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

TL;DR

EFlow introduces an efficient training framework for video diffusion transformers, significantly reducing computation and inference time while maintaining high-quality video generation.

Contribution

The paper presents Gated Local-Global Attention and Path-Drop Guided training, enabling fast few-step video generation from scratch.

Findings

Achieves 2.5x higher training throughput compared to standard solution-flow.

Attains 45.3x lower inference latency than standard iterative models.

Demonstrates competitive performance on Kinetics and large-scale text-to-video datasets.

Abstract

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.