When speaking of “render time”, do you actually mean training time + rendering of all pixels?
Marc-Andre Weibezahn

Yes, when I speak of “render time,” I mean training time + rendering all of the pixels. On resolutions under 1080p the brunt of the time is spent training — maybe a 3:1 ratio. Hill-climbing is definitely less than efficient. Though, I wonder if back-propagation would yield such varied, and interesting, results.

One clap, two clap, three clap, forty?

By clapping more or less, you can signal to us which stories really stand out.