almost certainly, i haven't tried the most recent models but i have used hy3d2 and hy3d2-fast a lot and they're quite light to inference. You're gonna spend mor

almost certainly, i haven't tried the most recent models but i have used hy3d2 and hy3d2-fast a lot and they're quite light to inference. You're gonna spend more time decoding the latent than you will inferencing. Takes about 6gb vram on my machine, I can't imagine these will be heavier.