N

Ner Portuguese Br Bert Cased

Developed by rhaymison
BERT-based Portuguese named entity recognition model supporting multiple entity categories
Downloads 53
Release Time : 3/19/2024

Model Overview

This model is a fine-tuned version of google-bert/bert-base-cased on the MultNERD dataset, specifically designed for Portuguese named entity recognition tasks, covering various entity categories such as locations, persons, organizations, etc.

Model Features

Multi-category Entity Recognition
Supports recognition of 16 entity categories, including locations (LOC), persons (PER), organizations (ORG), etc.
High Accuracy
Achieves excellent performance with an F1 score of 0.8889 and accuracy of 0.9810 on the evaluation set.
Based on BERT Architecture
Fine-tuned on the robust BERT-base-cased architecture, offering strong contextual understanding capabilities.

Model Capabilities

Named entity recognition in Portuguese texts
Multi-category entity classification
Long-text processing (supports up to 512 tokens)

Use Cases

News Analysis
News Figure Identification
Identifying key figures and institution names from news texts
Successfully identified entities such as political figure JAIR BOLSONARO in the example.
Geographic Information Extraction
Location Identification
Extracting geographic information from texts
Accurately identified Brazilian place names like Rio Grande do Sul in the example.
Legal Document Processing
Case Entity Extraction
Extracting involved parties and institution information from legal documents
Identified crime types and involved institution names in the example.
Featured Recommended AI Models
AIbase
Empowering the Future, Your AI Solution Knowledge Base
© 2025AIbase