project-monai / slicermonaiviz

MONAIViz - 3D Slicer Extension

MONAIViz

MONAIViz is an extension for 3D Slicer that helps to run chain of MONAI transforms and visualize every stage over an image/label.

Modules

The extension provides a single module - MONAIViz - which currently supports:

• import pre-processing definitions for available bundles from MONAI model zoo
• add/remove/re-order any MONAI transform to the list of transforms
• apply a sequence of transforms step-by-step over input image/label
• visualize image/label outputs for every transform run
• check the data/dictionary stats for every transform run

---

---

Installing Plugin

Installing 3D Slicer

To use MONAIViz with 3D Slicer, you'll need to download and install 3D Slicer. MONAIViz supports the stable and preview versions of 3D Slicer, version 5.3 or higher. For more information on installing 3D Slicer, check out the 3D Slicer Documentation

Installing MONAIViz Plugin

• Go to View -> Extension Manager -> Developer Tools -> MONAIViz
• Install MONAIViz plugin
• Restart 3D Slicer

Note: To update the plugin to the latest version, you have to uninstall the existing 3D Slicer version and download and install the new preview version of 3D Slicer again

Install Plugin in Developer Mode

• git clone [email protected]:Project-MONAI/SlicerMONAIViz.git
• Open 3D Slicer: Go to Edit -> Application Settings -> Modules -> Additional Module Paths
• Add New Module Path: <FULL_PATH>/SlicerMONAIViz/MONAIViz
• Restart 3D Slicer

Dependencies

MONAIViz depends on the following packages:

• PyTorch Slicer extension
• monai[itk,nibabel] python packages

Note: After opening the MONAIViz module in 3D Slicer for the first time, you will be asked to confirm the installation of the packages.

slicermonaiviz's Issues

apply transforms to a volume node

the plugin currently requires an image path

SlicerMONAIViz/MONAIViz/MONAIViz.py

Lines 501 to 503 in 9cbb50b

	def onRunTransform(self):
	if not self.ui.imagePathLineEdit.currentPath:
	slicer.util.errorDisplay("Image is not selected!")

and as a result, the transform sequence must start with a LoadImageD.

would be great to remove this requirement and the user can apply transforms to a volume node that is already in a scene.

Installation of PyYaml

yaml seems to be needed for most bundles, but neither MONAI nor the bundles install it.

It should be probably added as a dependency of MONAI or the bundles that require it, but as a workaround it could be added to the installation instructions.

support modifying transform arguments in different class types

Hi,
great tool so far! I am trying to apply a Lambdad function in my transform chain.
For example, after LoadImaged I want to apply the transform:

Lambdad(keys=["image"], func=lambda x: x.squeeze())

If I add the Lambdad transform in my stack in the MONAIViz panel, and I add the params keys=["image"] and func=lambda x: x.squeeze(), I receive the following error message:

Traceback (most recent call last):
  File "C:/Projects/SlicerMONAIViz/MONAIViz/MONAIViz.py", line 537, in onRunTransform
    t = eval(exp)
  File "<string>", line 1
    monai.transforms.Lambdad(keys=['image'], func=<function <lambda> at 0x0000017BD75F3280>)
                                                  ^
SyntaxError: invalid syntax

I am not sure what's happening under the hood in this module - would a lambda eval even be possible?
Thanks in advance!

Monai image transformation issue

while making exact transformations for training and val data the output dims are different and I'm interested to know how.

#setting piplines for train and validation

train_transforms = Compose(
    [
        LoadImaged(keys=["image", "label"]),
        EnsureChannelFirstd(keys=["image", "label"]),
        ScaleIntensityRanged(
            keys=["image"],
            a_min=-1024.0,
            a_max=1906.0,
            b_min=0.0,
            b_max=1.0,
            clip=True,
        ),
        Resized(keys=["image", "label"],spatial_size = (240,240,120)),
        CropForegroundd(keys=["image", "label"], source_key="image"),
        DivisiblePadd(keys=["image", "label"], k = 16),
        Orientationd(keys=["image", "label"], axcodes="RAS"),
        Spacingd(keys=["image", "label"], pixdim=(1.5, 1.5, 2.0), mode=("bilinear", "nearest")),
        RandCropByPosNegLabeld(
            keys=["image", "label"],
            label_key="label",
            spatial_size=(96, 96, 96),
            pos=1,
            neg=1,
            num_samples=4,
            image_key="image",
            allow_smaller = True,
            image_threshold=0)])


val_transforms = Compose(
    [
        LoadImaged(keys=["image", "label"]),
        EnsureChannelFirstd(keys=["image", "label"]),
        ScaleIntensityRanged(
            keys=["image"],
            a_min=-1024.0,
            a_max=1906.0,
            b_min=0.0,
            b_max=1.0,
            clip=True),
        Resized(keys=["image", "label"],spatial_size = (240,240,120)),
        CropForegroundd(keys=["image", "label"], source_key="image"),
        DivisiblePadd(keys=["image", "label"], k = 16),
        Orientationd(keys=["image", "label"], axcodes="RAS"),
        Spacingd(keys=["image", "label"], pixdim=(1.5, 1.5, 2.0), mode=("bilinear", "nearest"))])

original dims: (512, 512, 37)
train_loder output : ([313, 313, 131])
val_loader output: ([242, 242, 116])
roi = (96,96,96)

error message:

