Saved in:
| Main Authors: | , , , , , , , , , , , |
|---|---|
| Format: | Preprint |
| Published: |
2025
|
| Subjects: | |
| Online Access: | https://arxiv.org/abs/2503.10615 |
| Tags: |
Add Tag
No Tags, Be the first to tag this record!
|
| _version_ | 1866915202842230784 |
|---|---|
| author | Yang, Yi He, Xiaoxuan Pan, Hongkun Jiang, Xiyan Deng, Yan Yang, Xingtao Lu, Haoyu Yin, Dacheng Rao, Fengyun Zhu, Minfeng Zhang, Bo Chen, Wei |
| author_facet | Yang, Yi He, Xiaoxuan Pan, Hongkun Jiang, Xiyan Deng, Yan Yang, Xingtao Lu, Haoyu Yin, Dacheng Rao, Fengyun Zhu, Minfeng Zhang, Bo Chen, Wei |
| contents | Large Language Models have demonstrated remarkable reasoning capability in complex textual tasks. However, multimodal reasoning, which requires integrating visual and textual information, remains a significant challenge. Existing visual-language models often struggle to effectively analyze and reason visual content, resulting in suboptimal performance on complex reasoning tasks. Moreover, the absence of comprehensive benchmarks hinders the accurate assessment of multimodal reasoning capabilities. In this paper, we introduce R1-Onevision, a multimodal reasoning model designed to bridge the gap between visual perception and deep reasoning. To achieve this, we propose a cross-modal reasoning pipeline that transforms images into formal textural representations, enabling precise language-based reasoning. Leveraging this pipeline, we construct the R1-Onevision dataset which provides detailed, step-by-step multimodal reasoning annotations across diverse domains. We further develop the R1-Onevision model through supervised fine-tuning and reinforcement learning to cultivate advanced reasoning and robust generalization abilities. To comprehensively evaluate multimodal reasoning performance across different grades, we introduce R1-Onevision-Bench, a benchmark aligned with human educational stages, covering exams from junior high school to university and beyond. Experimental results show that R1-Onevision achieves state-of-the-art performance, outperforming models such as GPT-4o and Qwen2.5-VL on multiple challenging multimodal reasoning benchmarks. |
| format | Preprint |
| id |
arxiv_https___arxiv_org_abs_2503_10615 |
| institution | arXiv |
| publishDate | 2025 |
| record_format | arxiv |
| spellingShingle | R1-Onevision: Advancing Generalized Multimodal Reasoning through Cross-Modal Formalization Yang, Yi He, Xiaoxuan Pan, Hongkun Jiang, Xiyan Deng, Yan Yang, Xingtao Lu, Haoyu Yin, Dacheng Rao, Fengyun Zhu, Minfeng Zhang, Bo Chen, Wei Computer Vision and Pattern Recognition Large Language Models have demonstrated remarkable reasoning capability in complex textual tasks. However, multimodal reasoning, which requires integrating visual and textual information, remains a significant challenge. Existing visual-language models often struggle to effectively analyze and reason visual content, resulting in suboptimal performance on complex reasoning tasks. Moreover, the absence of comprehensive benchmarks hinders the accurate assessment of multimodal reasoning capabilities. In this paper, we introduce R1-Onevision, a multimodal reasoning model designed to bridge the gap between visual perception and deep reasoning. To achieve this, we propose a cross-modal reasoning pipeline that transforms images into formal textural representations, enabling precise language-based reasoning. Leveraging this pipeline, we construct the R1-Onevision dataset which provides detailed, step-by-step multimodal reasoning annotations across diverse domains. We further develop the R1-Onevision model through supervised fine-tuning and reinforcement learning to cultivate advanced reasoning and robust generalization abilities. To comprehensively evaluate multimodal reasoning performance across different grades, we introduce R1-Onevision-Bench, a benchmark aligned with human educational stages, covering exams from junior high school to university and beyond. Experimental results show that R1-Onevision achieves state-of-the-art performance, outperforming models such as GPT-4o and Qwen2.5-VL on multiple challenging multimodal reasoning benchmarks. |
| title | R1-Onevision: Advancing Generalized Multimodal Reasoning through Cross-Modal Formalization |
| topic | Computer Vision and Pattern Recognition |
| url | https://arxiv.org/abs/2503.10615 |