DentalSegmentator / nnUNet problems

Support
1

I’m having troubles with DentalSegmentator.

3D Slicer 5.10.0

The log:

2026/01/09 16:28:19.241 :: Start nnUNet install with requirements : nnunetv2

2026/01/09 16:28:19.280 :: - Installing nnunetv2 --no-deps...

2026/01/09 16:28:20.152 :: - Installing acvl-utils<0.3,>=0.2.3 --no-deps...

2026/01/09 16:28:20.797 :: - Installing batchgenerators --no-deps...

2026/01/09 16:28:21.471 :: - Installing scikit-image --no-deps...

2026/01/09 16:28:22.708 :: - Installing networkx>=3.0 --no-deps...

2026/01/09 16:28:24.741 :: - Installing imageio!=2.35.0,>=2.33 --no-deps...

2026/01/09 16:28:25.673 :: - Installing tifffile>=2022.8.12 --no-deps...

2026/01/09 16:28:26.624 :: - Installing lazy-loader>=0.4 --no-deps...

2026/01/09 16:28:27.257 :: - Installing scikit-learn --no-deps...

2026/01/09 16:28:30.490 :: - Installing joblib>=1.3.0 --no-deps...

2026/01/09 16:28:31.342 :: - Installing threadpoolctl>=3.2.0 --no-deps...

2026/01/09 16:28:32.032 :: - Installing future --no-deps...

2026/01/09 16:28:33.048 :: - Installing pandas --no-deps...

2026/01/09 16:28:39.524 :: - Installing pytz>=2020.1 --no-deps...

2026/01/09 16:28:40.498 :: - Installing tzdata>=2022.7 --no-deps...

2026/01/09 16:28:41.296 :: - Installing unittest2 --no-deps...

2026/01/09 16:28:42.123 :: - Installing argparse --no-deps...

2026/01/09 16:28:42.771 :: - Installing traceback2 --no-deps...

2026/01/09 16:28:43.422 :: - Installing linecache2 --no-deps...

2026/01/09 16:28:44.108 :: - Installing connected-components-3d --no-deps...

2026/01/09 16:28:45.016 :: - Installing blosc2>=3.0.0b4 --no-deps...

2026/01/09 16:28:46.027 :: - Installing ndindex --no-deps...

2026/01/09 16:28:46.870 :: - Installing msgpack --no-deps...

2026/01/09 16:28:47.632 :: - Installing platformdirs --no-deps...

2026/01/09 16:28:48.302 :: - Installing numexpr>=2.14.1 --no-deps...

2026/01/09 16:28:49.147 :: - Installing py-cpuinfo --no-deps...

2026/01/09 16:28:49.795 :: - Installing dynamic-network-architectures<0.5,>=0.4.1 --no-deps...

2026/01/09 16:28:50.504 :: - Installing timm --no-deps...

2026/01/09 16:28:52.209 :: - Installing pyyaml --no-deps...

2026/01/09 16:28:52.982 :: - Installing huggingface_hub --no-deps...

2026/01/09 16:28:54.416 :: - Installing filelock --no-deps...

2026/01/09 16:28:55.094 :: - Installing fsspec>=2023.5.0 --no-deps...

2026/01/09 16:28:56.339 :: - Installing hf-xet<2.0.0,>=1.2.0 --no-deps...

2026/01/09 16:28:57.376 :: - Installing httpx<1,>=0.23.0 --no-deps...

2026/01/09 16:28:58.164 :: - Installing anyio --no-deps...

2026/01/09 16:28:59.029 :: - Installing httpcore==1.* --no-deps...

2026/01/09 16:28:59.757 :: - Installing h11>=0.16 --no-deps...

2026/01/09 16:29:00.402 :: - Installing shellingham --no-deps...

2026/01/09 16:29:01.058 :: - Installing tqdm>=4.42.1 --no-deps...

2026/01/09 16:29:01.811 :: - Installing typer-slim --no-deps...

2026/01/09 16:29:02.559 :: - Installing click>=8.0.0 --no-deps...

2026/01/09 16:29:03.422 :: - Installing safetensors --no-deps...

2026/01/09 16:29:04.323 :: - Installing einops --no-deps...

