Closed dagewa closed 6 years ago
Compare stats for scaling runs omitting one dataset each time. Does this identify the worst dataset? Do the stats improve? How about excluding ranges of the worst images?
Compare stats for scaling runs omitting one dataset each time. Does this identify the worst dataset? Do the stats improve? How about excluding ranges of the worst images?