FlashRender: Few-step camera-controlled generative rendering via MeanFlow distillation

everex · hf · 2026-09-04

FlashRender accelerates generative video rendering to a few steps while keeping quality. It combines representation alignment, a mean-flow training objective, and on-policy distillation to achieve high-quality few-step camera-controlled synthesis. The model is available on Hugging Face.

Original post →

More from Multimodal

Multimodal channel →