---
title: "OmniDiagram — paper by Yang et al. (2026)"
description: "Most current systems that turn diagrams into code handle only one task or one output language. 34 references, cited by 1 paper here."
authors: ["Haoyue Yang", "Xuanle Zhao", "Xuexin Liu", "Feibing Jiang", "Yao Zhu"]
year: 2026
venue: "Findings of the Association for Computational Linguistics: ACL 2026"
source: "https://aclanthology.org/2026.findings-acl.809/"
doi: "10.18653/v1/2026.findings-acl.809"
licence: "CC BY 4.0 (ACL Anthology)"
licence_url: "https://creativecommons.org/licenses/by/4.0/"
summary_licence: "Elamur Cosmic License"
canonical: "https://research.elamur.ai/free/en/papers/yang-2026-omnidiagram-advancing-unified-diagram-code/"
---
[Elamur Research](https://md.research.elamur.ai/free/en/) · [this page in HTML](https://research.elamur.ai/free/en/papers/yang-2026-omnidiagram-advancing-unified-diagram-code/)

[home](https://md.research.elamur.ai/free/en/) / [papers](https://md.research.elamur.ai/free/en/papers/) / [2026](https://md.research.elamur.ai/free/en/papers/2026/page/15/) / OmniDiagram: Advancing Unified Diagram Code Generation via Visual Interrogation Reward

# OmniDiagram: Advancing Unified Diagram Code Generation via Visual Interrogation Reward

**Authors:** Haoyue Yang, [Xuanle Zhao](https://md.research.elamur.ai/free/en/authors/xuanle-zhao/), Xuexin Liu, Feibing Jiang and Yao Zhu · **Year:** 2026 · **Venue:** [Findings of the Association for Computational Linguistics: ACL 2026](https://md.research.elamur.ai/free/en/venues/findings/) ([Findings of the Association for Computational Linguistics 2026](https://md.research.elamur.ai/free/en/venues/findings/2026/)) · **Publisher:** [Association for Computational Linguistics](https://md.research.elamur.ai/free/en/publishers/association-for-computational-linguistics/) · **Licence:** [CC BY 4.0 (ACL Anthology)](https://creativecommons.org/licenses/by/4.0/)

**Links:** [Publisher's page](https://aclanthology.org/2026.findings-acl.809/) · [PDF](https://aclanthology.org/2026.findings-acl.809.pdf) · [DOI 10.18653/v1/2026.findings-acl.809](https://doi.org/10.18653/v1/2026.findings-acl.809) · [OpenAlex](https://openalex.org/W7166848653)

**Citations:** 0 citations (OpenAlex, 2026-10-14) · [1](https://md.research.elamur.ai/free/en/papers/yang-2026-omnidiagram-advancing-unified-diagram-code/cited-by/) in our library, of the 149,439 papers whose bibliographies we read

## Summary

Most current systems that turn diagrams into code handle only one task or one output language. That limits their use across the many kinds of structured diagrams. OmniDiagram is a single multimodal model that covers three tasks: Diagram-to-Code, Diagram Editing and Text-to-Code. It produces LaTeX, Mermaid and PlantUML.

The authors first built M3²Diagram, a 196k-sample corpus arranged as a 3 × 3 grid of tasks and languages. They synthesised it top-down with Gemini-2.5-Flash, moving from topic to scenario to structured data to code–image pairs. Execution-based error correction and visual checks then filtered the output.

Training runs in two stages. SFT comes first, then GRPO with a new reward called VIVA. Before training, VIVA prepares questions about each target diagram. During RL, the generated code is rendered and a VLM answers those questions about the rendered image. This gives fine-grained feedback on visual fidelity without needing reference code.

OmniDiagram reports state-of-the-art results on M3²Bench and on the Mermaid subset of VisPlotBench. This suggests that question-based visual rewards are a scalable way to align code with what actually appears in the rendered output.

