LLaVA-1.5 instruction fine-tuning data
收藏arXiv2025-09-30 收录
下载链接：
https://llava-prumerge.github.io/
下载链接
链接失效反馈官方服务：
资源简介：
该数据集用于对LLaVA-1.5模型进行微调，以减少视觉标记，它是指导性微调数据的一部分。此外，该数据集还用于对LLaVA-1.5进行微调，并评估多样化的视觉问答和推理基准测试任务。
This dataset is utilized to fine-tune the LLaVA-1.5 model for reducing visual tokens, and it is part of the instruction fine-tuning data. Furthermore, this dataset is also used for fine-tuning LLaVA-1.5 and evaluating diverse visual question answering and reasoning benchmark tasks.
