treamyracle/

distilled-indonesian-ner

This dataset is a high-quality synthetic Indonesian Named Entity Recognition (NER) dataset, distilled and self-corrected using a 120B parameter teacher model. It features annotations across 19 distinct entity classes formatted as character-level spans for token classification tasks.

	Dataset Details





	


	Dataset Structure




	


	Data Instances

An example instance in the dataset: { "text":… See the full description on the dataset page: https://huggingface.co/datasets/treamyracle/distilled-indonesian-ner.