Skip to content

Source paper

FireRed-Image-Edit-1.0 Techinical Report

Super Intelligence Team, Changhao Qiao, Chao Hui and 16 more

Super Intelligence Team, Changhao Qiao, Chao Hui, Chen Li, Cunzheng Wang, Dejia Song, Jiale Zhang, Jing Li, Qiang Xiang, Runqi Wang, Shuang Sun, Wei Zhu, Xu Tang, Yao Hu, Yibo Chen, Yuhao Huang, Yuxuan Duan, Zhiyi Chen, Ziyuan Guo

arXiv 2602.13344 · 20264 upvotes on Hugging Face

Research ideas from this paper (3)

Generated by AI. The paper’s authors did not write these.

AI summary

This paper introduces FireRed-Image-Edit, a state-of-the-art instruction-based image editing model trained on a rigorously curated dataset of over 100 million image pairs. The authors present a comprehensive end-to-end framework that optimizes data curation, diffusion transformer architecture, training efficiency, and reinforcement learning strategies to achieve high-fidelity, semantically accurate edits.

How it works (AI summary)

The model utilizes a Double-Stream Multi-Modal Diffusion Transformer (MM-DiT) architecture and a multi-stage training pipeline comprising pre-training, supervised fine-tuning, and RLHF using Direct Preference Optimization (DPO) and DiffusionNFT. It integrates novel efficiency optimizations like a Multi-Condition Aware Bucket Sampler and a facial landmark-based identity consistency loss to stabilize generation.