AIMClab/ChinaOpen
收藏Hugging Face2023-11-15 更新2024-03-04 收录
下载链接：
https://hf-mirror.com/datasets/AIMClab/ChinaOpen
下载链接
链接失效反馈官方服务：
资源简介：
---
license: cc-by-nc-sa-4.0
language:
- zh
size_categories:
- 1K<n<10K
---
## Dataset Description
- **Homepage:** [ChinaOpen homepage](https://ruc-aimc-lab.github.io/ChinaOpen/)
- **Paper:** [ChinaOpen: A Dataset for Open-World Multimodal Learning](https://doi.org/10.1145/3581783.3612156)
- **Point of Contact:** [Aozhu Chen](caz@ruc.edu.cn)
### Dataset Summary
ChinaOpen-1k is a dataset sourced from Bilibili, a popular Chinese video-sharing website. It is a manually annotated test set of videos, including manually checked user titles/tags, manually written captions, and manual labels describing the visual objects/actions/scenes shown in the content.
### Languages
Chinese and English
## Dataset Structure
All the files are put in a zip package.
```bash
├── ChinaOpen-1k
├── video01.mp4
├── video02.mp4
├── video03.mp4
├── [...]
└── ChinaOpen-1k-annotations.json
```
### Data Instances
Please refer to https://ruc-aimc-lab.github.io/ChinaOpen/#examples
提供机构：
AIMClab
原始信息汇总
数据集描述
数据集概述
ChinaOpen-1k 是一个从中国流行的视频分享网站 Bilibili 获取的数据集。它是一个手动标注的视频测试集，包括手动检查的用户标题/标签、手动编写的字幕以及描述视频内容中显示的视觉对象/动作/场景的手动标签。
语言
中文和英文
数据集结构
所有文件都放在一个压缩包中。 bash ├── ChinaOpen-1k ├── video01.mp4 ├── video02.mp4 ├── video03.mp4 ├── [...] └── ChinaOpen-1k-annotations.json