## Key points

- **M3²Diagram corpus:** 31k open-source samples plus 165k synthetic samples, kept from about 300k candidates. Each language covers roughly 15 diagram types, with a focus on flowchart variants.
- **M3²Bench:** a held-out evaluation set of 1.7k samples, each checked by human experts. The split uses clustering based on perceptual hashing.
- **Training ablations (3B model):**
  - Adding RL after SFT raises Diagram-to-Code execution rate from 88.6% to 93.0%.
  - It raises Text-to-Code execution rate from 86.5% to 90.3%.
  - RL without SFT collapses into producing only Mermaid, with only 30.2% Diagram-to-Code execution.
- **Reasoning-trajectory ablation (3B model):**
  - Reasoning data helps Diagram Editing.
  - It hurts the other two tasks.
- **Reward-model choice:** VIVA works almost equally well with Qwen3-VL-32B or Qwen3-VL-30B-A3B as the judge. This suggests the pre-generated questions matter more than the size of the verifier.

*Our summary of the paper, in our own words — not its text. The work belongs to its authors; read it at the links above.*

## Institutions

- [Zhejiang University](https://md.research.elamur.ai/free/en/universities/zhejiang-university/) — University, Hangzhou, China · 172 works · 12,988 citations
- [Chinese Academy of Sciences](https://md.research.elamur.ai/free/en/government/chinese-academy-of-sciences/) — Government body, Beijing, China · 194 works · 23,906 citations
- [Institute of Automation](https://md.research.elamur.ai/free/en/labs/institute-of-automation/) — Research institute, Beijing, China · 70 works · 14,251 citations
- [University of Chinese Academy of Sciences](https://md.research.elamur.ai/free/en/universities/university-of-chinese-academy-of-sciences/) — University, Beijing, China · 147 works · 8,480 citations

📚 Shelves (3): [🖥️ Computer Science](https://md.research.elamur.ai/free/en/shelves/computer-science/) · 16,667 works · [👁️ Computer Vision and Pattern Recognition](https://md.research.elamur.ai/free/en/shelves/computer-science/computer-vision-and-pattern-recognition/) · 2,869 works · [🗄️ Information Systems](https://md.research.elamur.ai/free/en/shelves/computer-science/information-systems/) · 1,437 works

🧭 Topics (1): [💻 Code & programming languages](https://md.research.elamur.ai/free/en/topics/code-and-programming/) · 129 works

## Cite

Cite as: Haoyue Yang, Xuanle Zhao, Xuexin Liu, Feibing Jiang and Yao Zhu. 2026. OmniDiagram: Advancing Unified Diagram Code Generation via Visual Interrogation Reward. Findings of the Association for Computational Linguistics: ACL 2026.

The paper itself, not this page. BibTeX from the library's record of the work.

```bibtex
@inproceedings{yang2026omnidiagram,
  title = {{OmniDiagram: Advancing Unified Diagram Code Generation via Visual Interrogation Reward}},
  author = {Haoyue Yang, Xuanle Zhao, Xuexin Liu, Feibing Jiang and Yao Zhu},
  booktitle = {Findings of the Association for Computational Linguistics: ACL 2026},
  year = {2026},
  doi = {10.18653/v1/2026.findings-acl.809},
  url = {https://doi.org/10.18653/v1/2026.findings-acl.809}
}
```

- [← previous Omni-Embed-Audio: Leveraging Multimodal LLMs for Robust Audio-Text Retrieval](https://md.research.elamur.ai/free/en/papers/yoo-2026-omni-embed-audio-leveraging-multimodal/)
- [next → On Safety Risks in Experience-Driven Self-Evolving Agents](https://md.research.elamur.ai/free/en/papers/zhao-2026-safety-risks-experience-driven-self/)
- [↑ back to papers of 2026](https://md.research.elamur.ai/free/en/papers/2026/page/15/)

## References

34 references, 9 of them with a card of ours. ↗ leads to the work, «find» searches for one we have no id for yet.

