metadata
tags:
- gliner2
- ner
- dataset-extraction
- lora
base_model: fastino/gliner2-base-v1
library_name: gliner2
GLiNER2 Dataset Mention Extractor
Fine-tuned GLiNER2 model for extracting structured dataset mentions from research documents.
Task
Given a document passage, extracts:
- Entity fields: dataset_name, acronym, producer, geography, description, etc.
- Classifications: dataset_tag (named/descriptive/vague), usage_context, is_used
Training
- Base model:
fastino/gliner2-base-v1 - Method: LoRA (r=16, alpha=32)
- Data: 1,197 synthetic training examples
Usage
from gliner2 import GLiNER2
extractor = GLiNER2.from_pretrained("fastino/gliner2-base-v1")
extractor.load_adapter("rafmacalaba/gliner2-datause-v2")
schema = (
extractor.create_schema()
.structure("dataset_mention")
.field("dataset_name", dtype="str")
.field("acronym", dtype="str")
.field("producer", dtype="str")
.field("geography", dtype="str")
.field("dataset_tag", dtype="str", choices=["named", "descriptive", "vague"])
.field("usage_context", dtype="str", choices=["primary", "supporting", "background"])
.field("is_used", dtype="str", choices=["True", "False"])
)
results = extractor.extract(text, schema)
dataset_mentions = results["dataset_mention"]