MONAI hint: if your transforms intentionally create images of different shapes, creating your `DataLoader` with `collate_fn=pad_list_data_collate` might solve this problem (check its documentation).

what are recommendations for handling such dimensions (512, 512, 33)?

monai UNET training error , diffreneces in tensor size

good day,

tensor size for image , label checked from train dataloader just before the training

data = first(train_loader)
data['image'][6].shape , data['label'][6].shape

(torch.Size([1, 90, 90, 40]), torch.Size([1, 90, 90, 40]))

Model:

device = torch.device("cuda")
model = UNet(
    spatial_dims=3,
    in_channels=1,
    out_channels=3,
    channels=(16, 32, 64, 128, 256),
    strides=(2, 2, 2, 2),
    num_res_units=2,
    norm=Norm.BATCH,
).to(device)

Error:

RuntimeError: Sizes of tensors must match except in dimension 1. Expected size 5 but got size 6 for tensor number 1 in the list.

any ideas on what to check? and what are the 5 and 6 dimensions?

Note: labels have background + 2 colors not 1
Thanks

Build and release this plugin as part of 3D Slicer Extension manager

https://slicer.readthedocs.io/en/latest/user_guide/extensions_manager.html
https://slicer.readthedocs.io/en/latest/developer_guide/extensions.html

[Linux] cannot edit transform

I was able to install it on Linux, but when trying to edit transform, I am getting the following error:

Selected Transform for Edit: 0
Traceback (most recent call last):
  File "/home/herzc/Slicer-5.4.0-linux-amd64/slicer.org/Extensions-31938/MONAIViz/lib/Slicer-5.4/qt-scripted-modules/MONAIViz.py", line 398, in onEditTransform
    with open(doc_html, "wb", encoding="utf-8") as fp:
ValueError: binary mode doesn't take an encoding argument

Image geometry is incorrect

Images are loaded with an anterior-posterior rotation or flip when using the module:

Eyes should be pointed up, as in this image:

monai unet training 2nd error on hepatic vessels dataset

transformation

train_transforms = Compose(
    [
        LoadImaged(keys=["image", "label"]),
        EnsureChannelFirstd(keys=["image", "label"]),
        ScaleIntensityRanged(
            keys=["image"],
            a_min=-1024.0,
            a_max=1906.0,
            b_min=0.0,
            b_max=1.0,
            clip=True,
        ),
            Spacingd(keys=["image", "label"], pixdim=(1.5, 1.5, 2.0), mode=("bilinear", "nearest")),
            Orientationd(keys=["image", "label"], axcodes="RAS"),
            Resized(keys=["image", "label"],spatial_size = (240,240,128)),
            DivisiblePadd(keys=["image", "label"], k = 64),
            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 ) ])
val_transforms = Compose(
    [
        LoadImaged(keys=["image", "label"]),
        EnsureChannelFirstd(keys=["image", "label"]),
        ScaleIntensityRanged(
            keys=["image"],
            a_min=-1024.0,
            a_max=1906.0,
            b_min=0.0,
            b_max=1.0,
            clip=True),
       
        #CropForegroundd(keys=["image", "label"], source_key="image"),
        
        Orientationd(keys=["image", "label"], axcodes="RAS"),
        Spacingd(keys=["image", "label"], pixdim=(1.5, 1.5, 2.0), mode=("bilinear", "nearest")),
        Resized(keys=["image", "label"],spatial_size = (240,240,128)),
        DivisiblePadd(keys=["image", "label"],k = 64)])

data shape from loader

data = first(train_loader)
data['image'].shape , data['label'].shape

(torch.Size([4, 1, 96, 96, 96]), torch.Size([4, 1, 96, 96, 96]))

Model:

model = UNet(
    spatial_dims=3,
    in_channels=1,
    out_channels=3,
    channels=(16, 32, 64, 128, 256),
    strides=(2, 2, 2, 2),
    num_res_units=2,
    norm=Norm.BATCH,
).to(device)

