- Are you in the terminal you mean to be in (i.e. did you ssh into a different machine or fail to do so?)
- If you are trying to launch over an ssh connection and failing, is your VPN on?
- Do you have a conda environment active? Should you?
- Is your port number in use? The default port for jupyter notebooks is 8888. You can use a different one with the
--port=XXXXflag when launching, e.g.jupyter notebook --port=1234. Ports are always integers between 0 and 65353 (but you don't generally want to use values less than 1023).
- Is this the kernel you mean to be using? Check the drop-down kernel menu to make sure.
- Try copying the code up to where your kernel dies into a python script and running from there instead. This can reveal certain errors that are sometimes hidden by the notebook kernel crashing.
- Are you running out of memory? This can be tough to diagnose. Try running
topin a separate commandline while you run your notebook, and if your memory use hits 100% this could be your issue.