I skipped naming convensions inside LBFGSB to match the uppercase matrix symbols of the paper.
Also some variables appear as undefined to flake8, because they are defined at the end of the loop, and only used once the gradient history is full. Not sure if this is bad practice in general. I could initialize the variables.
I skipped naming convensions inside LBFGSB to match the uppercase matrix symbols of the paper.
Also some variables appear as undefined to flake8, because they are defined at the end of the loop, and only used once the gradient history is full. Not sure if this is bad practice in general. I could initialize the variables.