Data generation pipeline

[CleyLyChen]

Hi, great work.
I want to follow your work, can you share me the prompts of the data generate pipeline with me, it would be really help.
Best wishes.

[xxlllz]

Thanks for your attention.

Overall, the prompts we use to generate data are based on  Multimodal Self-Instruct  repo https://github.com/zwq2018/Multi-modal-Self-instruct. Most prompts are available in their repository, but no chart-related content exists. We have made modifications to it and uploaded part of our prompts (the remaining part is still being organized) in our repo data_generator. You can use the Multimodal Instruct repo and the code we provided to generate relevant data.

By the way, we will open-source Chart2Code-160k as soon as possible.