README.md

---
license: apache-2.0
datasets:
- openbmb/RLAIF-V-Dataset
language:
- en
paper:
---

# Model Card for RLAIF-V

[GitHub ](https://github.com/RLHF-V/RLAIF-V) | [Paper](https://arxiv.org/abs/2405.17220)

**RLAIF-V-12B** is a multimodal large language model (MLLM) that exhibits **super GPT-4V trustworthiness**. The model is built up on OmniLMM from the [MiniCPM-V](https://github.com/OpenBMB/MiniCPM-V) series. 

We utilize a novel framework, [RLAIF-V](https://github.com/RLHF-V/RLAIF-V), which **aligns MLLMs in a fully open-source paradigm**. This framework maximally exploits the [open-source feedback](https://huggingface.co/datasets/HaoyeZhang/RLAIF-V-Dataset) from two key perspectives, including **high-quality feedback data** and an **online feedback learning algorithm**. 

<p align="center">
  <img src="https://cdn-uploads.huggingface.co/production/uploads/6566e0c493e30c8a60048eb3/T4hALrgNdXKHnkvb-27bA.png" alt="fig1" width="85%"/>
</p>

## Model Details

### Key Features

* 🏅 **Super GPT-4V Trustworthiness**: By learning from open-source AI feedback, RLAIF-V-12B achieves super GPT-4V trustworthiness in both generative and discriminative tasks.
* 💪 **Maintaining Well Performance on General Abilities**: On benchmarks tested with the general abilities (e.g. LLaVA Bench, MMStar), RLAIF-V-12B also exhibits good performance.

<p align="center">
  <img src="https://cdn-uploads.huggingface.co/production/uploads/6566e0c493e30c8a60048eb3/ypXZxb4HE-jDPJU9115bi.png" alt="fig1" width="90%"/>
</p>
<!-- ![image/png](https://cdn-uploads.huggingface.co/production/uploads/6566e0c493e30c8a60048eb3/ypXZxb4HE-jDPJU9115bi.png) -->

### Examples
<p align="center">
  <img src="https://cdn-uploads.huggingface.co/production/uploads/6566e0c493e30c8a60048eb3/yg-Ksp9qi8AodURSmX769.png" alt="fig2-1" width="81%"/>
  <img src="https://cdn-uploads.huggingface.co/production/uploads/6566e0c493e30c8a60048eb3/NSEkeBmH99B44rX8GTZig.png" alt="fig2-1" width="80%"/>
</p>

### Model Description
- **Related model:** [OmniLMM-12B](https://huggingface.co/openbmb/OmniLMM-12B)
- **Trained on data:** [RLAIF-V-Dataset](https://huggingface.co/datasets/HaoyeZhang/RLAIF-V-Dataset)

## Usage
Please look at [GitHub](https://github.com/RLHF-V/RLAIF-V) for more details about usage.



## Citation

If you find our model/code/paper helpful, please consider cite our papers 📝:

```bibtex
@article{yu2023rlhf,
  title={Rlhf-v: Towards trustworthy mllms via behavior alignment from fine-grained correctional human feedback},
  author={Yu, Tianyu and Yao, Yuan and Zhang, Haoye and He, Taiwen and Han, Yifeng and Cui, Ganqu and Hu, Jinyi and Liu, Zhiyuan and Zheng, Hai-Tao and Sun, Maosong and others},
  journal={arXiv preprint arXiv:2312.00849},
  year={2023}
}

@article{yu2024rlaifv,
  title={RLAIF-V: Aligning MLLMs through Open-Source AI Feedback for Super GPT-4V Trustworthiness}, 
  author={Yu, Tianyu and Zhang, Haoye and Yao, Yuan and Dang, Yunkai and Chen, Da and Lu, Xiaoman and Cui, Ganqu and He, Taiwen and Liu, Zhiyuan and Chua, Tat-Seng and Sun, Maosong},
  journal={arXiv preprint arXiv:2405.17220},
  year={2024},
}
```