1. Shuai Bai, Yuxuan Cai, Ruizhe Chen, Keqin Chen, Xionghui Chen, Zesen Cheng, Lianghao Deng, Wei Ding, Chang Gao, Chunjiang Ge, Wenbin Ge, Zhifang Guo, Qidong Huang, Jie Huang, Fei Huang, Binyuan Hui, Shutong Jiang, Zhaohai Li, Mingsheng Li, and 45 others. 2025a. [Qwen3-vl technical report](https://md.research.elamur.ai/free/en/papers/bai-2025-qwen3-vl-technical-report/). Preprint, a…
2. Shuai Bai, Keqin Chen, Xuejing Liu, Jialin Wang, Wenbin Ge, Sibo Song, Kai Dang, Peng Wang, Shijie Wang, Jun Tang, and 1 others. 2025b. [Qwen2.5-vl technical report](https://md.research.elamur.ai/free/en/papers/bai-2025-qwen2-5-vl-technical/). arXiv preprint arXiv:2502.13923.
3. Patrice Bechard, Chao Wang, Amirhossein Abaskohi, Juan Rodriguez, Christopher Pal, David Vazquez, Spandana Gella, Sai Rajeswar, and Perouz Taslakian. 2025. Starflow: Generating structured workflow outputs from sketch images. arXiv preprint arXiv:2503.21889. [find](https://scholar.google.com/scholar?q=Patrice+Bechard%2C+Chao+Wang%2C+Amirhossein+Abaskohi%2C+Juan+Rodriguez%2C+Christopher+Pal%2C+David+Vazquez%2C+Spandana+Gella%2C+Sai+Rajeswar%2C+and+Perouz+Taslakian%2E+2025%2E+Starflow%3A+Generating+structured+workflow+out)
4. Lei Chen, Xuanle Zhao, Zhixiong Zeng, Jing Huang, Liming Zheng, Yufeng Zhong, and Lin Ma. 2025a. Breaking the sft plateau: Multimodal structured reinforcement learning for chart-to-code generation. arXiv preprint arXiv:2508.13587. — [arXiv:2508.13587](https://arxiv.org/abs/2508.13587) [find](https://scholar.google.com/scholar?q=Lei+Chen%2C+Xuanle+Zhao%2C+Zhixiong+Zeng%2C+Jing+Huang%2C+Liming+Zheng%2C+Yufeng+Zhong%2C+and+Lin+Ma%2E+2025a%2E+Breaking+the+sft+plateau%3A+Multimodal+structured+reinforcement+learning+for+chart%2Dto%2Dcode+generation%2E+ar)
5. Liangyu Chen, Yichen Xu, Jianzhe Ma, Yuqi Liu, Donglu Yang, Liang Zhang, Wenxuan Wang, and Qin Jin. 2025b. Charteditor: A reinforcement learning framework for robust chart editing. arXiv preprint arXiv:2511.15266. — [arXiv:2511.15266](https://arxiv.org/abs/2511.15266) [find](https://scholar.google.com/scholar?q=Liangyu+Chen%2C+Yichen+Xu%2C+Jianzhe+Ma%2C+Yuqi+Liu%2C+Donglu+Yang%2C+Liang+Zhang%2C+Wenxuan+Wang%2C+and+Qin+Jin%2E+2025b%2E+Charteditor%3A+A+reinforcement+learning+framework+for+robust+chart+editing%2E+arXiv+preprint+arXi)
6. Aryaz Eghbali and Michael Pradel. 2022. Crystalbleu: precisely and efficiently measuring the similarity of code. In Proceedings of the 37th IEEE/ACM International Conference on Automated Software Engineering, pages 1–12. [find](https://scholar.google.com/scholar?q=Aryaz+Eghbali+and+Michael+Pradel%2E+2022%2E+Crystalbleu%3A+precisely+and+efficiently+measuring+the+similarity+of+code%2E+In+Proceedings+of+the+37th+IEEE%2FACM+International+Conference+on+Automated+Software+Engi)
