schema-driven

GLiNER2 is a unified transformer-based framework for multiple NLP tasks, offering improved efficiency and accessibility over large language models by supporting named entity recognition, text classification, and hierarchical data extraction in a single model.