2026/01/09 16:29:05.062 :: - Installing graphviz --no-deps...

2026/01/09 16:29:05.774 :: - Installing nibabel --no-deps...

2026/01/09 16:29:07.017 :: - Installing matplotlib --no-deps...

2026/01/09 16:29:09.733 :: - Installing contourpy>=1.0.1 --no-deps...

2026/01/09 16:29:10.556 :: - Installing cycler>=0.10 --no-deps...

2026/01/09 16:29:11.222 :: - Installing fonttools>=4.22.0 --no-deps...

2026/01/09 16:29:13.174 :: - Installing kiwisolver>=1.3.1 --no-deps...

2026/01/09 16:29:14.164 :: - Installing seaborn --no-deps...

2026/01/09 16:29:15.420 :: - Installing imagecodecs --no-deps...

2026/01/09 16:29:16.928 :: - Installing yacs --no-deps...

2026/01/09 16:29:17.890 :: - Installing batchgeneratorsv2>=0.3.0 --no-deps...

2026/01/09 16:29:18.850 :: - Installing fft-conv-pytorch --no-deps...

2026/01/09 16:29:19.577 :: PyTorch Python package is required. Installing... (it may take several minutes)

2026/01/09 16:29:48.294 :: nnUNet installation completed successfully.

2026/01/09 16:29:48.311 :: Downloading model weights...

2026/01/09 16:29:57.551 :: Transferring volume to nnUNet in /tmp/Slicer-eSDRuj

2026/01/09 16:30:11.062 :: Starting nnUNet with the following parameters:

2026/01/09 16:30:11.062 ::

2026/01/09 16:30:11.062 :: /home/user/Slicer-5.10.0-linux-amd64/lib/Python/bin/nnUNetv2_predict -i /tmp/Slicer-eSDRuj/input -o /tmp/Slicer-eSDRuj/output -d Dataset111_453CT -tr nnUNetTrainer -p nnUNetPlans -c 3d_fullres -f 0 -npp 1 -nps 1 -step_size 0.5 -device cpu -chk checkpoint_final.pth --disable_tta

2026/01/09 16:30:11.062 ::

2026/01/09 16:30:11.062 :: JSON parameters :

2026/01/09 16:30:11.062 :: {

2026/01/09 16:30:11.062 :: "folds": "0",

2026/01/09 16:30:11.062 :: "device": "cpu",

2026/01/09 16:30:11.062 :: "stepSize": 0.5,

2026/01/09 16:30:11.062 :: "disableTta": true,

2026/01/09 16:30:11.062 :: "nProcessPreprocessing": 1,

2026/01/09 16:30:11.062 :: "nProcessSegmentationExport": 1,

2026/01/09 16:30:11.062 :: "checkPointName": "",

2026/01/09 16:30:11.062 :: "modelPath": {

2026/01/09 16:30:11.062 :: "_path": "/home/user/Slicer-5.10.0-linux-amd64/slicer.org/Extensions-34045/DentalSegmentator/lib/Slicer-5.10/qt-scripted-modules/Resources/ML"

2026/01/09 16:30:11.062 :: }

2026/01/09 16:30:11.062 :: }

2026/01/09 16:30:12.469 :: nnUNet preprocessing...

2026/01/09 16:30:16.804 :: /home/user/Slicer-5.10.0-linux-amd64/lib/Python/lib/python3.12/site-packages/nnunetv2/utilities/plans_handling/plans_handler.py:37: UserWarning: Detected old nnU-Net plans format. Attempting to reconstruct network architecture parameters. If this fails, rerun nnUNetv2_plan_experiment for your dataset. If you use a custom architecture, please downgrade nnU-Net to the version you implemented this or update your implementation + plans.

2026/01/09 16:30:16.804 :: warnings.warn("Detected old nnU-Net plans format. Attempting to reconstruct network architecture "

2026/01/09 16:30:18.304 :: #######################################################################

2026/01/09 16:30:18.304 :: Please cite the following paper when using nnU-Net:

2026/01/09 16:30:18.304 :: Isensee, F., Jaeger, P. F., Kohl, S. A., Petersen, J., & Maier-Hein, K. H. (2021). nnU-Net: a self-configuring method for deep learning-based biomedical image segmentation. Nature methods, 18(2), 203-211.