7. John H Flavell. 1979. Metacognition and cognitive monitoring: A new area of cognitive–developmental inquiry. American psychologist, 34(10):906. [find](https://scholar.google.com/scholar?q=John+H+Flavell%2E+1979%2E+Metacognition+and+cognitive+monitoring%3A+A+new+area+of+cognitive%E2%80%93developmental+inquiry%2E+American+psychologist%2C+34%2810%29%3A906%2E)
8. Google DeepMind. 2025. Gemini 3 flash. [https://deepmind.google/models/gemini/flash/](https://deepmind.google/models/gemini/flash/). Accessed: 2025-12-28. [↗](https://deepmind.google/models/gemini/flash/)
9. Daya Guo, Dejian Yang, Haowei Zhang, Junxiao Song, Ruoyu Zhang, Runxin Xu, Qihao Zhu, Shirong Ma, Peiyi Wang, Xiao Bi, and 1 others. 2025. [Deepseek-r1: Incentivizing reasoning capability in llms via reinforcement learning](https://md.research.elamur.ai/free/en/papers/deepseek-ai-2025-deepseek-r1-incentivizing-reasoning-capability/). arXiv preprint arXiv:2501.12948.
10. Mengliang He, Jiayi Zeng, Yankai Jiang, Wei Zhang, Zeming Liu, Xiaoming Shi, and Aimin Zhou. 2025. Flow2code: Evaluating large language models for flowchart-based code generation capability. arXiv preprint arXiv:2506.02073. [find](https://scholar.google.com/scholar?q=Mengliang+He%2C+Jiayi+Zeng%2C+Yankai+Jiang%2C+Wei+Zhang%2C+Zeming+Liu%2C+Xiaoming+Shi%2C+and+Aimin+Zhou%2E+2025%2E+Flow2code%3A+Evaluating+large+language+models+for+flowchart%2Dbased+code+generation+capability%2E+arXiv+pre)
11. Akriti Jain, Pritika Ramu, Aparna Garimella, and Apoorv Saxena. 2025. Doc2chart: Intent-driven zero-shot chart generation from documents. In Proceedings of the 2025 Conference on Empirical Methods in Natural Language Processing, pages 34936–34951. [find](https://scholar.google.com/scholar?q=Akriti+Jain%2C+Pritika+Ramu%2C+Aparna+Garimella%2C+and+Apoorv+Saxena%2E+2025%2E+Doc2chart%3A+Intent%2Ddriven+zero%2Dshot+chart+generation+from+documents%2E+In+Proceedings+of+the+2025+Conference+on+Empirical+Methods+in+)
12. Xin Lai, Zhuotao Tian, Yukang Chen, Senqiao Yang, Xiangru Peng, and Jiaya Jia. 2024. [Step-dpo: Step-wise preference optimization for long-chain reasoning of llms](https://md.research.elamur.ai/free/en/papers/lai-2024-step-dpo-step-wise-preference/). arXiv preprint arXiv:2406.18629.
13. Harrison Lee, Samrat Phatale, Hassan Mansoor, Kellie Ren Lu, Thomas Mesnard, Johan Ferret, Colton Bishop, Ethan Hall, Victor Carbune, and Abhinav Rastogi. 2023. Rlaif: Scaling reinforcement learning from human feedback with ai feedback. [find](https://scholar.google.com/scholar?q=Harrison+Lee%2C+Samrat+Phatale%2C+Hassan+Mansoor%2C+Kellie+Ren+Lu%2C+Thomas+Mesnard%2C+Johan+Ferret%2C+Colton+Bishop%2C+Ethan+Hall%2C+Victor+Carbune%2C+and+Abhinav+Rastogi%2E+2023%2E+Rlaif%3A+Scaling+reinforcement+learning+f)