----------
epoch 1/100
epoch 1 average loss: 0.7484
----------
epoch 2/100
epoch 2 average loss: 0.7053
/usr/local/src/pytorch/aten/src/ATen/native/cuda/ScatterGatherKernel.cu:365: operator(): block: [1629,0,0], thread: [64,0,0] Assertion `idx_dim >= 0 && idx_dim < index_size && "index out of bounds"` failed.
/usr/local/src/pytorch/aten/src/ATen/native/cuda/ScatterGatherKernel.cu:365: operator(): block: [1629,0,0], thread: [68,0,0] Assertion `idx_dim >= 0 && idx_dim < index_size && "index out of bounds"` failed.
/usr/local/src/pytorch/aten/src/ATen/native/cuda/ScatterGatherKernel.cu:365: operator(): block: [1629,0,0], thread: [72,0,0] Assertion `idx_dim >= 0 && idx_dim < index_size && "index out of bounds"` failed.
/usr/local/src/pytorch/aten/src/ATen/native/cuda/ScatterGatherKernel.cu:365: operator(): block: [1312,0,0], thread: [65,0,0] Assertion `idx_dim >= 0 && idx_dim < index_size && "index out of bounds"` failed.
/usr/local/src/pytorch/aten/src/ATen/native/cuda/ScatterGatherKernel.cu:365: operator(): block: [1630,0,0], thread: [64,0,0] Assertion `idx_dim >= 0 && idx_dim < index_size && "index out of bounds"` failed.
/usr/local/src/pytorch/aten/src/ATen/native/cuda/ScatterGatherKernel.cu:365: operator(): block: [1630,0,0], thread: [68,0,0] Assertion `idx_dim >= 0 && idx_dim < index_size && "index out of bounds"` failed.
/usr/local/src/pytorch/aten/src/ATen/native/cuda/ScatterGatherKernel.cu:365: operator(): block: [1630,0,0], thread: [72,0,0] Assertion `idx_dim >= 0 && idx_dim < index_size && "index out of bounds"` failed.
/usr/local/src/pytorch/aten/src/ATen/native/cuda/ScatterGatherKernel.cu:365: operator(): block: [1630,0,0], thread: [76,0,0] Assertion `idx_dim >= 0 && idx_dim < index_size && "index out of bounds"` failed.
/usr/local/src/pytorch/aten/src/ATen/native/cuda/ScatterGatherKernel.cu:365: operator(): block: [1762,0,0], thread: [65,0,0] Assertion `idx_dim >= 0 && idx_dim < index_size && "index out of bounds"` failed.
/usr/local/src/pytorch/aten/src/ATen/native/cuda/ScatterGatherKernel.cu:365: operator(): block: [1762,0,0], thread: [69,0,0] Assertion `idx_dim >= 0 && idx_dim < index_size && "index out of bounds"` failed.
/usr/local/src/pytorch/aten/src/ATen/native/cuda/ScatterGatherKernel.cu:365: operator(): block: [1762,0,0], thread: [71,0,0] Assertion `idx_dim >= 0 && idx_dim < index_size && "index out of bounds"` failed.
/usr/local/src/pytorch/aten/src/ATen/native/cuda/ScatterGatherKernel.cu:365: operator(): block: [1762,0,0], thread: [73,0,0] Assertion `idx_dim >= 0 && idx_dim < index_size && "index out of bounds"` failed.
/usr/local/src/pytorch/aten/src/ATen/native/cuda/ScatterGatherKernel.cu:365: operator(): block: [1762,0,0], thread: [77,0,0] Assertion `idx_dim >= 0 && idx_dim < index_size && "index out of bounds"` failed.
/usr/local/src/pytorch/aten/src/ATen/native/cuda/ScatterGatherKernel.cu:365: operator(): block: [1762,0,0], thread: [81,0,0] Assertion `idx_dim >= 0 && idx_dim < index_size && "index out of bounds"` failed.
/usr/local/src/pytorch/aten/src/ATen/native/cuda/ScatterGatherKernel.cu:365: operator(): block: [1628,0,0], thread: [64,0,0] Assertion `idx_dim >= 0 && idx_dim < index_size && "index out of bounds"` failed.
/usr/local/src/pytorch/aten/src/ATen/native/cuda/ScatterGatherKernel.cu:365: operator(): block: [1628,0,0], thread: [67,0,0] Assertion `idx_dim >= 0 && idx_dim < index_size && "index out of bounds"` failed.
/usr/local/src/pytorch/aten/src/ATen/native/cuda/ScatterGatherKernel.cu:365: operator(): block: [1628,0,0], thread: [68,0,0] Assertion `idx_dim >= 0 && idx_dim < index_size && "index out of bounds"` failed.
/usr/local/src/pytorch/aten/src/ATen/native/cuda/ScatterGatherKernel.cu:365: operator(): block: [1628,0,0], thread: [71,0,0] Assertion `idx_dim >= 0 && idx_dim < index_size && "index out of bounds"` failed.
/usr/local/src/pytorch/aten/src/ATen/native/cuda/ScatterGatherKernel.cu:365: operator(): block: [1628,0,0], thread: [72,0,0] Assertion `idx_dim >= 0 && idx_dim < index_size && "index out of bounds"` failed.
/usr/local/src/pytorch/aten/src/ATen/native/cuda/ScatterGatherKernel.cu:365: operator(): block: [1761,0,0], thread: [29,0,0] Assertion `idx_dim >= 0 && idx_dim < index_size && "index out of bounds"` failed.
/usr/local/src/pytorch/aten/src/ATen/native/cuda/ScatterGatherKernel.cu:365: operator(): block: [1628,0,0], thread: [56,0,0] Assertion `idx_dim >= 0 && idx_dim < index_size && "index out of bounds"` failed.
/usr/local/src/pytorch/aten/src/ATen/native/cuda/ScatterGatherKernel.cu:365: operator(): block: [1628,0,0], thread: [59,0,0] Assertion `idx_dim >= 0 && idx_dim < index_size && "index out of bounds"` failed.
/usr/local/src/pytorch/aten/src/ATen/native/cuda/ScatterGatherKernel.cu:365: operator(): block: [1628,0,0], thread: [60,0,0] Assertion `idx_dim >= 0 && idx_dim < index_size && "index out of bounds"` failed.
/usr/local/src/pytorch/aten/src/ATen/native/cuda/ScatterGatherKernel.cu:365: operator(): block: [1312,0,0], thread: [50,0,0] Assertion `idx_dim >= 0 && idx_dim < index_size && "index out of bounds"` failed.
/usr/local/src/pytorch/aten/src/ATen/native/cuda/ScatterGatherKernel.cu:365: operator(): block: [1312,0,0], thread: [54,0,0] Assertion `idx_dim >= 0 && idx_dim < index_size && "index out of bounds"` failed.
/usr/local/src/pytorch/aten/src/ATen/native/cuda/ScatterGatherKernel.cu:365: operator(): block: [1312,0,0], thread: [58,0,0] Assertion `idx_dim >= 0 && idx_dim < index_size && "index out of bounds"` failed.
/usr/local/src/pytorch/aten/src/ATen/native/cuda/ScatterGatherKernel.cu:365: operator(): block: [1312,0,0], thread: [62,0,0] Assertion `idx_dim >= 0 && idx_dim < index_size && "index out of bounds"` failed.
/usr/local/src/pytorch/aten/src/ATen/native/cuda/ScatterGatherKernel.cu:365: operator(): block: [1630,0,0], thread: [52,0,0] Assertion `idx_dim >= 0 && idx_dim < index_size && "index out of bounds"` failed.
/usr/local/src/pytorch/aten/src/ATen/native/cuda/ScatterGatherKernel.cu:365: operator(): block: [1630,0,0], thread: [56,0,0] Assertion `idx_dim >= 0 && idx_dim < index_size && "index out of bounds"` failed.
/usr/local/src/pytorch/aten/src/ATen/native/cuda/ScatterGatherKernel.cu:365: operator(): block: [1630,0,0], thread: [60,0,0] Assertion `idx_dim >= 0 && idx_dim < index_size && "index out of bounds"` failed.
/usr/local/src/pytorch/aten/src/ATen/native/cuda/ScatterGatherKernel.cu:365: operator(): block: [1761,0,0], thread: [65,0,0] Assertion `idx_dim >= 0 && idx_dim < index_size && "index out of bounds"` failed.
