ValueError: operands could not be broadcast together with shapes (4,) (3,)

[talhaanwarch]

This is the inference code same as in spleen example

import matplotlib.pyplot as plt
model.eval()
device = torch.device("cuda:0")
model.to(device)
val_data=[]
with torch.no_grad():
    for i, val_data in enumerate(model.val_dataloader()):
        roi_size = (160, 160, 160)
        sw_batch_size = 4
        val_data['pred'] = sliding_window_inference(
            val_data["image"].to(device), roi_size, sw_batch_size, model
        )
        # plot the slice [:, :, 80]
        plt.figure("check", (18, 6))
        plt.subplot(1, 3, 1)
        plt.title(f"image {i}")
        plt.imshow(val_data["image"][0, 0, :, :, 80], cmap="gray")
        plt.subplot(1, 3, 2)
        plt.title(f"label {i}")
        plt.imshow(val_data["label"][0, 0, :, :, 80])
        plt.subplot(1, 3, 3)
        plt.title(f"output {i}")
        plt.imshow(torch.argmax(
            val_data['pred'], dim=1).detach().cpu()[0, :, :, 80])
        plt.show()
        post_transforms(val_data)

Code for val transform

val_transforms =T.Compose(
    [
        T.LoadImaged(keys=["image", "label"]),
        T.AddChanneld(keys=["image", "label"]),
        T.Spacingd(
            keys=["image", "label"],
            pixdim=(1.5, 1.5, 2.0),
            mode=("bilinear", "nearest"),
        ),
        T.Orientationd(keys=["image", "label"], axcodes="RAS"),
        T.ScaleIntensityRanged(
            keys=["image"], a_min=-57, a_max=164,
            b_min=0.0, b_max=1.0, clip=True,
        ),
        T.Resized(keys=["image", "label"],spatial_size =(224,224,-1)),
        T.DivisiblePadd(keys=["image", "label"],k=(-1,-1,96)),
        #T.CropForegroundd(keys=["image", "label"], source_key="image"),
        T.ToTensord(keys=["image", "label"]),
    ]
)

Here is the code to invert

post_transforms = T.Compose([
        T.Activationsd(keys="pred", sigmoid=True),
        T.AsDiscreted(keys="pred", threshold_values=True),
        T.Invertd(
            keys="pred",  # invert the `pred` data field, also support multiple fields
            transform=val_transforms,
            orig_keys="image",  # get the previously applied pre_transforms information on the `img` data field,
                              # then invert `pred` based on this information. we can use same info
                              # for multiple fields, also support different orig_keys for different fields
            #meta_keys="pred_meta_dict",  # key field to save inverted meta data, every item maps to `keys`
            #orig_meta_keys="img_meta_dict",  # get the meta data from `img_meta_dict` field when inverting,
                                             # for example, may need the `affine` to invert `Spacingd` transform,
                                             # multiple fields can use the same meta data to invert
            meta_key_postfix="meta_dict",  # if `meta_keys=None`, use "{keys}_{meta_key_postfix}" as the meta key,
                                           # if `orig_meta_keys=None`, use "{orig_keys}_{meta_key_postfix}",
                                           # otherwise, no need this arg during inverting
            nearest_interp=True,  # change to use "nearest" mode in interpolation when inverting
            #to_tensor=True,  # convert to PyTorch Tensor after inverting
        ),
        T.SaveImaged(keys="pred", meta_keys="pred_meta_dict", output_dir="/content/", output_postfix="seg", resample=False),
    ])

It give following trace

---------------------------------------------------------------------------
ValueError                                Traceback (most recent call last)
/usr/local/lib/python3.7/dist-packages/monai/transforms/transform.py in apply_transform(transform, data, map_items, unpack_items)
     83             return [_apply_transform(transform, item, unpack_items) for item in data]
---> 84         return _apply_transform(transform, data, unpack_items)
     85     except Exception as e:

9 frames
/usr/local/lib/python3.7/dist-packages/monai/transforms/transform.py in _apply_transform(transform, parameters, unpack_parameters)
     51 
---> 52     return transform(parameters)
     53 

/usr/local/lib/python3.7/dist-packages/monai/transforms/croppad/dictionary.py in inverse(self, data)
    295             current_size = np.array(d[key].shape[1:])
--> 296             roi_start = np.floor((current_size - orig_size) / 2)
    297             roi_end = orig_size + roi_start

ValueError: operands could not be broadcast together with shapes (4,) (3,) 

The above exception was the direct cause of the following exception:

RuntimeError                              Traceback (most recent call last)
/usr/local/lib/python3.7/dist-packages/monai/transforms/transform.py in apply_transform(transform, data, map_items, unpack_items)
     83             return [_apply_transform(transform, item, unpack_items) for item in data]
---> 84         return _apply_transform(transform, data, unpack_items)
     85     except Exception as e:

/usr/local/lib/python3.7/dist-packages/monai/transforms/transform.py in _apply_transform(transform, parameters, unpack_parameters)
     51 
---> 52     return transform(parameters)
     53 

/usr/local/lib/python3.7/dist-packages/monai/transforms/post/dictionary.py in __call__(self, data)
    539             with allow_missing_keys_mode(self.transform):  # type: ignore
--> 540                 inverted = self.transform.inverse(input_dict)
    541 

/usr/local/lib/python3.7/dist-packages/monai/transforms/compose.py in inverse(self, data)
    168         for t in reversed(invertible_transforms):
--> 169             data = apply_transform(t.inverse, data, self.map_items, self.unpack_items)
    170         return data

/usr/local/lib/python3.7/dist-packages/monai/transforms/transform.py in apply_transform(transform, data, map_items, unpack_items)
    107                 _log_stats(data=data)
--> 108         raise RuntimeError(f"applying transform {transform}") from e
    109 

RuntimeError: applying transform <bound method DivisiblePadd.inverse of <monai.transforms.croppad.dictionary.DivisiblePadd object at 0x7f8a3f708d10>>

The above exception was the direct cause of the following exception:

RuntimeError                              Traceback (most recent call last)
<ipython-input-62-04d6c4e6e61a> in <module>()
     24             val_data['pred'], dim=1).detach().cpu()[0, :, :, 80])
     25         plt.show()
---> 26         post_transforms(val_data)
     27         #val_shapes.append(val_data["image"].shape)

/usr/local/lib/python3.7/dist-packages/monai/transforms/compose.py in __call__(self, input_)
    157     def __call__(self, input_):
    158         for _transform in self.transforms:
--> 159             input_ = apply_transform(_transform, input_, self.map_items, self.unpack_items)
    160         return input_
    161 

/usr/local/lib/python3.7/dist-packages/monai/transforms/transform.py in apply_transform(transform, data, map_items, unpack_items)
    106             else:
    107                 _log_stats(data=data)
--> 108         raise RuntimeError(f"applying transform {transform}") from e
    109 
    110 

RuntimeError: applying transform <monai.transforms.post.dictionary.Invertd object at 0x7f89960a0650>

Here is gist
https://colab.research.google.com/gist/talhaanwarch/2b309feae0f855f222fafe37eec33a22/3d-segmentation.ipynb

[Nic-Ma]

Hi @talhaanwarch ,

Thanks for your interest here.
As we introduced several breaking changes in MONAI v0.6 to improve the flexibility, please check the wiki:
https://github.com/Project-MONAI/MONAI/wiki/v0.5-to-v0.6-migration-guide
and the latest example code:
https://github.com/Project-MONAI/tutorials/blob/0.6.0/3d_segmentation/torch/unet_inference_dict.py
to adjust your inference program.

Thanks.