IndoSRL is a semantic role labeling datasets in Indonesian. This dataset provides complete predicate-argument structure for the existing predicates in a sentence. This dataset provides SRL annotations on top of Indonesian GSD corpus from Universal dependencies. This dataset can be used for four tasks: Predicate identification, predicate sense disambiguation, argument identification and argument classification. This data is available in conllup format.
NusaCatalogue: https://indonlp.github.io/nusa-catalogue/card.html?indosrl