Instructions to use nicolasembleton/gliner2.5-multi-v1-onnx with libraries, inference providers, notebooks, and local apps. Follow these links to get started.
- Libraries
- GLiNER2
How to use nicolasembleton/gliner2.5-multi-v1-onnx with GLiNER2:
from gliner2 import GLiNER2 model = GLiNER2.from_pretrained("nicolasembleton/gliner2.5-multi-v1-onnx") # Extract entities text = "Apple CEO Tim Cook announced iPhone 15 in Cupertino yesterday." result = extractor.extract_entities(text, ["company", "person", "product", "location"]) print(result) - Notebooks
- Google Colab
- Kaggle
| { | |
| "architecture": "boundary", | |
| "export_version": 2, | |
| "base_model": "fastino/gliner2.5-multi-v1", | |
| "candidate_budget": 192, | |
| "pair_temperature": 1.0, | |
| "opset": 17, | |
| "inputs": [ | |
| "input_ids", | |
| "attention_mask", | |
| "text_word_indices", | |
| "text_word_mask", | |
| "query_marker_indices", | |
| "query_marker_mask" | |
| ], | |
| "outputs": [ | |
| "start_logits", | |
| "end_logits", | |
| "pair_indices", | |
| "pair_logits", | |
| "pair_valid" | |
| ], | |
| "notes": "v2 graph: proposer + shared pool + pair reranker included. Host decode: keep pair_valid candidates with sigmoid(pair_logits / pair_temperature) >= threshold, dedupe (start,end), resolve overlaps per label, boundaries are word indices (s,e) half-open over words." | |
| } |