14. Fang Liu, Yang Liu, Lin Shi, Houkun Huang, Ruifeng Wang, Zhen Yang, Li Zhang, Zhongqi Li, and Yuchi Ma. 2024. Exploring and evaluating hallucinations in llm-powered code generation. arXiv preprint arXiv:2404.00971. — [arXiv:2404.00971](https://arxiv.org/abs/2404.00971) [find](https://scholar.google.com/scholar?q=Fang+Liu%2C+Yang+Liu%2C+Lin+Shi%2C+Houkun+Huang%2C+Ruifeng+Wang%2C+Zhen+Yang%2C+Li+Zhang%2C+Zhongqi+Li%2C+and+Yuchi+Ma%2E+2024%2E+Exploring+and+evaluating+hallucinations+in+llm%2Dpowered+code+generation%2E+arXiv+preprint+arX)
15. Tianqi Luo, Chuhan Huang, Leixian Shen, Boyan Li, Shuyu Shen, Wei Zeng, Nan Tang, and Yuyu Luo. 2025. nvbench 2.0: Resolving ambiguity in text-to-visualization through stepwise reasoning. arXiv preprint arXiv:2503.12880. — [arXiv:2503.12880](https://arxiv.org/abs/2503.12880) [find](https://scholar.google.com/scholar?q=Tianqi+Luo%2C+Chuhan+Huang%2C+Leixian+Shen%2C+Boyan+Li%2C+Shuyu+Shen%2C+Wei+Zeng%2C+Nan+Tang%2C+and+Yuyu+Luo%2E+2025%2E+nvbench+2%2E0%3A+Resolving+ambiguity+in+text%2Dto%2Dvisualization+through+stepwise+reasoning%2E+arXiv+prepri)
16. Yuansheng Ni, Songcheng Cai, Xiangchao Chen, Jiarong Liang, Zhiheng Lyu, Jiaqi Deng, Kai Zou, Ping Nie, Fei Yuan, Xiang Yue, and 1 others. 2025. Viscoder2: Building multi-language visualization coding agents. arXiv preprint arXiv:2510.23642. — [arXiv:2510.23642](https://arxiv.org/abs/2510.23642) [find](https://scholar.google.com/scholar?q=Yuansheng+Ni%2C+Songcheng+Cai%2C+Xiangchao+Chen%2C+Jiarong+Liang%2C+Zhiheng+Lyu%2C+Jiaqi+Deng%2C+Kai+Zou%2C+Ping+Nie%2C+Fei+Yuan%2C+Xiang+Yue%2C+and+1+others%2E+2025%2E+Viscoder2%3A+Building+multi%2Dlanguage+visualization+coding)
17. OpenAI. 2025. Introducing GPT-4.1 in the API. [https://openai.com/index/gpt-4-1/](https://openai.com/index/gpt-4-1/). Accessed: 2025-12-28. [↗](https://openai.com/index/gpt-4-1/)
18. Mizanur Rahman, Md Tahmid Rahman Laskar, Shafiq Joty, and Enamul Hoque. 2025. Text2vis: A challenging and diverse benchmark for generating multimodal visualizations from text. In Proceedings of the 2025 Conference on Empirical Methods in Natural Language Processing, pages 31837–31862. [find](https://scholar.google.com/scholar?q=Mizanur+Rahman%2C+Md+Tahmid+Rahman+Laskar%2C+Shafiq+Joty%2C+and+Enamul+Hoque%2E+2025%2E+Text2vis%3A+A+challenging+and+diverse+benchmark+for+generating+multimodal+visualizations+from+text%2E+In+Proceedings+of+the+20)
