H3 excels at generating complex space scenes

SIR_NVAX_A_LOT · reddit · 2026-08-24

A showcase of the H3 model's text-to-video capabilities, highlighting its strength in rendering rich, intricate environments, particularly space scenes. Generated using T2VA, int8, and 20 steps.

Original post →

More from Multimodal

Multimodal channel →