G

Gliner Small News V2.1

Developed by EmergentMethods
A fine-tuned version based on GLiNER, optimized for news domain entity recognition, achieving up to 7.5% higher zero-shot accuracy across 18 benchmark tests
Downloads 34
Release Time : 4/25/2024

Model Overview

This model excels at entity extraction from long news texts. The underlying dataset was constructed with a global perspective by enforcing country/language/topic/time diversity, with all fine-tuning data synthetically generated

Model Features

Cross-domain Topic Recognition
Specially optimized for entity extraction in long news texts
Global Perspective Data
Training data enforces country/language/topic/time diversity
Synthetic Data Generation
Uses WizardLM and Llama3 for news translation/summarization and entity annotation

Model Capabilities

News text entity recognition
Multilingual text processing (via translation)
Zero-shot transfer learning

Use Cases

News Analysis
News Event Entity Extraction
Extract key information such as people, locations, and time from news reports
Accurately identified entities like people, locations, and organizations in the Ciudad Juárez arrest case
Content Understanding
Cross-language News Analysis
Perform entity recognition on translated news texts
Featured Recommended AI Models
AIbase
Empowering the Future, Your AI Solution Knowledge Base
© 2025AIbase