This dataset is part of the NounAtlas project, aiming to enhance Nominal Semantic Role Labeling (SRL) by providing a comprehensive inventory of nominal predicates organized into semantically-coherent frames.
The NounAtlas SRL Corpus contains sentences annotated with nominal predicates and their corresponding semantic roles. This dataset is split into three subsets: training, development, and test.
Train: 22,452 sentences
Dev: 2,806… See the full description on the dataset page:
https://huggingface.co/datasets/sapienzanlp/nounatlas_srl_corpus.
