From the model card
Fields below are copied from the tags and counters on the HuggingFace repository fastino/gliner2-base-v1 at our last fetch. They are set by the uploader, not verified by us; rows with no tag are omitted. How this page is made.
- Publisher (HF namespace)
- fastino
- Weight formats
- safetensors
- License tag
apache-2.0— read the license file in the repo before relying on it- Language tags
- English (en)
- Papers cited
- arXiv:2507.18546
- Downloads (HF counter at last fetch)
- 396,307
- Likes (HF counter at last fetch)
- 99
- Model card
- https://huggingface.co/fastino/gliner2-base-v1
Use cases
- Zero-shot NER for entity types not seen during training
- Extracting structured JSON from unstructured product descriptions
- Intent and sentiment classification from customer messages
- Relation extraction between entities in short texts
- Rapid prototyping of information extraction pipelines without task-specific fine-tuning
Pros
- Single model handles NER, relation extraction, intent, and sentiment
- Zero-shot entity types via natural language type definitions at inference time
- 86 likes with active community use across diverse extraction scenarios
- Lightweight enough for CPU deployment in moderate-throughput pipelines
Cons
- GLiNER2 custom code is not in standard HuggingFace transformers; requires gliner2 package
- Base size underperforms fine-tuned task-specific models on narrow domains
- No explicit license declared; verify before production use
- Complex nested entity structures and long-range relations remain difficult
Tags
gliner2safetensorsextractorText classificationNamed Entity RecognitionRelation ExtractionIntent classificationSentiment AnalysisTopic classificationStructured extractionJson extractionenarxiv:2507.18546license:apache-2.0region:us