GLiNER2

GLiNER2

Unified schema-based information extraction

Description

Extracting entities, labels and structured fields from text means chaining models or paying for LLM calls. GLiNER2 is a lightweight schema-based information extraction model handling many tasks in one call.

Define a schema and run NER, classification and structured extraction quickly on CPU.

Features



Unified:Entities and fields.

Classification:Same pass.

Schemas:Custom fields.

CPU:Lightweight.