EFlow: Fast Few-Step Video Generator Training from Scratch via Efficient Solution Flow

Fuente: arXiv
Saved in:
Bibliographic Details
Main Authors: Park, Dogyun, Li, Yanyu, Tulyakov, Sergey, Kag, Anil
Format: Preprint
Published: 2026
Subjects:
Online Access:
Tags: Add Tag
No Tags, Be the first to tag this record!
_version_ 1866908918868869120
author Park, Dogyun
Li, Yanyu
Tulyakov, Sergey
Kag, Anil
author_facet Park, Dogyun
Li, Yanyu
Tulyakov, Sergey
Kag, Anil
contents Scaling video diffusion transformers is fundamentally bottlenecked by two compounding costs: the expensive quadratic complexity of attention per step, and the iterative sampling steps. In this work, we propose EFlow, an efficient few-step training framework, that tackles these bottlenecks simultaneously. To reduce sampling steps, we build on a solution-flow objective that learns a function mapping a noised state at time t to time s. Making this formulation computationally feasible and high-quality at video scale, however, demands two complementary innovations. First, we propose Gated Local-Global Attention, a token-droppable hybrid block which is efficient, expressive, and remains highly stable under aggressive random token-dropping, substantially reducing per-step compute. Second, we develop an efficient few-step training recipe. We propose Path-Drop Guided training to replace the expensive guidance target with a computationally cheap, weak path. Furthermore, we augment this with a Mean-Velocity Additivity regularizer to ensure high fidelity at extremely low step counts. Together, our EFlow enables a practical from-scratch training pipeline, achieving up to 2.5x higher training throughput over standard solution-flow, and 45.3x lower inference latency than standard iterative models with competitive performance on Kinetics and large-scale text-to-video datasets.
format Preprint
id arxiv_https___arxiv_org_abs_2603_27086
institution arXiv
publishDate 2026
record_format arxiv
spellingShingle EFlow: Fast Few-Step Video Generator Training from Scratch via Efficient Solution Flow
Park, Dogyun
Li, Yanyu
Tulyakov, Sergey
Kag, Anil
Computer Vision and Pattern Recognition
Scaling video diffusion transformers is fundamentally bottlenecked by two compounding costs: the expensive quadratic complexity of attention per step, and the iterative sampling steps. In this work, we propose EFlow, an efficient few-step training framework, that tackles these bottlenecks simultaneously. To reduce sampling steps, we build on a solution-flow objective that learns a function mapping a noised state at time t to time s. Making this formulation computationally feasible and high-quality at video scale, however, demands two complementary innovations. First, we propose Gated Local-Global Attention, a token-droppable hybrid block which is efficient, expressive, and remains highly stable under aggressive random token-dropping, substantially reducing per-step compute. Second, we develop an efficient few-step training recipe. We propose Path-Drop Guided training to replace the expensive guidance target with a computationally cheap, weak path. Furthermore, we augment this with a Mean-Velocity Additivity regularizer to ensure high fidelity at extremely low step counts. Together, our EFlow enables a practical from-scratch training pipeline, achieving up to 2.5x higher training throughput over standard solution-flow, and 45.3x lower inference latency than standard iterative models with competitive performance on Kinetics and large-scale text-to-video datasets.
title EFlow: Fast Few-Step Video Generator Training from Scratch via Efficient Solution Flow
topic Computer Vision and Pattern Recognition
url https://arxiv.org/abs/2603.27086