/usr/local/src/pytorch/aten/src/ATen/native/cuda/ScatterGatherKernel.cu:365: operator(): block: [1761,0,0], thread: [69,0,0] Assertion `idx_dim >= 0 && idx_dim < index_size && "index out of bounds"` failed.
/usr/local/src/pytorch/aten/src/ATen/native/cuda/ScatterGatherKernel.cu:365: operator(): block: [1761,0,0], thread: [71,0,0] Assertion `idx_dim >= 0 && idx_dim < index_size && "index out of bounds"` failed.
/usr/local/src/pytorch/aten/src/ATen/native/cuda/ScatterGatherKernel.cu:365: operator(): block: [1761,0,0], thread: [73,0,0] Assertion `idx_dim >= 0 && idx_dim < index_size && "index out of bounds"` failed.
/usr/local/src/pytorch/aten/src/ATen/native/cuda/ScatterGatherKernel.cu:365: operator(): block: [1761,0,0], thread: [77,0,0] Assertion `idx_dim >= 0 && idx_dim < index_size && "index out of bounds"` failed.
/usr/local/src/pytorch/aten/src/ATen/native/cuda/ScatterGatherKernel.cu:365: operator(): block: [1895,0,0], thread: [64,0,0] Assertion `idx_dim >= 0 && idx_dim < index_size && "index out of bounds"` failed.
/usr/local/src/pytorch/aten/src/ATen/native/cuda/ScatterGatherKernel.cu:365: operator(): block: [1894,0,0], thread: [64,0,0] Assertion `idx_dim >= 0 && idx_dim < index_size && "index out of bounds"` failed.
/usr/local/src/pytorch/aten/src/ATen/native/cuda/ScatterGatherKernel.cu:365: operator(): block: [1894,0,0], thread: [68,0,0] Assertion `idx_dim >= 0 && idx_dim < index_size && "index out of bounds"` failed.
/usr/local/src/pytorch/aten/src/ATen/native/cuda/ScatterGatherKernel.cu:365: operator(): block: [1894,0,0], thread: [72,0,0] Assertion `idx_dim >= 0 && idx_dim < index_size && "index out of bounds"` failed.
/usr/local/src/pytorch/aten/src/ATen/native/cuda/ScatterGatherKernel.cu:365: operator(): block: [2018,0,0], thread: [65,0,0] Assertion `idx_dim >= 0 && idx_dim < index_size && "index out of bounds"` failed.
/usr/local/src/pytorch/aten/src/ATen/native/cuda/ScatterGatherKernel.cu:365: operator(): block: [2018,0,0], thread: [67,0,0] Assertion `idx_dim >= 0 && idx_dim < index_size && "index out of bounds"` failed.
/usr/local/src/pytorch/aten/src/ATen/native/cuda/ScatterGatherKernel.cu:365: operator(): block: [2018,0,0], thread: [69,0,0] Assertion `idx_dim >= 0 && idx_dim < index_size && "index out of bounds"` failed.
/usr/local/src/pytorch/aten/src/ATen/native/cuda/ScatterGatherKernel.cu:365: operator(): block: [2018,0,0], thread: [71,0,0] Assertion `idx_dim >= 0 && idx_dim < index_size && "index out of bounds"` failed.
/usr/local/src/pytorch/aten/src/ATen/native/cuda/ScatterGatherKernel.cu:365: operator(): block: [2018,0,0], thread: [73,0,0] Assertion `idx_dim >= 0 && idx_dim < index_size && "index out of bounds"` failed.
/usr/local/src/pytorch/aten/src/ATen/native/cuda/ScatterGatherKernel.cu:365: operator(): block: [2018,0,0], thread: [75,0,0] Assertion `idx_dim >= 0 && idx_dim < index_size && "index out of bounds"` failed.
/usr/local/src/pytorch/aten/src/ATen/native/cuda/ScatterGatherKernel.cu:365: operator(): block: [2018,0,0], thread: [77,0,0] Assertion `idx_dim >= 0 && idx_dim < index_size && "index out of bounds"` failed.
/usr/local/src/pytorch/aten/src/ATen/native/cuda/ScatterGatherKernel.cu:365: operator(): block: [2018,0,0], thread: [81,0,0] Assertion `idx_dim >= 0 && idx_dim < index_size && "index out of bounds"` failed.
/usr/local/src/pytorch/aten/src/ATen/native/cuda/ScatterGatherKernel.cu:365: operator(): block: [2018,0,0], thread: [85,0,0] Assertion `idx_dim >= 0 && idx_dim < index_size && "index out of bounds"` failed.
/usr/local/src/pytorch/aten/src/ATen/native/cuda/ScatterGatherKernel.cu:365: operator(): block: [1894,0,0], thread: [60,0,0] Assertion `idx_dim >= 0 && idx_dim < index_size && "index out of bounds"` failed.
/usr/local/src/pytorch/aten/src/ATen/native/cuda/ScatterGatherKernel.cu:365: operator(): block: [2021,0,0], thread: [49,0,0] Assertion `idx_dim >= 0 && idx_dim < index_size && "index out of bounds"` failed.
/usr/local/src/pytorch/aten/src/ATen/native/cuda/ScatterGatherKernel.cu:365: operator(): block: [2021,0,0], thread: [53,0,0] Assertion `idx_dim >= 0 && idx_dim < index_size && "index out of bounds"` failed.
/usr/local/src/pytorch/aten/src/ATen/native/cuda/ScatterGatherKernel.cu:365: operator(): block: [2021,0,0], thread: [57,0,0] Assertion `idx_dim >= 0 && idx_dim < index_size && "index out of bounds"` failed.
/usr/local/src/pytorch/aten/src/ATen/native/cuda/ScatterGatherKernel.cu:365: operator(): block: [2021,0,0], thread: [61,0,0] Assertion `idx_dim >= 0 && idx_dim < index_size && "index out of bounds"` failed.
/usr/local/src/pytorch/aten/src/ATen/native/cuda/ScatterGatherKernel.cu:365: operator(): block: [2017,0,0], thread: [45,0,0] Assertion `idx_dim >= 0 && idx_dim < index_size && "index out of bounds"` failed.
/usr/local/src/pytorch/aten/src/ATen/native/cuda/ScatterGatherKernel.cu:365: operator(): block: [2017,0,0], thread: [49,0,0] Assertion `idx_dim >= 0 && idx_dim < index_size && "index out of bounds"` failed.
/usr/local/src/pytorch/aten/src/ATen/native/cuda/ScatterGatherKernel.cu:365: operator(): block: [2017,0,0], thread: [53,0,0] Assertion `idx_dim >= 0 && idx_dim < index_size && "index out of bounds"` failed.
/usr/local/src/pytorch/aten/src/ATen/native/cuda/ScatterGatherKernel.cu:365: operator(): block: [2017,0,0], thread: [55,0,0] Assertion `idx_dim >= 0 && idx_dim < index_size && "index out of bounds"` failed.
/usr/local/src/pytorch/aten/src/ATen/native/cuda/ScatterGatherKernel.cu:365: operator(): block: [2017,0,0], thread: [57,0,0] Assertion `idx_dim >= 0 && idx_dim < index_size && "index out of bounds"` failed.
/usr/local/src/pytorch/aten/src/ATen/native/cuda/ScatterGatherKernel.cu:365: operator(): block: [2017,0,0], thread: [59,0,0] Assertion `idx_dim >= 0 && idx_dim < index_size && "index out of bounds"` failed.
/usr/local/src/pytorch/aten/src/ATen/native/cuda/ScatterGatherKernel.cu:365: operator(): block: [2017,0,0], thread: [61,0,0] Assertion `idx_dim >= 0 && idx_dim < index_size && "index out of bounds"` failed.
/usr/local/src/pytorch/aten/src/ATen/native/cuda/ScatterGatherKernel.cu:365: operator(): block: [2017,0,0], thread: [63,0,0] Assertion `idx_dim >= 0 && idx_dim < index_size && "index out of bounds"` failed.
---------------------------------------------------------------------------
RuntimeError                              Traceback (most recent call last)
File <timed exec>:48

