GLiNER4j ONNX — GLiNER2 Base

ONNX export of fastino/gliner2-base-v1 for Java inference via ONNX Runtime.

Part of the GLiNER4j project.

Supported Tasks

Task	Description
Named Entity Recognition	Extract typed entity spans from text with confidence scores
Text Classification	Assign labels to text with multi-label support and confidence scores

Both tasks support entity/label descriptions for improved accuracy and per-call overrides without model reloading.

Repository Structure

├── gliner4j_config.json        # Shared model configuration
├── tokenizer.json              # Shared HuggingFace tokenizer
├── tokenizer_config.json
├── onnx/                       # Base FP32 (~830 MB)
│   ├── ner_full.onnx
│   └── classifier_full.onnx
├── onnx_fp16/                  # FP16 (~416 MB, ~50% smaller)
│   ├── ner_full.onnx
│   └── classifier_full.onnx
└── onnx_quantized/             # INT8 dynamic quantization (~208 MB, ~75% smaller)
    ├── ner_full.onnx
    └── classifier_full.onnx

An onnx_optimized_cpu/ folder with the same two files may also be present (ONNX Runtime graph-optimized for CPU).

Model Architecture

Each variant ships two merged, self-contained ONNX graphs — one per task:

Graph	Description
`ner_full.onnx`	Transformer encoder + span representation + count-aware scoring head (NER)
`classifier_full.onnx`	Transformer encoder + classifier head MLP (Classification)

The graphs are fused at export time from the encoder and task heads; the intermediate split modules are not published.

Variants

Variant	Folder	Precision	Size	Use case
Base	`onnx/`	FP32	~830 MB	Maximum accuracy
FP16	`onnx_fp16/`	FP16	~416 MB, ~50% smaller	Good accuracy/size trade-off
Quantized	`onnx_quantized/`	INT8 (QUInt8, per-channel)	~208 MB, ~75% smaller	Smallest footprint, fastest on CPU

To download a specific variant only:

huggingface-cli download <repo> --include "onnx_fp16/*" "*.json"

Configuration

Parameter	Value
Hidden size	768
Max span width	8
Max count	20
Span mode	SpanMarkerV0
Token pooling	first
ONNX opset	17

Usage

Use with GLiNER4j, a Java library for GLiNER2 inference via ONNX Runtime.

Named Entity Recognition

var entities = List.of(
    new EntityDefinition("person", "Names of individuals"),
    new EntityDefinition("organization", "Company or institution names")
);
var gliner = GLiNER4jNER.load(modelDir, entities);
Map<String, List<EntitySpan>> results = gliner.extract("John works at Google.");

Text Classification

var labels = List.of(
    new ClassificationLabel("positive", "Expresses positive sentiment"),
    new ClassificationLabel("negative", "Expresses negative sentiment")
);
var classifier = GLiNER4jClassifier.load(modelDir, labels);
List<ClassificationResult> results = classifier.classify("Great product!");

Model Variants

// FP16 variant
var gliner = GLiNER4jNER.load(modelDir, entities, "onnx_fp16");

// Quantized variant
var gliner = GLiNER4jNER.load(modelDir, entities, "onnx_quantized");

License

Apache License 2.0