Show HN: We compressed a 31B video+audio diffusion model to 14B, 6 steps, INT8

Sign in to join the discussion — your thebay.events account works here.