Use new training container path

Also use pr.create.job.RunnerFit syntax. This clears the output of the cells since I couldn't run the fit itself on the workshop nodes yet.

Two short feedback points:

1. you frequently print datacontainers used for input. If you do not you're supposed to see a nice interactive widget, did that not work for you?
2. At the end you manually do an E-V curve, the 'pyironic' would be to do a Murnaghan job as is shown in the introduction notebooks.

Feel free to ignore these though, given the time constraints we have. Point 2 could actually be an exercise for the participants.

EDIT: I also got an error on importing ipympl, do we need to put that into the environment?