My thoughts
#1
by Jam747 - opened
I am using a two pass workflow, (low and upscaled latent), I set the 0.8 at the first pass and 0.3 at the upscaled pass, but I don not see any difference, probably I am missing something.
When H3 already handles a scene well, the difference may be very small / when the base model begins to struggle, the difference can become much more obvious.