Junteng/Chart_CLIP
Image Feature Extraction
•
Updated
Error code: JobManagerCrashedError
Need help to make the dataset viewer work? Make sure to review how to configure the dataset viewer, and open a discussion for direct support.
This repository contains the CLIP Training data implementation from our paper "On the Perception Bottleneck of VLMs for Chart Understanding".
Code: https://github.com/hkust-nlp/Vision4Chart
If you find this data useful in your research, please consider citing our paper:
@misc{liu2025perceptionbottleneckvlmschart,
title={On the Perception Bottleneck of VLMs for Chart Understanding},
author={Junteng Liu and Weihao Zeng and Xiwen Zhang and Yijun Wang and Zifei Shan and Junxian He},
year={2025},
eprint={2503.18435},
archivePrefix={arXiv},
primaryClass={cs.CV},
url={https://arxiv.org/abs/2503.18435},
}