numind/NuNER_Zero
Token Classification
•
Updated
•
53.3k
•
70
The best compact Zero-Shot NER models with MIT license
Note The Zero-shot NER model * +3.1% more capable than GLiNER-large-v2.1 * reliably works with long entities (5 tokens+) * trained on diverse dataset tailored for real-life use cases
Note The Zero-shot NER model * +4.5% more capable than GLiNER-large-v2.1 * trained on diverse dataset tailored for real-life use cases
Note The Zero-shot NER model * 4096 context window * reliably works with long entities (5 tokens+) * +1% more capable than GLiNER-large-v2.1 * trained on diverse dataset tailored for real-life use cases
Note Paper citation for this work