Nvmidl-convert problem

Hi, I want to fine tune pre-trained model for liver segmentation.
It should be trained on Task03_Liver dataset from Medical Decathlon.

I download 3Dircadb dataset which contains dicom images.
I am trying to convert them to nifti format. The problem is that it cant do conversion.
Paths are correct. If I rename it from image_0 to image_0.dcm it can read it. But it will create only one files caled .nii for all images in folder.

nvmidl-dataconvert -d /workspace/data/3Dircadb_liver/labelsTr -r 1 -s .dcm -e .nii.gz -o /workspace/data/Task03_Liver_nii/labelsTr

Do you have any idea what could be wrong?
I am obtaining this announcement (I have more then 100 input images): Converting dicom series 1 of 1 in .
Thanks for your help.

Hi
Thanks for your interest in clara train
If I remember correctly the medical decathlon dataset is already in nifti so you should not convert dcm to nii but instead change the resolution by converting from nii to nii with the -r 1 flag

If I am mistaken and you to have dicom images. Our tool does basic conversion so may be it can’t read these set of dicom. if that is the case then may be you can use a tool like dcm2niix NITRC
to convert dcm to nifti then use our tool to do the resampling

you can see the tool in action in this notebook clara-train-examples/TCIADownloader.ipynb at master · NVIDIA/clara-train-examples · GitHub
see step 0.2 to download the converting tool and use it