Datasets:
metadata
license: openrail
task_categories:
- feature-extraction
- text2text-generation
language:
- en
size_categories:
- 10K<n<100K
Title and Keywords Dataset
The titles are taken from Medium articles dataset , and the keywords are extracted by our team at 🤖 https://exnrt.com