WIT by Google AI
美国
其他数据集

WIT by Google AI 翻译站点

基于Wikipedia的图像文本数据集,用于多模式多语言机器学习

标签:
爱站权重:PC 百度权重移动 百度移动权重

WIT(基于Wikipedia的图像文本)数据集是一个大型多模式的多语言数据集,其中包含37m+ Image-Text Sets在100多种语言上具有11M+唯一图像。

#### 动机
多模式粘性语言模型依靠丰富的数据集来帮助他们学习建模图像和文本之间的关系。如最近的工作所示,拥有大型图像文本数据集可以显着提高性能。此外,在现有数据集中缺乏语言覆盖范围(主要是英语)也阻碍了多语言多模式空间的研究 - Google AI认为这是一个丢失的机会,因为在利用图像(作为语言敏捷的媒介)中显示的潜力(作为一种语言媒介)来帮助提高我们的多语言文本理解。

为了解决这些挑战并提高多语言的多模式学习的研究,Google AI创建了基于Wikipedia的图像文本(WIT)数据集。 WIT是通过从Wikipedia文章和Wikimedia图像链接中提取与图像(例如,如上图所示)相关的多个不同文本(例如,如上图所示)创建的。伴随着严格的过滤,仅保留高质量的图像文本集。

The resulting dataset contains over 37.6 million image-text sets – making WIT the largest multimodal dataset (publicly available at the time of this writing) with unparalleled multilingual coverage – with 12K+ examples in each of 108 languages (53 languages have 100K+ image-text pairs )。

sljf

原文:

WIT (Wikipedia-based Image Text) Dataset is a large multimodal multilingual dataset comprising 37M+ image-text sets with 11M+ unique images across 100+ languages.

#### Motivation
Multimodal visio-linguistic models rely on a rich dataset to help them learn to model the relationship between images and texts. Having large image-text datasets can significantly improve performance, as shown by recent works. Furthermore the lack of language coverage in existing datasets (which are mostly only in English) also impedes research in the multilingual multimodal space – Google AI consider this a lost opportunity given the potential shown in leveraging images (as a language-agnostic medium) to help improve our multilingual textual understanding.

To address these challenges and advance research on multilingual, multimodal learning Google AI created the Wikipedia-based Image Text (WIT) Dataset. WIT is created by extracting multiple different texts associated with an image (e.g., as shown in the above image) from Wikipedia articles and Wikimedia image links. This was accompanied by rigorous filtering to only retain high quality image-text sets.

The resulting dataset contains over 37.6 million image-text sets – making WIT the largest multimodal dataset (publicly available at the time of this writing) with unparalleled multilingual coverage – with 12K+ examples in each of 108 languages (53 languages have 100K+ image-text pairs).

Sljf

数据统计

数据评估

WIT by Google AI浏览人数已经达到284,如你需要查询该站的相关权重信息,可以点击"爱站数据""Chinaz数据"进入;以目前的网站数据参考,建议大家请以爱站数据为准,更多网站价值评估因素如:WIT by Google AI的访问速度、搜索引擎收录以及索引量、用户体验等;当然要评估一个站的价值,最主要还是需要根据您自身的需求以及需要,一些确切的数据则需要找WIT by Google AI的站长进行洽谈提供。如该站的IP、PV、跳出率等!

关于WIT by Google AI特别声明

本站GPT 案例导航提供的WIT by Google AI都来源于网络,不保证外部链接的准确性和完整性,同时,对于该外部链接的指向,不由GPT 案例导航实际控制,在2023年3月9日 下午10:20收录时,该网页上的内容,都属于合规合法,后期网页的内容如出现违规,可以直接联系网站管理员进行删除,GPT 案例导航不承担任何责任。

相关导航