The GeoChat Instruct dataset is a multimodal instruction tracking dataset designed for remote sensing. It was developed by a research team at Mohamed bin Zayed University of AI in the paper “GeoChat: Grounded Large Vision-Language Model for Remote Sensing》, which was published in CVPR2024. The main content of the paper is to introduce the construction and application of the GeoChat model and the GeoChat_Instruct dataset. This dataset was generated using Vicuna-v1.5 and an automated pipeline, contains nearly 318k instructions, and is designed to extend multimodal instruction adaptation to the remote sensing domain for training multi-task conversational assistants. The GeoChat_Instruct dataset was created to address the lack of multimodal instruction adaptation dialogue datasets in the remote sensing domain, and is used to fine-tune a remote sensing domain visual language model called GeoChat.

Citation

@misc{kuckreja2023geochat, title={GeoChat: Grounded Large Vision-Language Model for Remote Sensing}, author={Kartik Kuckreja and Muhammad Sohail Danish and Muzammal Naseer and Abhijit Das and Salman Khan and Fahad Shahbaz Khan}, year={2023}, eprint={2311.15826}, archivePrefix={arXiv}, primaryClass={cs.CV} }