File /opt/conda/lib/python3.10/site-packages/monai/metrics/metric.py:344, in CumulativeIterationMetric.__call__(self, y_pred, y, **kwargs)
    324 def __call__(
    325     self, y_pred: TensorOrList, y: TensorOrList | None = None, **kwargs: Any
    326 ) -> torch.Tensor | Sequence[torch.Tensor | Sequence[torch.Tensor]]:
    327     """
    328     Execute basic computation for model prediction and ground truth.
    329     It can support  both `list of channel-first Tensor` and `batch-first Tensor`.
   (...)
    342         a `batch-first` tensor (BC[HWD]) or a list of `batch-first` tensors.
    343     """
--> 344     ret = super().__call__(y_pred=y_pred, y=y, **kwargs)
    345     if isinstance(ret, (tuple, list)):
    346         self.extend(*ret)

File /opt/conda/lib/python3.10/site-packages/monai/metrics/metric.py:73, in IterationMetric.__call__(self, y_pred, y, **kwargs)
     71 # handling a list of channel-first data
     72 if isinstance(y_pred, (list, tuple)) or isinstance(y, (list, tuple)):
---> 73     return self._compute_list(y_pred, y, **kwargs)
     74 # handling a single batch-first data
     75 if isinstance(y_pred, torch.Tensor):