2026/01/09 16:30:18.304 :: #######################################################################

2026/01/09 16:30:18.304 ::

2026/01/09 16:30:18.304 :: perform_everything_on_device=True is only supported for cuda devices! Setting this to False

2026/01/09 16:30:18.538 :: There are 1 cases in the source folder

2026/01/09 16:30:18.538 :: I am processing 0 out of 1 (max process ID is 0, we start counting with 0!)

2026/01/09 16:30:18.538 :: There are 1 cases that I would like to predict

2026/01/09 16:31:10.971 :: Traceback (most recent call last):

2026/01/09 16:31:10.971 :: File "/home/user/Slicer-5.10.0-linux-amd64/lib/Python/bin/nnUNetv2_predict", line 8, in <module>

2026/01/09 16:31:14.232 :: Loading inference results...

2026/01/09 16:31:17.081 :: Error loading results :

2026/01/09 16:31:17.081 :: Failed to load the segmentation.

2026/01/09 16:31:17.081 :: Something went wrong during the nnUNet processing.

2026/01/09 16:31:17.081 :: Please check the logs for potential errors and contact the library maintainers.
2

I tried to remove as many 3D Slicer related files from my computer and re installed it along with DentalSegmentator. Now my issue is that it says this infinitely:

nnUNet is already installed (2.6.2) and compatible with requested version (nnunetv2).

3

You have another nnUNet installation on your system, probably python is my guess. Are you using nnUNet for anything else? If not, these steps should clear up your problem:

  1. From The Extensions Manager, Uninstall Dental Segmentator, NNUNet, and any other Segmentator Extensions.
  2. Close 3D Slicer
  3. Then open Powershell as admin and run the command: python -m pip uninstall nnunet.
  4. Run: Find-Package -Name nnunet
  5. If nothing shows up for nnunet packages, re-install Dental Segmentator and try again.
4

Has anyone tried this and had it work? I couldn’t fix it and had to revert to the previous version of Slicer, 5.8.

5

It works well for me on Windows:

