B

Bert Base NER

Developed by optimum
BERT-base based named entity recognition model, capable of identifying four types of entities including locations, organizations, and person names
Downloads 69
Release Time : 3/24/2022

Model Overview

This model is a fine-tuned version of BERT-base, specifically designed for English named entity recognition tasks, achieving industry-leading performance on the CoNLL-2003 dataset

Model Features

High-precision Entity Recognition
Achieved 91.3 F1 score on the CoNLL-2003 test set
Multi-category Recognition
Supports four types of entity recognition: LOC/ORG/PER/MISC
Subword Token Processing
Handles subword segmentation issues based on BERT architecture

Model Capabilities

Identify named entities in text
Distinguish entity types (location/organization/person name/others)
Process incomplete word fragments

Use Cases

Text Analysis
News Entity Extraction
Automatically identify organization names and individuals from news texts
Test set accuracy 90.7%
Geographic Information Annotation
Identify geographic location information in text
Location recognition F1 score 91.9%
Featured Recommended AI Models
AIbase
Empowering the Future, Your AI Solution Knowledge Base
© 2025AIbase