kgnlp/meld-open

MELD Open MELD is a multilingual and multi-domain dataset for Named Entity Recognition (NER) constructed from 60 existing datasets. It includes gold-standard annotations across 60 languages and 14 domains. This dataset is a subset of 43 datasets for which licenses permit the redistribution of data in a new format. See the MELD GitHub repository for more details. Note: This version of MELD Open retains the original labels from its source datasets. For normalized labels, use… See the full description on the dataset page: https://huggingface.co/datasets/kgnlp/meld-open.