19. Juan A Rodriguez, Haotian Zhang, Abhay Puri, Aarash Feizi, Rishav Pramanik, Pascal Wichmann, Arnab Mondal, Mohammad Reza Samsami, Rabiul Awal, Perouz Taslakian, and 1 others. 2025. Rendering-aware reinforcement learning for vector graphics generation. arXiv preprint arXiv:2505.20793. — [arXiv:2505.20793](https://arxiv.org/abs/2505.20793) [find](https://scholar.google.com/scholar?q=Juan+A+Rodriguez%2C+Haotian+Zhang%2C+Abhay+Puri%2C+Aarash+Feizi%2C+Rishav+Pramanik%2C+Pascal+Wichmann%2C+Arnab+Mondal%2C+Mohammad+Reza+Samsami%2C+Rabiul+Awal%2C+Perouz+Taslakian%2C+and+1+others%2E+2025%2E+Rendering%2Daware+rei)
20. Zhihong Shao, Peiyi Wang, Qihao Zhu, Runxin Xu, Junxiao Song, Xiao Bi, Haowei Zhang, Mingchuan Zhang, YK Li, Yang Wu, and 1 others. 2024. [Deepseekmath: Pushing the limits of mathematical reasoning in open language models](https://md.research.elamur.ai/free/en/papers/shao-2024-deepseekmath-pushing-limits-mathematical-reasoning/). arXiv preprint arXiv:2402.03300.
21. Qiushi Sun, Jingyang Gong, Yang Liu, Qiaosheng Chen, Lei Li, Kai Chen, Qipeng Guo, Ben Kao, and Fei Yuan. 2025. Januscoder: Towards a foundational visual-programmatic interface for code intelligence. arXiv preprint arXiv:2510.23538. — [arXiv:2510.23538](https://arxiv.org/abs/2510.23538) [find](https://scholar.google.com/scholar?q=Qiushi+Sun%2C+Jingyang+Gong%2C+Yang+Liu%2C+Qiaosheng+Chen%2C+Lei+Li%2C+Kai+Chen%2C+Qipeng+Guo%2C+Ben+Kao%2C+and+Fei+Yuan%2E+2025%2E+Januscoder%3A+Towards+a+foundational+visual%2Dprogrammatic+interface+for+code+intelligence%2E+)
22. Wentao Tan, Qiong Cao, Chao Xue, Yibing Zhan, Changxing Ding, and Xiaodong He. 2025. Chartmaster: Advancing chart-to-code generation with real-world charts and chart similarity reinforcement learning. arXiv preprint arXiv:2508.17608. — [arXiv:2508.17608](https://arxiv.org/abs/2508.17608) [find](https://scholar.google.com/scholar?q=Wentao+Tan%2C+Qiong+Cao%2C+Chao+Xue%2C+Yibing+Zhan%2C+Changxing+Ding%2C+and+Xiaodong+He%2E+2025%2E+Chartmaster%3A+Advancing+chart%2Dto%2Dcode+generation+with+real%2Dworld+charts+and+chart+similarity+reinforcement+learning%2E)
23. Weiyun Wang, Zhangwei Gao, Lixin Gu, Hengjun Pu, Long Cui, Xingguang Wei, Zhaoyang Liu, Linglin Jing, Shenglong Ye, Jie Shao, and 1 others. 2025. [Internvl3](https://md.research.elamur.ai/free/en/papers/wang-2025-internvl3-5-advancing-open/). 5: Advancing open-source multimodal models in versatility, reasoning, and efficiency. arXiv preprint arXiv:2508.18265.
