ReasonEdit：面向推理增强的图像编辑模型
ReasonEdit: Towards Reasoning-Enhanced Image Editing Models
把冻结的 MLLM 编码器解锁为可推理模块，用「思考—编辑—反思」循环让模型理解抽象指令并自动纠错，为指令式编辑提供了一条不依赖更多配对数据、而靠推理提升精度的可复用范式。
原文依据 our proposed framework enables image editing in a thinking-editing-reflection loop
摘要译文
由模型翻译 · 术语统一 · 请对照原文核对关键结论近期图像编辑模型取得了显著进展。一种常见的架构设计将多模态大语言模型（MLLM）编码器与扩散解码器耦合，如 Step1X-Edit 和 Qwen-Image-Edit 等系统所示，其中 MLLM 对参考图像和指令进行编码，但在训练过程中保持冻结。在本工作中，我们证明释放 MLLM 的推理能力可以进一步拓展编辑模型的边界。具体而言，我们探索了思考与反思两种推理机制，它们分别增强指令理解与编辑准确性。基于此，我们提出的框架能够在「思考—编辑—反思」循环中完成图像编辑：思考机制利用 MLLM 的世界知识来解读抽象指令，而反思机制则审视编辑结果、自动纠正非预期的改动，并判定停止轮次。大量实验表明，我们的推理方法取得了显著的性能提升：当以 Step1X-Edit 初始化我们的 DiT 时（ReasonEdit-S），在 ImgEdit（+4.4%）、GEdit（+3.1%）和 Kris（+11.5%）上均有提升；当与 Qwen-Image-Edit 结合时（ReasonEdit-Q），在 GEdit 和 Kris 上也优于此前的开源方法。代码与检查点将开源。
英文原文摘要 1338 字符
Recent advances in image editing models have shown remarkable progress. A common architectural design couples a multimodal large language model (MLLM) encoder with a diffusion decoder, as seen in systems such as Step1X-Edit and Qwen-Image-Edit, where the MLLM encodes both the reference image and the instruction but remains frozen during training.In this work, we demonstrate that unlocking the reasoning capabilities of MLLM can further push the boundaries of editing models. Specifically, we explore two reasoning mechanisms, thinking and reflection, which enhance instruction understanding and editing accuracy.Based on that, our proposed framework enables image editing in a thinking-editing-reflection loop: the thinking mechanism leverages the world knowledge of MLLM to interpret abstract instructions, while the reflection reviews editing results, automatically corrects unintended manipulations, and identifies the stopping round.Extensive experiments demonstrate that our reasoning approach achieves significant performance gains, with improvements of ImgEdit(+4.4%), GEdit(+3.1%), and Kris(+11.5%) when initializing our DiT from the Step1X-Edit(ReasonEdit-S), and also outperforms previous open-source methods on both GEdit and Kris when integrated with Qwen-Image-Edit(ReasonEdit-Q).Code and checkpoints will be open-sourced.