2026/01/12 11:55:45.742 :: nnUNet is already installed (2.6.2) and compatible with requested version (nnunetv2).
2026/01/12 11:55:45.750 :: Downloading model weights...
2026/01/12 11:56:03.416 :: Transferring volume to nnUNet in C:/Users/andra/AppData/Local/Temp/Slicer-daAwew
2026/01/12 11:56:04.283 :: Starting nnUNet with the following parameters:
2026/01/12 11:56:04.283 :: 
2026/01/12 11:56:04.283 :: C:\Users\andra\AppData\Local\slicer.org\3D Slicer 5.10.0\lib\Python\Scripts\nnUNetv2_predict.exe -i C:/Users/andra/AppData/Local/Temp/Slicer-daAwew/input -o C:/Users/andra/AppData/Local/Temp/Slicer-daAwew/output -d Dataset111_453CT -tr nnUNetTrainer -p nnUNetPlans -c 3d_fullres -f 0 -npp 1 -nps 1 -step_size 0.5 -device cuda -chk checkpoint_final.pth --disable_tta
2026/01/12 11:56:04.283 :: 
2026/01/12 11:56:04.283 :: JSON parameters :
2026/01/12 11:56:04.283 :: {
2026/01/12 11:56:04.283 ::     "folds": "0",
2026/01/12 11:56:04.283 ::     "device": "cuda",
2026/01/12 11:56:04.283 ::     "stepSize": 0.5,
2026/01/12 11:56:04.283 ::     "disableTta": true,
2026/01/12 11:56:04.283 ::     "nProcessPreprocessing": 1,
2026/01/12 11:56:04.283 ::     "nProcessSegmentationExport": 1,
2026/01/12 11:56:04.283 ::     "checkPointName": "",
2026/01/12 11:56:04.283 ::     "modelPath": {
2026/01/12 11:56:04.283 ::         "_path": "C:\\Users\\andra\\AppData\\Local\\slicer.org\\3D Slicer 5.10.0\\slicer.org\\Extensions-34045\\DentalSegmentator\\lib\\Slicer-5.10\\qt-scripted-modules\\Resources\\ML"
2026/01/12 11:56:04.283 ::     }
2026/01/12 11:56:04.283 :: }
2026/01/12 11:56:04.312 :: nnUNet preprocessing...
2026/01/12 11:56:19.029 :: C:\Users\andra\AppData\Local\slicer.org\3D Slicer 5.10.0\lib\Python\Lib\site-packages\nnunetv2\utilities\plans_handling\plans_handler.py:37: UserWarning: Detected old nnU-Net plans format. Attempting to reconstruct network architecture parameters. If this fails, rerun nnUNetv2_plan_experiment for your dataset. If you use a custom architecture, please downgrade nnU-Net to the version you implemented this or update your implementation + plans.
2026/01/12 11:56:19.029 ::   warnings.warn("Detected old nnU-Net plans format. Attempting to reconstruct network architecture "
2026/01/12 11:57:07.486 :: #######################################################################
2026/01/12 11:57:07.486 :: Please cite the following paper when using nnU-Net:
2026/01/12 11:57:07.486 :: Isensee, F., Jaeger, P. F., Kohl, S. A., Petersen, J., & Maier-Hein, K. H. (2021). nnU-Net: a self-configuring method for deep learning-based biomedical image segmentation. Nature methods, 18(2), 203-211.
2026/01/12 11:57:07.486 :: #######################################################################
2026/01/12 11:57:07.486 :: 
2026/01/12 11:57:07.486 :: There are 1 cases in the source folder
2026/01/12 11:57:07.486 :: I am processing 0 out of 1 (max process ID is 0, we start counting with 0!)
2026/01/12 11:57:07.486 :: There are 1 cases that I would like to predict
2026/01/12 11:57:07.486 :: 
2026/01/12 11:57:07.486 :: Predicting volume:
2026/01/12 11:57:07.486 :: perform_everything_on_device: True
2026/01/12 11:57:07.486 :: 0%|          | 0/210 [00:00<?, ?it/s]
2026/01/12 11:57:08.566 :: 0%|          | 1/210 [00:01<03:45,  1.08s/it]
2026/01/12 11:57:08.790 :: 1%|          | 2/210 [00:01<01:59,  1.74it/s]
2026/01/12 11:57:09.119 :: 1%|1         | 3/210 [00:01<01:35,  2.16it/s]
2026/01/12 11:57:09.261 :: 2%|1         | 4/210 [00:01<01:09,  2.97it/s]
...
2026/01/12 11:57:36.607 :: 99%|#########8| 207/210 [00:29<00:00,  7.33it/s]
2026/01/12 11:57:36.742 :: 99%|#########9| 208/210 [00:29<00:00,  7.37it/s]
2026/01/12 11:57:36.882 :: 100%|#########9| 209/210 [00:29<00:00,  7.30it/s]
2026/01/12 11:57:37.009 :: 100%|##########| 210/210 [00:29<00:00,  7.47it/s]
2026/01/12 11:57:37.009 :: 100%|##########| 210/210 [00:29<00:00,  7.11it/s]
2026/01/12 11:57:55.780 :: sending off prediction to background worker for resampling and export
2026/01/12 11:57:55.780 :: done with volume
2026/01/12 11:57:56.939 :: Loading inference results...
2026/01/12 11:57:58.427 :: Post processing results...
2026/01/12 11:57:58.438 :: Remove small voxels for Maxilla & Upper Skull...
2026/01/12 11:57:59.833 :: Remove small voxels for Mandible...
2026/01/12 11:58:00.758 :: Remove small voxels for Upper Teeth...
2026/01/12 11:58:01.539 :: Remove small voxels for Lower Teeth...
2026/01/12 11:58:02.298 :: Post processing done.
2026/01/12 11:58:02.346 :: Inference ended successfully.

One difference I see is that you get this message:

2026/01/09 16:30:18.304 :: perform_everything_on_device=True is only supported for cuda devices! Setting this to False

What nnUnet version do you have installed (you can find it in nnUNet module in Slicer in the “nnUNet Install” section)?
What GPU do you have?
Does segmentation work if you select “Device” → “cpu” in the module?