24. ZENG Xingchen, Zhewei Su, Hengming Zhang, Juyong Jiang, Jiazhi Xia, and Wei Zeng. Davinci: Reinforcing visual-structural syntax in mllms for generalized scientific diagram parsing. In The Fourteenth International Conference on Learning Representations. Cheng Yang, Chufan Shi, Yaxin Liu, Bo Shui, Jun… — [arXiv:2406.09961](https://arxiv.org/abs/2406.09961) [find](https://scholar.google.com/scholar?q=ZENG+Xingchen%2C+Zhewei+Su%2C+Hengming+Zhang%2C+Juyong+Jiang%2C+Jiazhi+Xia%2C+and+Wei+Zeng%2E+Davinci%3A+Reinforcing+visual%2Dstructural+syntax+in+mllms+for+generalized+scientific+diagram+parsing%2E+In+The+Fourteenth+I)
25. Yiying Yang, Wei Cheng, Sijin Chen, Xianfang Zeng, Fukun Yin, Jiaxu Zhang, Liao Wang, Gang Yu, Xingjun Ma, and Yu-Gang Jiang. 2025a. Omnisvg: A unified scalable vector graphics generation model. In The Thirty-ninth Annual Conference on Neural Information Processing Systems. [find](https://scholar.google.com/scholar?q=Yiying+Yang%2C+Wei+Cheng%2C+Sijin+Chen%2C+Xianfang+Zeng%2C+Fukun+Yin%2C+Jiaxu+Zhang%2C+Liao+Wang%2C+Gang+Yu%2C+Xingjun+Ma%2C+and+Yu%2DGang+Jiang%2E+2025a%2E+Omnisvg%3A+A+unified+scalable+vector+graphics+generation+model%2E+In+Th)
26. Yue Yang, Ajay Patel, Matt Deitke, Tanmay Gupta, Luca Weihs, Andrew Head, Mark Yatskar, Chris Callison-Burch, Ranjay Krishna, Aniruddha Kembhavi, and 1 others. 2025b. Scaling text-rich image understanding via code-guided synthetic multimodal data generation. arXiv preprint arXiv:2502.14846. [find](https://scholar.google.com/scholar?q=Yue+Yang%2C+Ajay+Patel%2C+Matt+Deitke%2C+Tanmay+Gupta%2C+Luca+Weihs%2C+Andrew+Head%2C+Mark+Yatskar%2C+Chris+Callison%2DBurch%2C+Ranjay+Krishna%2C+Aniruddha+Kembhavi%2C+and+1+others%2E+2025b%2E+Scaling+text%2Drich+image+understan)
27. Sukmin Yun, Rusiru Thushara, Mohammad Bhat, Yongxin Wang, Mingkai Deng, Jinhong Wang, Tianhua Tao, Junbo Li, Haonan Li, Preslav Nakov, and 1 others. 2024. Web2code: A large-scale webpage-to-code dataset and evaluation framework for multimodal llms. Advances in neural information processing systems,… [find](https://scholar.google.com/scholar?q=Sukmin+Yun%2C+Rusiru+Thushara%2C+Mohammad+Bhat%2C+Yongxin+Wang%2C+Mingkai+Deng%2C+Jinhong+Wang%2C+Tianhua+Tao%2C+Junbo+Li%2C+Haonan+Li%2C+Preslav+Nakov%2C+and+1+others%2E+2024%2E+Web2code%3A+A+large%2Dscale+webpage%2Dto%2Dcode+datas)
28. Guangming Sheng, Chi Zhang, Zilingfeng Ye, Xibin Wu, Wang Zhang, Ru Zhang, Yanghua Peng, Haibin Lin, and Chuan Wu. 2024. [Hybridflow: A flexible and efficient rlhf framework](https://md.research.elamur.ai/free/en/papers/sheng-2024-hybridflow-flexible-efficient-rlhf-framework/). arXiv preprint arXiv: 2409.19256.
