FlashRender: Few-Step Generative Rendering via Camera-Controlled Video MeanFlow

Hugging Face Daily Papers Papers

Summary

FlashRender is a few-step generative rendering framework that accelerates video synthesis by aligning representations and using mean-flow objectives, achieving comparable quality to multi-step methods with significantly reduced sampling cost.

We present FlashRender, a few-step generative rendering framework that retakes a source video along a target camera trajectory in seconds. We identify sampling-step-dependent camera control as a prominent manifestation of discretization error in existing multi-step generative rendering models and show that resolving this inconsistency substantially lowers denoising trajectory curvature, facilitating subsequent step distillation. To this end, we introduce Representation Transformation and Alignment (RETA), which aligns hidden source-video representations with target-video features from a frozen visual geometry model. This directly encodes the geometric transformation within the source-video stream, enabling sampling-step-consistent camera control. We then fine-tune the model with the MeanFlow objective on the lower-curvature denoising trajectory induced by RETA, allowing the model to more effectively address discretization error. Finally, we apply on-policy flow map distillation to correct self-rollout errors under fixed few-step sampling. Extensive experiments show that RETA, MeanFlow, and on-policy flow map distillation play complementary roles in few-step generative rendering. Together, they enable our approach to match multi-step baselines in video quality and geometric consistency at 25x lower sampling cost while achieving superior camera controllability, even under out-of-distribution target camera trajectories.
Original Article
View Cached Full Text

Cached at: 09/04/26, 03:55 AM

Paper page - FlashRender: Few-Step Generative Rendering via Camera-Controlled Video MeanFlow

Source: https://huggingface.co/papers/2609.03563

Abstract

FlashRender accelerates generative video rendering via representation alignment, a mean-flow objective, and on-policy distillation to achieve high-quality few-step camera-controlled synthesis.

We presentFlashRender, a few-stepgenerative renderingframework that retakes a source video along a target camera trajectory in seconds. We identifysampling-step-dependent camera controlas a prominent manifestation ofdiscretization errorin existing multi-stepgenerative renderingmodels and show that resolving this inconsistency substantially lowers denoising trajectory curvature, facilitating subsequent step distillation. To this end, we introduceRepresentation Transformation and Alignment (RETA), which aligns hidden source-video representations with target-video features from a frozenvisual geometry model. This directly encodes the geometric transformation within the source-video stream, enabling sampling-step-consistent camera control. We then fine-tune the model with theMeanFlow objectiveon the lower-curvature denoising trajectory induced by RETA, allowing the model to more effectively addressdiscretization error. Finally, we applyon-policy flow map distillationto correct self-rollout errors under fixedfew-step sampling. Extensive experiments show that RETA, MeanFlow, andon-policy flow map distillationplay complementary roles in few-stepgenerative rendering. Together, they enable our approach to match multi-step baselines in video quality and geometric consistency at 25x lower sampling cost while achieving superior camera controllability, even under out-of-distribution target camera trajectories.

View arXiv pageView PDFProject pageAdd to collection

Get this paper in your agent:

hf papers read 2609\.03563

Don’t have the latest CLI?curl \-LsSf https://hf\.co/cli/install\.sh \| bash

Models citing this paper0

No model linking this paper

Cite arxiv.org/abs/2609.03563 in a model README.md to link it from this page.

Datasets citing this paper0

No dataset linking this paper

Cite arxiv.org/abs/2609.03563 in a dataset README.md to link it from this page.

Spaces citing this paper0

No Space linking this paper

Cite arxiv.org/abs/2609.03563 in a Space README.md to link it from this page.

Collections including this paper0

No Collection including this paper

Add this paper to acollectionto link it from this page.

Similar Articles

Generative World Renderer at the Speed of Play

Hugging Face Daily Papers

This paper introduces AlayaRenderer-Flash, a real-time generative world renderer that accelerates rendering from 0.56 FPS to 31.54 FPS using a few-step autoregressive streaming model and lightweight distilled codecs, enabling interactive play with a physics engine.

Flexible Video Diffusion (3 minute read)

TLDR AI

Flex-Forcing introduces a unified framework for video diffusion that supports both autoregressive and bidirectional generation modes, offering flexible control for video generation tasks.

Streaming Video Generation with Streaming Force Control

Hugging Face Daily Papers

StreamForce is a causal, unified video generation model that provides real-time, physically grounded responses to time-varying forces through a distillation pipeline and autoregressive architecture, achieving state-of-the-art performance in force adherence and motion realism.