Juncheng Li^1,2*, Kaihang Pan¹*, Zhiqi Ge¹*, Minghe Gao¹*, Hanwang Zhang³, Wei Ji², Wenqiao Zhang², Tat-Seng Chua², Siliang Tang^1†, Yueting Zhuang^1†

¹Zhejiang University, ²National University of Singapore, ³Nanyang Technological University

*Equal Contribution. ^†Corresponding Authors

To facilitate research in interleaved vision-language instruction following, we build I4 (semantically Interconnected, Interleaved Image-Text Instruction-Following), an extensive large-scale benchmark of 31 tasks with diverse instructions in a unified instruction-response format, covering 20 diverse scenarios.

I4 has three important properties:

• Interleaved vision-language context: all the instructions contain sequences of inter-related images and texts, such as storyboards with scripts, textbooks with diagrams.
• Diverse forms of complex instructions: the instructions range from predicting dialogue for comics, to discovering differences between surveillance images, and to conversational embodied tasks.
• Vast range of instruction-following scenarios: the benchmark covers multiple application scenarios, including cartoons, industrial images, driving recording, etc.

Cheetor is a Transformer-based multi-modal large language model empowered by controllable knowledge re-injection, which can effectively handle a wide variety of interleaved vision-language instructions.

Cheetor demonstrates strong abilities to perform reasoning over complicated interleaved vision-language instructions. For instance, in (a), Cheetor is able to keenly identify the connections between the images and thereby infer the reason that causes this unusual phenomenon. In (b, c), Cheetor can reasonably infer the relations among the images and understand the metaphorical implications they want to convey. In (e, f), Cheetor exhibits the ability to comprehend absurd objects through multi-modal conversations with humans.

1. Installation

Git clone our repository and creating conda environment:

git clone https://github.com/DCDmllm/Cheetah.git
cd Cheetah/Cheetah
conda create -n cheetah python=3.8
conda activate cheetah
pip install -r requirement.txt

2. Prepare Vicuna Weights and Llama2 weights

The current version of Cheetor supports Vicuna-7B and LLaMA2-7B as the language model. Please first follow the instructions to prepare Vicuna-v0 7B weights and follow the instructions to prepare LLaMA-2-Chat 7B weights.

Then modify the llama_model in the Cheetah/cheetah/configs/models/cheetah_vicuna.yaml to the folder that contains Vicuna weights and modify the llama_model in the Cheetah/cheetah/configs/models/cheetah_llama2.yaml to the folder that contains LLaMA2 weights.

3. Prepare the pretrained checkpoint for Cheetor

Download the pretrained checkpoints of Cheetah according to the language model you prepare:

For the checkpoint aligned with Vicuna 7B, please set the path to the pretrained checkpoint in the evaluation config file in Cheetah/eval_configs/cheetah_eval_vicuna.yaml at Line 10.

For the checkpoint aligned with LLaMA2 7B, please set the path to the pretrained checkpoint in the evaluation config file in Cheetah/eval_configs/cheetah_eval_llama2.yaml at Line 10.

Besides, Cheetor reuses the pretrained Q-former from BLIP-2 that matches FlanT5-XXL.

4. How to use Cheetor

Examples of using our Cheetah model are provided in files Cheetah/test_cheetah_vicuna.py and Cheetah/test_cheetah_llama2.py. You can test your own samples following the format shown in these two files. And you can run the test code in the following way (taking the Vicuna version of Cheetah as an example):

python test_cheetah_vicuna.py --cfg-path eval_configs/cheetah_eval_vicuna.yaml --gpu-id 0

And in the near future, we will also demonstrate how to launch the gradio demo of Cheetor locally.

If you found this work useful, please consider giving this repository a star and citing our paper as followed:

@misc{li2023empowering,
      title={Empowering Vision-Language Models to Follow Interleaved Vision-Language Instructions}, 
      author={Juncheng Li and Kaihang Pan and Zhiqi Ge and Minghe Gao and Hanwang Zhang and Wei Ji and Wenqiao Zhang and Tat-Seng Chua and Siliang Tang and Yueting Zhuang},
      year={2023},
      eprint={2308.04152},
      archivePrefix={arXiv},
      primaryClass={cs.CV}
}