markuplm
Datasets
All datasets matching “markuplm”markuplm_scimetadata
Dataset Card for Scientific Works Metadata Extraction
📖 Description
Dataset for training models (including MarkupLM) on the task of metadata extraction from scientific papers in HTML format. Contains BIO tagging for 27 metadata types.
Task: Token Classification for entity extraction:
['title', 'author', 'date', 'doi', 'issn', 'eissn', 'journal', 'publisher', 'pages', 'first_page', 'last_page', 'language', 'volume', 'issue', 'abstract', 'affiliation', 'keyword'… See the full description on the dataset page: https://huggingface.co/datasets/endthesame/markuplm_scimetadata.markuplm-toy-datasetmarkuplm-propertiesmarkuplm-toy-dataset