File /opt/conda/lib/python3.10/site-packages/monai/metrics/metric.py:97, in IterationMetric._compute_list(self, y_pred, y, **kwargs)
     83 """
     84 Execute the metric computation for `y_pred` and `y` in a list of "channel-first" tensors.
     85 
   (...)
     94 Note: subclass may enhance the operation to have multi-thread support.
     95 """
     96 if y is not None:
---> 97     ret = [
     98         self._compute_tensor(p.detach().unsqueeze(0), y_.detach().unsqueeze(0), **kwargs)
     99         for p, y_ in zip(y_pred, y)
    100     ]
    101 else:
    102     ret = [self._compute_tensor(p_.detach().unsqueeze(0), None, **kwargs) for p_ in y_pred]

File /opt/conda/lib/python3.10/site-packages/monai/metrics/metric.py:98, in <listcomp>(.0)
     83 """
     84 Execute the metric computation for `y_pred` and `y` in a list of "channel-first" tensors.
     85 
   (...)
     94 Note: subclass may enhance the operation to have multi-thread support.
     95 """
     96 if y is not None:
     97     ret = [
---> 98         self._compute_tensor(p.detach().unsqueeze(0), y_.detach().unsqueeze(0), **kwargs)
     99         for p, y_ in zip(y_pred, y)
    100     ]
    101 else:
    102     ret = [self._compute_tensor(p_.detach().unsqueeze(0), None, **kwargs) for p_ in y_pred]

File /opt/conda/lib/python3.10/site-packages/monai/metrics/meandice.py:95, in DiceMetric._compute_tensor(self, y_pred, y)
     93     raise ValueError(f"y_pred should have at least 3 dimensions (batch, channel, spatial), got {dims}.")
     94 # compute dice (BxC) for each channel for each batch
---> 95 return self.dice_helper(y_pred=y_pred, y=y)

File /opt/conda/lib/python3.10/site-packages/monai/metrics/meandice.py:260, in DiceHelper.__call__(self, y_pred, y)
    258         x_pred = (y_pred[b, 0] == c) if (y_pred.shape[1] == 1) else y_pred[b, c].bool()
    259         x = (y[b, 0] == c) if (y.shape[1] == 1) else y[b, c]
--> 260         c_list.append(self.compute_channel(x_pred, x))
    261     data.append(torch.stack(c_list))
    262 data = torch.stack(data, dim=0).contiguous()  # type: ignore

File /opt/conda/lib/python3.10/site-packages/monai/metrics/meandice.py:219, in DiceHelper.compute_channel(self, y_pred, y)
    217 """"""
    218 y_o = torch.sum(y)
--> 219 if y_o > 0:
    220     return (2.0 * torch.sum(torch.masked_select(y, y_pred))) / (y_o + torch.sum(y_pred))
    221 if self.ignore_empty:

File /opt/conda/lib/python3.10/site-packages/monai/data/meta_tensor.py:282, in MetaTensor.__torch_function__(cls, func, types, args, kwargs)
    280 if kwargs is None:
    281     kwargs = {}
--> 282 ret = super().__torch_function__(func, types, args, kwargs)
    283 # if `out` has been used as argument, metadata is not copied, nothing to do.
    284 # if "out" in kwargs:
    285 #     return ret
    286 if _not_requiring_metadata(ret):

File /opt/conda/lib/python3.10/site-packages/torch/_tensor.py:1295, in Tensor.__torch_function__(cls, func, types, args, kwargs)
   1292     return NotImplemented
   1294 with _C.DisableTorchFunctionSubclass():
-> 1295     ret = func(*args, **kwargs)
   1296     if func in get_default_nowrap_functions():
   1297         return ret

RuntimeError: CUDA error: device-side assert triggered
CUDA kernel errors might be asynchronously reported at some other API call, so the stacktrace below might be incorrect.
For debugging consider passing CUDA_LAUNCH_BLOCKING=1.
Compile with `TORCH_USE_CUDA_DSA` to enable device-side assertions.

Automatic MONAI install

Instead of asking the user to install MONAI manually, we should display a confirmation popup: OK to install MONAI or cancel to install it manually.

Undefined methods called if no monai installation was found

After installing the extension and running it for the first time, monai python package is not installed.

The following two methods do not exist:

SlicerMONAIViz/MONAIViz/MONAIViz.py

Line 842 in 9dd9dc2

with self.showWaitCursor(), self.peakPythonConsole():

Refactor dictionary transform detection

This piece of code:

SlicerMONAIViz/MONAIViz/MONAIViz.py

Lines 440 to 445 in 27c6fa8

	if t[-1] == "d": # this is a dictionary transform
	# now exclude some transforms whose name happens to end with d
	if t not in ["AffineGrid", "Decollated", "RandAffineGrid", "RandDeformGrid"]:
	image_key = slicer.util.settingsValue("SlicerMONAIViz/imageKey", "image")
	label_key = slicer.util.settingsValue("SlicerMONAIViz/labelKey", "label")
	v = f"keys=['{image_key}', '{label_key}']"

could be refactored to use this:

SlicerMONAIViz/MONAIViz/MONAIVizLib/utils.py

Line 46 in 27c6fa8

"dictionary": o.__module__.endswith("dictionary"),

That should make it more reliable and future proof.

Review/test/doc update to use `version 5.4 or higher`

          Since we will be releasing Slicer 5.4 this week, we could probably change this to `version 5.4 or higher`

Originally posted by @jcfr in #23 (comment)

monai import errors

After installing dependencies with MONAIVizLogic.installMONAI() and then importing monai, I am getting the following errors.

>>> import monai
Traceback (most recent call last):
  File "/Applications/Slicer-5.4.0_stable.app/Contents/lib/Python/lib/python3.9/site-packages/monai/utils/module.py", line 210, in load_submodules
    mod = import_module(name)
  File "/Applications/Slicer-5.4.0_stable.app/Contents/lib/Python/lib/python3.9/importlib/__init__.py", line 127, in import_module
    return _bootstrap._gcd_import(name[level:], package, level)
  File "/Applications/Slicer-5.4.0_stable.app/Contents/lib/Python/lib/python3.9/site-packages/monai/apps/__init__.py", line 15, in <module>
    from .mmars import MODEL_DESC, RemoteMMARKeys, download_mmar, get_model_spec, load_from_mmar
  File "/Applications/Slicer-5.4.0_stable.app/Contents/lib/Python/lib/python3.9/site-packages/monai/apps/mmars/__init__.py", line 14, in <module>
    from .mmars import download_mmar, get_model_spec, load_from_mmar
  File "/Applications/Slicer-5.4.0_stable.app/Contents/lib/Python/lib/python3.9/site-packages/monai/apps/mmars/mmars.py", line 29, in <module>
    import monai.networks.nets as monai_nets
  File "/Applications/Slicer-5.4.0_stable.app/Contents/lib/Python/lib/python3.9/site-packages/monai/networks/nets/__init__.py", line 101, in <module>
    from .swin_unetr import PatchMerging, PatchMergingV2, SwinUNETR
  File "/Applications/Slicer-5.4.0_stable.app/Contents/lib/Python/lib/python3.9/site-packages/monai/networks/nets/swin_unetr.py", line 21, in <module>
    import torch.utils.checkpoint as checkpoint
  File "/Applications/Slicer-5.4.0_stable.app/Contents/lib/Python/lib/python3.9/site-packages/torch/utils/checkpoint.py", line 18, in <module>
    from torch.testing._internal.logging_tensor import LoggingTensorMode, capture_logs
  File "/Applications/Slicer-5.4.0_stable.app/Contents/lib/Python/lib/python3.9/site-packages/torch/testing/_internal/logging_tensor.py", line 2, in <module>
    from torch.utils._pytree import tree_map
ImportError: cannot import name 'tree_map' from 'torch.utils._pytree' (/Applications/Slicer-5.4.0_stable.app/Contents/lib/Python/lib/python3.9/site-packages/torch/utils/_pytree.py)

