README.md

---
language:
- en
---
# Adapting Multimodal Large Language Models to Domains via Post-Training

This repository provides an implementation preview of our paper: [On Domain-Specific Post-Training for Multimodal Large Language Models](https://huggingface.co/papers/2411.19930).

Our code will be available at [https://github.com/bigai-ai/QA-Synthesizer](https://github.com/bigai-ai/QA-Synthesizer)


We investigate domain adaptation of MLLMs through post-training, focusing on data synthesis, training pipelines, and task evaluation. 
**(1) Data Synthesis**: Using open-source models, we develop a visual instruction synthesizer that effectively generates diverse visual instruction tasks from domain-specific image-caption pairs. **Our synthetic tasks surpass those generated by manual rules, GPT-4, and GPT-4V in enhancing the domain-specific performance of MLLMs.** 
**(2) Training Pipeline**: While the two-stage training--initially on image-caption pairs followed by visual instruction tasks--is commonly adopted for developing general MLLMs, we apply a single-stage training pipeline to enhance task diversity for domain-specific post-training. 
**(3) Task Evaluation**: We conduct experiments in two domains, biomedicine and food, by post-training MLLMs of different sources and scales (e.g., Qwen2-VL-2B, LLaVA-v1.6-8B, Llama-3.2-11B), and then evaluating MLLM performance on various domain-specific tasks.

<p align='left'>
    <img src="https://cdn-uploads.huggingface.co/production/uploads/650801ced5578ef7e20b33d4/-Jp7pAsCR2Tj4WwfwsbCo.png" width="600">
</p>


<p align='left'>
    <img src="https://cdn-uploads.huggingface.co/production/uploads/650801ced5578ef7e20b33d4/bRu85CWwP9129bSCRzos2.png" width="1000">
</p>


***************** **Updates** ********************  
- [2024/12/11] Released [food visual instructions](https://huggingface.co/datasets/AdaptLLM/food-visual-instructions) for post-training MLLMs
- [2024/12/10] Released evaluation benchmark datasets for biomedicine and food domains: [biomed-VQA-benchmark](https://huggingface.co/datasets/AdaptLLM/biomed-VQA-benchmark), [food-VQA-benchmark](https://huggingface.co/datasets/AdaptLLM/food-VQA-benchmark).
- [2024/12/9] Released AdaMLLM developed from llava-next-llama3-8b: [AdaMLLM-med-8B](https://huggingface.co/AdaptLLM/biomed-LLaVA-NeXT-Llama3-8B), [AdaMLLM-food-8B](https://huggingface.co/AdaptLLM/food-LLaVA-NeXT-Llama3-8B).
- [2024/12/7] Released [visual-instruction-synthesizer](https://huggingface.co/AdaptLLM/visual-instruction-synthesizer) used to synthesize task triplets based on image-caption pairs.
- [2024/12/6] Released AdaMLLM developed from Qwen2-VL-2B and Llama-3.2-11B-Vision: [AdaMLLM-med-2B](https://huggingface.co/AdaptLLM/biomed-Qwen2-VL-2B-Instruct), [AdaMLLM-food-2B](https://huggingface.co/AdaptLLM/food-Qwen2-VL-2B-Instruct), [AdaMLLM-med-11B](https://huggingface.co/AdaptLLM/biomed-Llama-3.2-11B-Vision-Instruct),  [AdaMLLM-food-11B](https://huggingface.co/AdaptLLM/food-Llama-3.2-11B-Vision-Instruct), 
- [2024/12/05] Released [biomedicine visual instructions](https://huggingface.co/datasets/AdaptLLM/biomed-visual-instructions) for post-training MLLMs
- [2024/11/29] Released our paper