29. Xuanle Zhao, Deyang Jiang, Zhixiong Zeng, Lei Chen, Haibo Qiu, Jing Huang, Yufeng Zhong, Liming Zheng, Yilin Cao, and Lin Ma. 2025a. Vincicoder: Unifying multimodal code generation via coarse-to-fine visual reinforcement learning. arXiv preprint arXiv:2511.00391. — [arXiv:2511.00391](https://arxiv.org/abs/2511.00391) [find](https://scholar.google.com/scholar?q=Xuanle+Zhao%2C+Deyang+Jiang%2C+Zhixiong+Zeng%2C+Lei+Chen%2C+Haibo+Qiu%2C+Jing+Huang%2C+Yufeng+Zhong%2C+Liming+Zheng%2C+Yilin+Cao%2C+and+Lin+Ma%2E+2025a%2E+Vincicoder%3A+Unifying+multimodal+code+generation+via+coarse%2Dto%2Dfine+)
30. Shubhankar Singh, Purvi Chaurasia, Yerram Varun, Pranshu Pandya, Vatsal Gupta, Vivek Gupta, and Dan Roth. 2024. Flowvqa: Mapping multimodal logic in visual question answering with flowcharts. arXiv preprint arXiv:2406.19237. [find](https://scholar.google.com/scholar?q=Shubhankar+Singh%2C+Purvi+Chaurasia%2C+Yerram+Varun%2C+Pranshu+Pandya%2C+Vatsal+Gupta%2C+Vivek+Gupta%2C+and+Dan+Roth%2E+2024%2E+Flowvqa%3A+Mapping+multimodal+logic+in+visual+question+answering+with+flowcharts%2E+arXiv+pr)
31. Xuanle Zhao, Xuexin Liu, Haoyue Yang, Xianzhen Luo, Fanhu Zeng, Jianling Li, Qi Shi, and Chi Chen. 2025b. Chartedit: How far are mllms from automating chart analysis? evaluating mllms’ capability via chart editing. arXiv preprint arXiv:2505.11935. [find](https://scholar.google.com/scholar?q=Xuanle+Zhao%2C+Xuexin+Liu%2C+Haoyue+Yang%2C+Xianzhen+Luo%2C+Fanhu+Zeng%2C+Jianling+Li%2C+Qi+Shi%2C+and+Chi+Chen%2E+2025b%2E+Chartedit%3A+How+far+are+mllms+from+automating+chart+analysis%3F+evaluating+mllms%E2%80%99+capability+via+)
32. Xuanle Zhao, Xianzhen Luo, Qi Shi, Chi Chen, Shuo Wang, Zhiyuan Liu, and Maosong Sun. 2025c. [Chartcoder: Advancing multimodal large language model for chart-to-code generation](https://md.research.elamur.ai/free/en/papers/zhao-2025-chartcoder-advancing-multimodal-large-language/). arXiv preprint arXiv:2501.06598.
33. Yuze Zhao, Jintao Huang, Jinghan Hu, Xingjun Wang, Yunlin Mao, Daoze Zhang, Zeyinzi Jiang, Zhikai Wu, Baole Ai, Ang Wang, and 1 others. 2025d. Swift: a scalable lightweight infrastructure for fine-tuning. In Proceedings of the AAAI Conference on Artificial Intelligence, volume 39, pages 29733–29735. [find](https://scholar.google.com/scholar?q=Yuze+Zhao%2C+Jintao+Huang%2C+Jinghan+Hu%2C+Xingjun+Wang%2C+Yunlin+Mao%2C+Daoze+Zhang%2C+Zeyinzi+Jiang%2C+Zhikai+Wu%2C+Baole+Ai%2C+Ang+Wang%2C+and+1+others%2E+2025d%2E+Swift%3A+a+scalable+lightweight+infrastructure+for+fine%2Dtun)
34. Jinguo Zhu, Weiyun Wang, Zhe Chen, Zhaoyang Liu, Shenglong Ye, Lixin Gu, Hao Tian, Yuchen Duan, Weijie Su, Jie Shao, and 1 others. 2025. [Internvl3: Exploring advanced training and test-time recipes for open-source multimodal models](https://md.research.elamur.ai/free/en/papers/zhu-2025-internvl3-exploring-advanced-training/). arXiv preprint arXiv:2504.10479.

*Updated 31 October 2026*