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

Traceback (most recent call last):
  File "<console>", line 1, in <module>
  File "/Applications/Slicer-5.4.0_stable.app/Contents/lib/Python/lib/python3.9/site-packages/monai/__init__.py", line 58, in <module>
    load_submodules(sys.modules[__name__], False, exclude_pattern=excludes)
  File "/Applications/Slicer-5.4.0_stable.app/Contents/lib/Python/lib/python3.9/site-packages/monai/utils/module.py", line 220, in load_submodules
    raise type(e)(f"{e}\n{msg}").with_traceback(e.__traceback__) from e  # raise with modified message
  File "/Applications/Slicer-5.4.0_stable.app/Contents/lib/Python/lib/python3.9/site-packages/monai/utils/module.py", line 210, in load_submodules
    mod = import_module(name)
  File "/Applications/Slicer-5.4.0_stable.app/Contents/lib/Python/lib/python3.9/importlib/__init__.py", line 127, in import_module
    return _bootstrap._gcd_import(name[level:], package, level)
  File "/Applications/Slicer-5.4.0_stable.app/Contents/lib/Python/lib/python3.9/site-packages/monai/apps/__init__.py", line 15, in <module>
    from .mmars import MODEL_DESC, RemoteMMARKeys, download_mmar, get_model_spec, load_from_mmar
  File "/Applications/Slicer-5.4.0_stable.app/Contents/lib/Python/lib/python3.9/site-packages/monai/apps/mmars/__init__.py", line 14, in <module>
    from .mmars import download_mmar, get_model_spec, load_from_mmar
  File "/Applications/Slicer-5.4.0_stable.app/Contents/lib/Python/lib/python3.9/site-packages/monai/apps/mmars/mmars.py", line 29, in <module>
    import monai.networks.nets as monai_nets
  File "/Applications/Slicer-5.4.0_stable.app/Contents/lib/Python/lib/python3.9/site-packages/monai/networks/nets/__init__.py", line 101, in <module>
    from .swin_unetr import PatchMerging, PatchMergingV2, SwinUNETR
  File "/Applications/Slicer-5.4.0_stable.app/Contents/lib/Python/lib/python3.9/site-packages/monai/networks/nets/swin_unetr.py", line 21, in <module>
    import torch.utils.checkpoint as checkpoint
  File "/Applications/Slicer-5.4.0_stable.app/Contents/lib/Python/lib/python3.9/site-packages/torch/utils/checkpoint.py", line 18, in <module>
    from torch.testing._internal.logging_tensor import LoggingTensorMode, capture_logs
  File "/Applications/Slicer-5.4.0_stable.app/Contents/lib/Python/lib/python3.9/site-packages/torch/testing/_internal/logging_tensor.py", line 2, in <module>
    from torch.utils._pytree import tree_map
ImportError: cannot import name 'tree_map' from 'torch.utils._pytree' (/Applications/Slicer-5.4.0_stable.app/Contents/lib/Python/lib/python3.9/site-packages/torch/utils/_pytree.py)

Multiple versions of MONAI may have been installed?
Please see the installation guide: https://docs.monai.io/en/stable/installation.html

Slicer Version

[DEBUG][Qt] 12.10.2023 11:07:20 [] (unknown:0) - Session start time .......: 2023-10-12 11:07:20
[DEBUG][Qt] 12.10.2023 11:07:20 [] (unknown:0) - Slicer version ...........: 5.4.0 (revision 31938 / 311cb26) macosx-amd64 - installed release
[DEBUG][Qt] 12.10.2023 11:07:21 [] (unknown:0) - Operating system .........: macOS / 13.5 / 22G74 / UTF-8 - 64-bit
[DEBUG][Qt] 12.10.2023 11:07:21 [] (unknown:0) - Memory ...................: 65536 MB physical, 0 MB virtual
[DEBUG][Qt] 12.10.2023 11:07:21 [] (unknown:0) - CPU ......................:  Apple M2 Max, 12 cores, 12 logical processors
[DEBUG][Qt] 12.10.2023 11:07:21 [] (unknown:0) - VTK configuration ........: OpenGL2 rendering, Sequential threading
[DEBUG][Qt] 12.10.2023 11:07:21 [] (unknown:0) - Qt configuration .........: version 5.15.8, with SSL, requested OpenGL 3.2 (core profile)
[DEBUG][Qt] 12.10.2023 11:07:21 [] (unknown:0) - Internationalization .....: disabled, language=
[DEBUG][Qt] 12.10.2023 11:07:21 [] (unknown:0) - Developer mode ...........: enabled
[DEBUG][Qt] 12.10.2023 11:07:21 [] (unknown:0) - Application path .........: /Applications/Slicer-5.4.0_stable.app/Contents/MacOS