nexuswho/gliner-onnx-small

GLiNER ONNX - Small

Pre-converted ONNX model for GLiNER zero-shot NER.

Model Details

• Base Model: urchade/gliner_small-v2.1
• Format: ONNX with INT8 quantization
• Use Case: Fast CPU inference for entity extraction

Usage

from gliner import GLiNER

model = GLiNER.from_pretrained(
    "nexuswho/gliner-onnx-small",
    load_onnx_model=True,
    onnx_model_file="model_quantized.onnx",
)

entities = model.predict_entities(
    "AWS Lambda integrates with Amazon S3",
    ["cloud service", "storage service"],
)

Files

• model_quantized.onnx - INT8 quantized ONNX model (recommended)
• model.onnx - Full precision ONNX model
• config.json, tokenizer.json, etc. - Model configuration

License

Apache 2.0 (same as base GLiNER model)