测试训练>> nuswide>> 返回
项目作者： wenting-zhao

项目描述 ：
Images and labels for NUSWIDE (unofficial).
高级语言： Python
项目地址: git://github.com/wenting-zhao/nuswide.git
创建时间： 2020-11-18T22:55:17Z
项目社区：https://github.com/wenting-zhao/nuswide
开源协议：MIT License
下载


nuswide

This repo provides images and labels for NUSWIDE (unofficially). Since the release of NUSWIDE, many original flickr images have gone down. I have collected as many images as possible, and hope this will benefit the community. Any contribution is welcomed! For now, it is only for multi-label classification. NUSWIDE, however, can also be used for image tagging.

A pytorch interface for NUSWIDE classification dataloader is also provided. Used as follows:

  1. >>> import nuswide
  2. >>> dataset = nuswide.NUSWIDEClassification('.', 'trainval')
  3. [dataset] read ./classification_labels/classification_trainval.csv
  4. [dataset] NUSWIDE classification set=trainval number of classes=81  number of images=134025
  5. >>> dataset = nuswide.NUSWIDEClassification('.', 'test')
  6. [dataset] read ./classification_labels/classification_test.csv
  7. [dataset] NUSWIDE classification set=test number of classes=81  number of images=89470

I downloaded images from https://github.com/thuml/HashNet/tree/master/pytorch. I processed the labels from https://lms.comp.nus.edu.sg/wp-content/uploads/2019/research/nuswide/NUS-WIDE.html. Many thanks to them!