跳到正文
HuggingFace Daily Papers(社區熱門論文)·· 19 天前AI 評分45

StepX-OPD:重新思考擴散模型的多教師在策略蒸餾

StepX-OPD: Rethinking Multi-Teacher On-Policy Distillation in Diffusion Models

AI 導讀

StepX-OPD 將多教師在策略蒸餾與少步採樣整合訓練,讓文字生成圖像擴散模型同時習得多項能力並以少量步驟生成。研究在 SD3.5-Medium 和 FLUX.1-dev 上評估,StepX-OPD 在比較的四步方法中取得最高標準化平均分,並超越四十步 Flow-OPD。於 SD3.5-Medium 上,其 GPU 訓練時數較先後進行 DiffusionOPD 和步數蒸餾減少約 37%。

正文

Published on Nov 3

Authors:

,

,

,

,

,

,

Abstract

StepX-OPD jointly trains a text-to-image diffusion model to acquire capabilities and few-step sampling simultaneously, outperforming sequential approaches while reducing training cost.

A deployable text-to-image diffusion model must hold several capabilities at once, such as instruction following, text rendering and human preference, and must sample in a few steps. Capabilities and few-step sampling are built in separate stages: a full-step model first consolidates capabilities through multi-teacher on-policy distillation, then serves as a fixed teacher for step distillation. Every full-step model we evaluate scores lower once compressed, and the second stage cannot begin until the first finishes. We introduce StepX-OPD, which merges the two stages: one student acquires every capability while learning to sample in a few steps, so frozen specialists supervise it on the states its own rollout visits rather than on a longer trajectory it never runs. Because a decomposition of the rollout's KL weights velocity mismatch most heavily near the clean end, we also regress the student's velocity on the low-noise part of the rollout. Separately, we find that capability integration makes models write prompt words into images on prompts that never asked for text, and we give a variant that lowers it. We evaluate on SD3.5-Medium and FLUX.1-dev across instruction following, text rendering, image quality and human preference. On both architectures, StepX-OPD attains the highest normalized average score among the four-step methods compared and exceeds the forty-step Flow-OPD model, and on SD3.5-Medium it trains with about 37% fewer GPU-hours than sequential DiffusionOPD followed by step distillation.

View arXiv page View PDF Project page GitHub 2 Add to collection

Models citing this paper 1

ShimaRinC/StepX-OPD-SD3.5

Text-to-Image •

Updated 3 days ago

•

16

Datasets citing this paper 0

No dataset linking this paper

Cite arxiv.org/abs/2611.04386 in a dataset README.md to link it from this page.

Spaces citing this paper 0

No Space linking this paper

Cite arxiv.org/abs/2611.04386 in a Space README.md to link it from this page.

Collections including this paper 0

No Collection including this paper

Add this paper to a collection to link it from this page.

來源:HuggingFace Daily Papers(社區熱門論文) · huggingface.co