From: The performance of BERT as data representation of text clustering
Dataset | Description | Number of class | Total number of data |
---|---|---|---|
AG News | Contains news titles and content from AG News media categorized by news topics | 4 | 4000 |
Yahoo! Answers | Contains questions asked on Yahoo! Answers along with their answers which are categorized based on the topic of the question | 10 | 10000 |
Reuters | Contains documents extracted from Reuters-21578, which is data containing news documents from the Reuters mass media in 1987 | 2 | 5859 |