargs)
    533         for hook in self._forward_hooks.values():
    534             hook_result = hook(self, input, result)

c:\dev\monai\pyenv\lib\site-packages\torch\nn\modules\container.py in forward(self, input)
     98     def forward(self, input):
     99         for module in self:
--> 100             input = module(input)
    101         return input
    102 

c:\dev\monai\pyenv\lib\site-packages\torch\nn\modules\module.py in __call__(self, *input, **kwargs)
    530             result = self._slow_forward(*input, **kwargs)
    531         else:
--> 532             result = self.forward(*input, **kwargs)
    533         for hook in self._forward_hooks.values():
    534             hook_result = hook(self, input, result)

c:\dev\monai\pyenv\lib\site-packages\monai\networks\layers\simplelayers.py in forward(self, x)
     37 
     38     def forward(self, x: torch.Tensor) -> torch.Tensor:
---> 39         return torch.cat([x, self.submodule(x)], self.cat_dim)
     40 
     41 

RuntimeError: invalid argument 0: Sizes of tensors must match except in dimension 1. Got 7 and 8 in dimension 3 at C:\w\1\s\windows\pytorch\aten\src\TH/generic/THTensor.cpp:612

[YuanTingHsieh]

Hi
The problem is the data shape changed after down-sample then up-sample.

You can take a look at the answer here: https://stackoverflow.com/questions/60063797/not-understanding-the-data-flow-in-unet-like-architetures-and-having-problems-wi
and here: http://makeyourownneuralnetwork.blogspot.com/2020/02/calculating-output-size-of-convolutions.html

So you will need to do some calculations and get a valid shape
Then either pad or rescale to that shape should solve this problem

[Ontheroad123]

Thank you! That was the problem. Adding DivisiblePadd with k=32 solves it.

I meet same error, can you show detail in code?