WIT
收藏极市2025-05-30 更新2025-05-31 收录
下载链接:
https://www.cvmart.net/dataSets/detail/1253
下载链接
链接失效反馈官方服务:
资源简介:
基于 Wikipedia 的图像文本 (WIT) 数据集是一个大型多模态数据集多语言数据集
应用场景:
提供机构:
极市相关数据集
wit
# Dataset Card for WIT ## Table of Contents - [Table of Contents](#table-of-contents) - [Dataset Description](#dataset-description) - [Dataset Summary](#dataset-summary) - [Dataset Preprocessing
魔搭社区2026-01-06 更新300
Witness
BACKGROUND The artwork is an archival digital print, titled Witness. Set in a panoramic landscape of the Central Victorian goldfields region in Australian, the image includes an Indian man dressed in
Research Data Australia100
WITS Dataset
This dataset is an extension of MASAC, a multimodal, multi-party, Hindi-English code-mixed dialogue dataset compiled from the popular Indian TV show, ‘Sarabhai v/s Sarabhai’. WITS was created by augm
paperswithcode.com60
WITS
该数据集名为WITS,包含了来自印度电视剧的多模态、多方参与、语码混合的讽刺性对话。每个讽刺实例都附有相应的自然语言语码混合解释。文本输入采用罗马字编码混合格式,包含了讽刺、幽默和情绪的标注。此外,该数据集还经过了增强,创建了如sWITS、hWITS和eWITS等变体,以适应不同的任务需求,如讽刺解释、讽刺检测、幽默识别和情绪识别。
arXiv130



