Take stock of all models that currently can't be trained

Full test run (this might need a bit of an adjustment on the tests), to run through all the models available in `ssm-simulators`.

This is to take stock of all failing models to define next steps.

Start by running the full test suite across all models available in `ssm-simulators`, adjusting the tests if required to exercise them consistently. Record which models fail to train and use those failures to define the next steps; completion is a complete inventory of failing models.

Written by the indexing model from the issue text.