Multi-Modal CelebA-HQ
收藏OpenDataLab2026-05-24 更新2024-05-09 收录
下载链接:
https://opendatalab.org.cn/OpenDataLab/Multi-Modal_CelebA-HQ
下载链接
链接失效反馈官方服务:
资源简介:
多模式-CelebA-HQ是一个大规模的人脸图像数据集,具有30,000通过遵循CelebA-HQ从CelebA数据集选择的高分辨率人脸图像。每个图像都具有高质量的分割蒙版,草图,描述性文本和具有透明背景的图像。
多模式-CelebA-HQ可用于训练和评估文本到图像生成,文本引导图像处理,草图到图像生成,图像标题和VQA的算法。该数据集是在TediGAN中提出并使用的。
Multi-Modal CelebA-HQ is a large-scale face image dataset consisting of 30,000 high-resolution face images selected from the CelebA dataset in accordance with the CelebA-HQ construction standards. Each sample is paired with high-quality segmentation masks, sketches, descriptive text captions, and images with transparent backgrounds.
Multi-Modal CelebA-HQ can be used to train and evaluate algorithms for tasks including text-to-image generation, text-guided image processing, sketch-to-image generation, image captioning, and VQA. This dataset was first proposed and utilized in TediGAN.
提供机构:
OpenDataLab
创建时间:
2022-11-02
搜集汇总
数据集介绍

以上内容由遇见数据集搜集并总结生成



