Segmentation causes Jupyter kernel to die and restart

[estep-1]

I've been trying to use the interactive segmentation. It works when I use one foreground point, but if I add more points the kernel dies. The same thing happens if I use input points instead of the interactive segmentation. I can run the automatic segmentation, but my kernel dies if I try to run anything afterwards. I have 56 GB of memory and I'm on a GPU. What could be causing the kernel to die? Is it something to do with the specs of my computer, or could it be the environment installation? Any insight is appreciated! :)