Saved in:
| Main Authors: | , , , , , , , , , , |
|---|---|
| Format: | Preprint |
| Published: |
2025
|
| Subjects: | |
| Online Access: | https://arxiv.org/abs/2508.10424 |
| Tags: |
Add Tag
No Tags, Be the first to tag this record!
|
| _version_ | 1866916898491334656 |
|---|---|
| author | Liu, Shanyuan Zhu, Jian Lu, Junda Gong, Yue Li, Liuzhuozheng Cheng, Bo Ma, Yuhang Wu, Liebucha Wu, Xiaoyu Leng, Dawei Yin, Yuhui |
| author_facet | Liu, Shanyuan Zhu, Jian Lu, Junda Gong, Yue Li, Liuzhuozheng Cheng, Bo Ma, Yuhang Wu, Liebucha Wu, Xiaoyu Leng, Dawei Yin, Yuhui |
| contents | Diffusion Transformers (DiTs) have demonstrated exceptional capabilities in text-to-image synthesis. However, in the domain of controllable text-to-image generation using DiTs, most existing methods still rely on the ControlNet paradigm originally designed for UNet-based diffusion models. This paradigm introduces significant parameter overhead and increased computational costs. To address these challenges, we propose the Nano Control Diffusion Transformer (NanoControl), which employs Flux as the backbone network. Our model achieves state-of-the-art controllable text-to-image generation performance while incurring only a 0.024\% increase in parameter count and a 0.029\% increase in GFLOPs, thus enabling highly efficient controllable generation. Specifically, rather than duplicating the DiT backbone for control, we design a LoRA-style (low-rank adaptation) control module that directly learns control signals from raw conditioning inputs. Furthermore, we introduce a KV-Context Augmentation mechanism that integrates condition-specific key-value information into the backbone in a simple yet highly effective manner, facilitating deep fusion of conditional features. Extensive benchmark experiments demonstrate that NanoControl significantly reduces computational overhead compared to conventional control approaches, while maintaining superior generation quality and achieving improved controllability. |
| format | Preprint |
| id |
arxiv_https___arxiv_org_abs_2508_10424 |
| institution | arXiv |
| publishDate | 2025 |
| record_format | arxiv |
| spellingShingle | NanoControl: A Lightweight Framework for Precise and Efficient Control in Diffusion Transformer Liu, Shanyuan Zhu, Jian Lu, Junda Gong, Yue Li, Liuzhuozheng Cheng, Bo Ma, Yuhang Wu, Liebucha Wu, Xiaoyu Leng, Dawei Yin, Yuhui Computer Vision and Pattern Recognition Diffusion Transformers (DiTs) have demonstrated exceptional capabilities in text-to-image synthesis. However, in the domain of controllable text-to-image generation using DiTs, most existing methods still rely on the ControlNet paradigm originally designed for UNet-based diffusion models. This paradigm introduces significant parameter overhead and increased computational costs. To address these challenges, we propose the Nano Control Diffusion Transformer (NanoControl), which employs Flux as the backbone network. Our model achieves state-of-the-art controllable text-to-image generation performance while incurring only a 0.024\% increase in parameter count and a 0.029\% increase in GFLOPs, thus enabling highly efficient controllable generation. Specifically, rather than duplicating the DiT backbone for control, we design a LoRA-style (low-rank adaptation) control module that directly learns control signals from raw conditioning inputs. Furthermore, we introduce a KV-Context Augmentation mechanism that integrates condition-specific key-value information into the backbone in a simple yet highly effective manner, facilitating deep fusion of conditional features. Extensive benchmark experiments demonstrate that NanoControl significantly reduces computational overhead compared to conventional control approaches, while maintaining superior generation quality and achieving improved controllability. |
| title | NanoControl: A Lightweight Framework for Precise and Efficient Control in Diffusion Transformer |
| topic | Computer Vision and Pattern Recognition |
| url | https://arxiv.org/abs/2508.10424 |