Scientific multimodal instruction tuning with large language and vision models. - View it on GitHub