update the max_lr param for the supervised marker map from 0.0001 to 0.001. It turns out this was what was occasionally causing worse results for some of the data sets for the Supervised MM. I had miscopied this parameter from the notebooks.
add a small script to make it easier to print the results as i get them
Testing
ran the models, they now seem to be performing how they are in the notebooks
Changes
Testing