OneReward: Unified Mask-Guided Image Generation via Multi-Task Human Preference Learning

Fuente: arXiv
Salvato in:
Dettagli Bibliografici
Autori principali: Gong, Yuan, Wang, Xionghui, Wu, Jie, Wang, Shiyin, Wang, Yitong, Wu, Xinglong
Natura: Preprint
Pubblicazione: 2025
Soggetti:
Accesso online:
Tags: Aggiungi Tag
Nessun Tag, puoi essere il primo ad aggiungerne!!
_version_ 1866911128091623424
author Gong, Yuan
Wang, Xionghui
Wu, Jie
Wang, Shiyin
Wang, Yitong
Wu, Xinglong
author_facet Gong, Yuan
Wang, Xionghui
Wu, Jie
Wang, Shiyin
Wang, Yitong
Wu, Xinglong
contents In this paper, we introduce OneReward, a unified reinforcement learning framework that enhances the model's generative capabilities across multiple tasks under different evaluation criteria using only \textit{One Reward} model. By employing a single vision-language model (VLM) as the generative reward model, which can distinguish the winner and loser for a given task and a given evaluation criterion, it can be effectively applied to multi-task generation models, particularly in contexts with varied data and diverse task objectives. We utilize OneReward for mask-guided image generation, which can be further divided into several sub-tasks such as image fill, image extend, object removal, and text rendering, involving a binary mask as the edit area. Although these domain-specific tasks share same conditioning paradigm, they differ significantly in underlying data distributions and evaluation metrics. Existing methods often rely on task-specific supervised fine-tuning (SFT), which limits generalization and training efficiency. Building on OneReward, we develop Seedream 3.0 Fill, a mask-guided generation model trained via multi-task reinforcement learning directly on a pre-trained base model, eliminating the need for task-specific SFT. Experimental results demonstrate that our unified edit model consistently outperforms both commercial and open-source competitors, such as Ideogram, Adobe Photoshop, and FLUX Fill [Pro], across multiple evaluation dimensions. Code and model are available at: https://one-reward.github.io
format Preprint
id arxiv_https___arxiv_org_abs_2508_21066
institution arXiv
publishDate 2025
record_format arxiv
spellingShingle OneReward: Unified Mask-Guided Image Generation via Multi-Task Human Preference Learning
Gong, Yuan
Wang, Xionghui
Wu, Jie
Wang, Shiyin
Wang, Yitong
Wu, Xinglong
Computer Vision and Pattern Recognition
In this paper, we introduce OneReward, a unified reinforcement learning framework that enhances the model's generative capabilities across multiple tasks under different evaluation criteria using only \textit{One Reward} model. By employing a single vision-language model (VLM) as the generative reward model, which can distinguish the winner and loser for a given task and a given evaluation criterion, it can be effectively applied to multi-task generation models, particularly in contexts with varied data and diverse task objectives. We utilize OneReward for mask-guided image generation, which can be further divided into several sub-tasks such as image fill, image extend, object removal, and text rendering, involving a binary mask as the edit area. Although these domain-specific tasks share same conditioning paradigm, they differ significantly in underlying data distributions and evaluation metrics. Existing methods often rely on task-specific supervised fine-tuning (SFT), which limits generalization and training efficiency. Building on OneReward, we develop Seedream 3.0 Fill, a mask-guided generation model trained via multi-task reinforcement learning directly on a pre-trained base model, eliminating the need for task-specific SFT. Experimental results demonstrate that our unified edit model consistently outperforms both commercial and open-source competitors, such as Ideogram, Adobe Photoshop, and FLUX Fill [Pro], across multiple evaluation dimensions. Code and model are available at: https://one-reward.github.io
title OneReward: Unified Mask-Guided Image Generation via Multi-Task Human Preference Learning
topic Computer Vision and Pattern Recognition
url https://arxiv.org/abs/2508.21066