Number of Iterations for every k value

[kelvinmo0513]

Hi Team,

This is a great body of work. I'm using cNMF on a few extremely large single-cell datasets (~150GB) and it is taking forever even when using parallel computing. Would you recommend lowering the number of iterations for every k (10-25) or do you suggest keeping it at default 100? In other words, is there anything else I can tune to make it less time consuming?

Thank you!

Best,
Kelvin