paper-with-me

홈 › Papers

pi-Flow: Policy-Based Few-Step Generation via Imitation Distillation

2025-10-16 · Hansheng Chen, Kai Zhang, Hao Tan, Leonidas Guibas, Gordon Wetzstein, Sai Bi arxiv

Few-step diffusion or flow-based generative models typically distill a velocity-predicting teacher into a student that predicts a shortcut towards denoised data. This format mismatch has led to complex distillation procedures that often suffer from a quality-diversity trade-off. To address this, we propose policy-based flow models ($π$-Flow). $π$-Flow modifies the output layer of a student flow model to predict a network-free policy at one timestep. The policy then produces dynamic flow velocities at future substeps with negligible overhead, enabling fast and accurate ODE integration on these substeps without extra network evaluations. To match the policy's ODE trajectory to the teacher's, we introduce a novel imitation distillation approach, which matches the policy's velocity to the teacher's along the policy's trajectory using a standard $\ell_2$ flow matching loss. By simply mimicking the teacher's behavior, $π$-Flow enables stable and scalable training and avoids the quality-diversity trade-off. On ImageNet 256$^2$, it attains a 1-NFE FID of 2.85, outperforming previous 1-NFE models of the same DiT architecture. On FLUX.1-12B and Qwen-Image-20B at 4 NFEs, $π$-Flow achieves substantially better diversity than state-of-the-art DMD models, while maintaining teacher-level quality.

📄 PDF Abstract BibTeX arXiv:2510.14974

Code (0)

등록된 구현이 없습니다.

Similar Papers 제목 키워드 기반

AnyFlow: Any-Step Video Diffusion Model with On-Policy Flow Map Distillation

2026-05-13 · Yuchao Gu, Guian Fang, Yuxin Jiang, Weijia Mao 외 arxiv

Few-step video generation has been significantly advanced by consistency distillation. However, the performance of consistency-distilled models often degrades as more sampling steps are allocated at test time, limiting t…

Video Generation

Fast and Highly Expressive Policy Learning for Offline Reinforcement Learning via Bootstrapped Flow Q-Learning

2026-06-09 · Thanh Nguyen, Tri Ton, Hongbin Choe, Tung M. Luu 외 arxiv

Diffusion-based Q-learning has emerged as a powerful paradigm for offline reinforcement learning, but its reliance on multi-step denoising makes both training and inference computationally expensive and brittle. Recent e…

Reinforcement Learning

One-Step Generative Policies with Q-Learning: A Reformulation of MeanFlow

2025-11-17 · Zeyuan Wang, Da Li, Yulin Chen, Ye Shi 외 arxiv

We introduce a one-step generative policy for offline reinforcement learning that maps noise directly to actions via a residual reformulation of MeanFlow, making it compatible with Q-learning. While one-step Gaussian pol…

Reinforcement Learning

On-Policy Adversarial Flow Distillation for Autoregressive Video Generation

2026-05-25 · Yang Luo, Shengju Qian, Xiaohang Tang, Zirui Zhu 외 arxiv

Autoregressive video generators are attractive for streaming, long-horizon, and interactive applications, but distilling strong black-box teachers into causal students remains difficult. The student must learn under its …

Reinforcement LearningVideo Generation

One-Step Flow Q-Learning: Addressing the Diffusion Policy Bottleneck in Offline Reinforcement Learning

2025-08-19 · Thanh Nguyen, Chang D. Yoo arxiv

Diffusion Q-Learning (DQL) has established diffusion policies as a high-performing paradigm for offline reinforcement learning, but its reliance on multi-step denoising for action generation renders both training and inf…

Reinforcement Learning