Multiple datasets with different labels.

[j-onne]

Hi,
First, thank you for making nnUNet!
For my project i'm segmenting different features in MRI image in the knee. I have two datasets. One that contains labels 1 to 5 (and has a total of 176 volumes) and another dataset that contains labels 1, 2, 6 and 7 (total of 500 volumes). Both dataset have the same size MRI images, same voxel spacing and same mri sequence.
list of all labels:
label 1 = cartilage femur
label 2 = cartilage tibia
label 3 = cartilage patella
label 4 = meniscus lateral
label 5 = meniscus medial
label 6 = femur (bone)
label 7 = tibia (bone)

Is it possible to train both datasets in one network? Or will this cause the network to 'think' that sometimes there is no bone, when indeed there is bone there it is just not labeled?

Hope my question is clear. please let me know if any additional information is needed.