Segmentation process gets stuck sometimes

[mik0699]

Hello, first of all thanks for your amazing work! I am actually encountering a weird problem. Sometimes the segmentation process gets stuck and does not proceed anymore, it just loops infinitely and I have to stop the execution manually. I am using the Pyhton API and the roi_subset attribute. The weird aspect is that, with the same data, if I add some classes to the roi_subset or, for example, I add the fast option, the segmentation finishes, while in other cases in does not. I can't understand why this happens only sometimes and what could be the problem because the data I use are the same. The segmentation process gets stuck in this point:

There are 1 cases in the source folder I am process 0 out of 1 (max process ID is 0, we start counting with 0!) There are 1 cases that I would like to predict

and does not proceed anymore, it just shows me this message.

Thanks in advance for your help.

[wasserth]

Maybe some of the nnU-Net processes run out of memory. Does it work with a smaller image?

[mik0699]

Thanks for the reply, I solved it using GPU instead of CPU, so maybe the problem was exactly that one you mentioned