Hmm, one thing you might try is to double check that the jupyter
command you ran to launch the notebook is installed in your fastai
virtualenv/conda environment, versus being some other installation of jupyter
. That is, double check that running which jupyter
at the command line prints something like some/path/to/envs/fastai/bin/jupyter
, as opposed to something global like ~/anaconda/bin/jupyter
. You could also run pip freeze
from within your virtualenv and see if jupyter is in there.
I’ve made that mistake before!