v-mllm-13b

v-MLLM Model Card

Model details

Model type: v-MLLM is an open-source MLLM trained on Visual-Modality Instruction (VIM) corpus, it can robustly follow the text-modality instructions and visual-modality instructions.

Model date: v-MLLM-13B was trained in January 2024.

Github for more information: https://github.com/VIM-Bench/VIM_TOOL

License

v-MLLM is licensed under the LLAMA 2 Community License, Copyright (c) Meta Platforms, Inc. All Rights Reserved.

Intended use

Primary intended uses: The primary use of v-MLLM is for research on multimodal large language models.

Primary intended users: The primary intended users of the model are researchers in computer vision, natural language processing, machine learning, and artificial intelligence.

Training dataset

• 846k VIM corpus based on LVIS-Instruct4V corpus.

Citation

Please kindly cite our paper if you find our resources useful:

@misc{li2024text,
      title={Text as Images: Can Multimodal Large Language Models Follow Printed Instructions in Pixels?}, 
      author={Xiujun Li and Yujie Lu and Zhe Gan and Jianfeng Gao and William Yang Wang and Yejin Choi},
      year={2024},
      eprint={2311.17647},
      archivePrefix={arXiv},
      primaryClass={cs.CV}
}
@misc{lu2023vim,
      title={VIM: Probing Multimodal Large Language Models for Visual Embedded Instruction Following}, 
      author={Yujie Lu and Xiujun Li and William Yang Wang and Yejin Choi},
      year={2023},
      eprint={2311.17647},
      archivePrefix={arXiv},
      primaryClass={cs.CV}
}