Update benchmarks

[htjb]

Updating the benchmarks and adding in physical benchmark examples.

1. Need to re-run figure 4 and the analysis that went into table 1 with equal numbers of hyperparameters across the different methods and given the discussion in this margarine PR and this margarine issue.
2. Want to run a test to demonstrate that the PNFs are not as costly as the individual MAFs with the toy benchmarks.
3. Need to test the clustering on physical examples
   • This might involve implementing a new clustering example