Florence
收藏arXiv2025-09-30 收录
下载链接：
https://www.micc.unifi.it/resources/datasets/florence-3d-faces/
下载链接
链接失效反馈官方服务：
资源简介：
该数据集包含了9亿个图像-文本对，旨在提升对象级别的表征能力。它已经在各种基准测试中取得了最新的领先成果，其规模达到了9亿对，任务定位于视觉-语言预训练。
This dataset contains 900 million image-text pairs, aiming to enhance object-level representation capabilities. It has achieved state-of-the-art performance across various benchmark tests, with a scale of 900 million pairs and is targeted at visual-language pre-training.
