I am using min_mover (lbfgs_armijo_nonmonotone) for minimizing the energy function of a protein for structure prediction in pyrosetta.
In trying to get an energy plot over iterations, i set min_mover.max_iter(1) to 1. After each iteration of min_mover, I pause the optimization, store the current energy, and continue.
However, I've found that repeating min_mover.max_iter(1) in a for-loop 100 times is not the same as running min_mover.max_iter(100). Why is that? It seems like min_mover.max_iter(100) finds significantly lower energies than min_mover.max_iter(1) repeated 100 times. Is this expected behaviour? Thanks for any help!
Category:
Post Situation:
> In trying to get an energy plot over iterations
This post https://www.rosettacommons.org/node/11250 may be helpful —contains code that dumps the pose at every _n_ iteration.
This is for diagnostic use only as you quickly get a godzillean file footprint.