Can we have some sort of image output samples underway?
Run parameters on the PCA?
Progress of the learning? (In the next iteration we will go into depth with this, are we overfitting the data? What is the expected out-of-sample error with of VC dimension etc.)
Investigate: