DeepGen 1.0: A Lightweight Unified Multimodal Model for Advancing Image Generation and Editing
Fuente:
arXiv
Salvato in:
| Autori principali: | , , , , , , , , , , , , , , , , , , , |
|---|---|
| Natura: | Preprint |
| Pubblicazione: |
2026
|
| Soggetti: | |
| Accesso online: | |
| Tags: |
Aggiungi Tag
Nessun Tag, puoi essere il primo ad aggiungerne!!
|
| _version_ | 1866911446073344000 |
|---|---|
| author | Wang, Dianyi Li, Ruihang Han, Feng Ma, Chaofan Song, Wei Wang, Siyuan Wang, Yibin Xin, Yi Liu, Hongjian Zhang, Zhixiong Ding, Shengyuan Wang, Tianhang Cheng, Zhenglin Lin, Tao Jin, Cheng Yu, Kaicheng Chen, Jingjing Wang, Wenjie Wei, Zhongyu Wang, Jiaqi |
| author_facet | Wang, Dianyi Li, Ruihang Han, Feng Ma, Chaofan Song, Wei Wang, Siyuan Wang, Yibin Xin, Yi Liu, Hongjian Zhang, Zhixiong Ding, Shengyuan Wang, Tianhang Cheng, Zhenglin Lin, Tao Jin, Cheng Yu, Kaicheng Chen, Jingjing Wang, Wenjie Wei, Zhongyu Wang, Jiaqi |
| contents | Current unified multimodal models for image generation and editing typically rely on massive parameter scales (e.g., >10B), entailing prohibitive training costs and deployment footprints. In this work, we present DeepGen 1.0, a lightweight 5B unified model that achieves comprehensive capabilities competitive with or surpassing much larger counterparts. To overcome the limitations of compact models in semantic understanding and fine-grained control, we introduce Stacked Channel Bridging (SCB), a deep alignment framework that extracts hierarchical features from multiple VLM layers and fuses them with learnable 'think tokens' to provide the generative backbone with structured, reasoning-rich guidance. We further design a data-centric training strategy spanning three progressive stages: (1) Alignment Pre-training on large-scale image-text pairs and editing triplets to synchronize VLM and DiT representations, (2) Joint Supervised Fine-tuning on a high-quality mixture of generation, editing, and reasoning tasks to foster omni-capabilities, and (3) Reinforcement Learning with MR-GRPO, which leverages a mixture of reward functions and supervision signals, resulting in substantial gains in generation quality and alignment with human preferences, while maintaining stable training progress and avoiding visual artifacts. Despite being trained on only ~50M samples, DeepGen 1.0 achieves leading performance across diverse benchmarks, surpassing the 80B HunyuanImage by 28% on WISE and the 27B Qwen-Image-Edit by 37% on UniREditBench. By open-sourcing our training code, weights, and datasets, we provide an efficient, high-performance alternative to democratize unified multimodal research. |
| format | Preprint |
| id |
arxiv_https___arxiv_org_abs_2602_12205 |
| institution | arXiv |
| publishDate | 2026 |
| record_format | arxiv |
| spellingShingle | DeepGen 1.0: A Lightweight Unified Multimodal Model for Advancing Image Generation and Editing Wang, Dianyi Li, Ruihang Han, Feng Ma, Chaofan Song, Wei Wang, Siyuan Wang, Yibin Xin, Yi Liu, Hongjian Zhang, Zhixiong Ding, Shengyuan Wang, Tianhang Cheng, Zhenglin Lin, Tao Jin, Cheng Yu, Kaicheng Chen, Jingjing Wang, Wenjie Wei, Zhongyu Wang, Jiaqi Computer Vision and Pattern Recognition Artificial Intelligence Current unified multimodal models for image generation and editing typically rely on massive parameter scales (e.g., >10B), entailing prohibitive training costs and deployment footprints. In this work, we present DeepGen 1.0, a lightweight 5B unified model that achieves comprehensive capabilities competitive with or surpassing much larger counterparts. To overcome the limitations of compact models in semantic understanding and fine-grained control, we introduce Stacked Channel Bridging (SCB), a deep alignment framework that extracts hierarchical features from multiple VLM layers and fuses them with learnable 'think tokens' to provide the generative backbone with structured, reasoning-rich guidance. We further design a data-centric training strategy spanning three progressive stages: (1) Alignment Pre-training on large-scale image-text pairs and editing triplets to synchronize VLM and DiT representations, (2) Joint Supervised Fine-tuning on a high-quality mixture of generation, editing, and reasoning tasks to foster omni-capabilities, and (3) Reinforcement Learning with MR-GRPO, which leverages a mixture of reward functions and supervision signals, resulting in substantial gains in generation quality and alignment with human preferences, while maintaining stable training progress and avoiding visual artifacts. Despite being trained on only ~50M samples, DeepGen 1.0 achieves leading performance across diverse benchmarks, surpassing the 80B HunyuanImage by 28% on WISE and the 27B Qwen-Image-Edit by 37% on UniREditBench. By open-sourcing our training code, weights, and datasets, we provide an efficient, high-performance alternative to democratize unified multimodal research. |
| title | DeepGen 1.0: A Lightweight Unified Multimodal Model for Advancing Image Generation and Editing |
| topic | Computer Vision and Pattern Recognition Artificial Intelligence |
| url | https://arxiv.org/abs/2602.12205 |