## Resources
**🤗 We share our data and models with example usages, feel free to open any issues or discussions! 🤗**

| Model                                                                       | Repo ID in HF 🤗                           | Domain       | Base Model              | Training Data                                                                                  | Evaluation Benchmark |
|:----------------------------------------------------------------------------|:--------------------------------------------|:--------------|:-------------------------|:------------------------------------------------------------------------------------------------|-----------------------|
| [Visual Instruction Synthesizer](https://huggingface.co/AdaptLLM/visual-instruction-synthesizer) | AdaptLLM/visual-instruction-synthesizer     | -  | open-llava-next-llama3-8b    | TBD | -                   |
| [AdaMLLM-med-2B](https://huggingface.co/AdaptLLM/biomed-Qwen2-VL-2B-Instruct) | AdaptLLM/biomed-Qwen2-VL-2B-Instruct     | Biomedicine  | Qwen2-VL-2B-Instruct    | [biomed-visual-instructions](https://huggingface.co/datasets/AdaptLLM/biomed-visual-instructions) | [biomed-VQA-benchmark](https://huggingface.co/datasets/AdaptLLM/biomed-VQA-benchmark)                   |
| [AdaMLLM-food-2B](https://huggingface.co/AdaptLLM/food-Qwen2-VL-2B-Instruct) | AdaptLLM/food-Qwen2-VL-2B-Instruct     | Food  | Qwen2-VL-2B-Instruct    | [food-visual-instructions](https://huggingface.co/datasets/AdaptLLM/food-visual-instructions) | [food-VQA-benchmark](https://huggingface.co/datasets/AdaptLLM/food-VQA-benchmark)                   |
| [AdaMLLM-med-8B](https://huggingface.co/AdaptLLM/biomed-LLaVA-NeXT-Llama3-8B) | AdaptLLM/biomed-LLaVA-NeXT-Llama3-8B     | Biomedicine  | open-llava-next-llama3-8b    | [biomed-visual-instructions](https://huggingface.co/datasets/AdaptLLM/biomed-visual-instructions) | [biomed-VQA-benchmark](https://huggingface.co/datasets/AdaptLLM/biomed-VQA-benchmark)                   |
| [AdaMLLM-food-8B](https://huggingface.co/AdaptLLM/food-LLaVA-NeXT-Llama3-8B) |AdaptLLM/food-LLaVA-NeXT-Llama3-8B     | Food  | open-llava-next-llama3-8b    | [food-visual-instructions](https://huggingface.co/datasets/AdaptLLM/food-visual-instructions) |  [food-VQA-benchmark](https://huggingface.co/datasets/AdaptLLM/food-VQA-benchmark)                   |
| [AdaMLLM-med-11B](https://huggingface.co/AdaptLLM/biomed-Llama-3.2-11B-Vision-Instruct) | AdaptLLM/biomed-Llama-3.2-11B-Vision-Instruct     | Biomedicine  | Llama-3.2-11B-Vision-Instruct    | [biomed-visual-instructions](https://huggingface.co/datasets/AdaptLLM/biomed-visual-instructions) | [biomed-VQA-benchmark](https://huggingface.co/datasets/AdaptLLM/biomed-VQA-benchmark)                   |
| [AdaMLLM-food-11B](https://huggingface.co/AdaptLLM/food-Llama-3.2-11B-Vision-Instruct) | AdaptLLM/food-Llama-3.2-11B-Vision-Instruct     | Food | Llama-3.2-11B-Vision-Instruct    | [food-visual-instructions](https://huggingface.co/datasets/AdaptLLM/food-visual-instructions) |  [food-VQA-benchmark](https://huggingface.co/datasets/AdaptLLM/food-VQA-benchmark)                   |


## Contact  
Daixuan Cheng: `[email protected]`

## About

AdaMLLM represents our latest advancement in building domain-specific foundation models through post-training on synthetic supervised tasks derived from unsupervised contexts.

<p align='left'>
    <img src="https://cdn-uploads.huggingface.co/production/uploads/650801ced5578ef7e20b33d4/2aPl6mKIyHeQp8SO4TXAk.png" width="700">
</p>


- **[AdaptLLM](https://huggingface.co/papers/2309.09530): Adapt LLM to domains**  
  We employ rule-based methods to extract tasks from domain-specific corpora, reformatting them into reading comprehension tasks for continued pre-training. Our 7B finance model outperforms domain-specific models of much larger scales, such as BloombergGPT-50B.

- **[AdaMLLM](https://huggingface.co/papers/2411.19930): Adapt Multimodal LLM to domains**  
  We extend supervised task synthesis to multimodality, introducing a unified visual instruction synthesizer to extract instruction-response pairs from domain-specific image-caption pairs. Our synthetic tasks outperform those generated by manual rules, GPT-4, and GPT-4V in improving domain-specific performance for MLLMs.



## Citation
If you find our work helpful, please cite us.

[AdaMLLM](https://huggingface.co/papers/2411.19930)
```bibtex
@article{adamllm,
  title={On Domain-Specific Post-Training for Multimodal Large Language Models},
  author={Cheng, Daixuan and Huang, Shaohan and Zhu, Ziyu and Zhang, Xintong and Zhao, Wayne Xin and Luan, Zhongzhi and Dai, Bo and Zhang, Zhenliang},
  journal={arXiv preprint arXiv:2411.19930},
  year={2024}
}
```

[AdaptLLM](https://huggingface.co/papers/2309.09530) (ICLR 2024)
```bibtex
@inproceedings{
adaptllm,
title={Adapting Large Language Models via Reading Comprehension},
author={Daixuan Cheng and Shaohan Huang and Furu Wei},
booktitle={The Twelfth International Conference on Learning Representations},
year={2024},
url={https://openreview.net/forum?id=y886UXPEZ0}
}
```