Bertimbau-base-lener_br Open-source Model - Free Deployment, Accurate Marking of Portuguese Text Entities

Bertimbau Base Lener Br

Developed by Luciano

A named entity recognition model fine-tuned on the lener_br dataset based on BERTimbau (Portuguese BERT), designed for entity tagging tasks in Portuguese texts.

Sequence Labeling

Transformers

OtherOpen Source License:MIT #Portuguese NER #Legal Entity Recognition #High-precision BERT

Downloads 2,303

Release Time : 3/2/2022

Model Overview

This model is a BERT model specifically optimized for Portuguese named entity recognition tasks, excelling on the lener_br dataset and suitable for legal domain entity recognition applications.

Model Features

High-precision Portuguese NER

Achieves 98.24% accuracy and 98.74% F1 score on the lener_br test set

Optimized for Legal Domain

Fine-tuned specifically for named entity recognition in legal texts

Based on BERTimbau

Uses the Portuguese-optimized BERT architecture to capture language-specific features

Model Capabilities

Entity recognition in Portuguese texts

Entity extraction from legal documents

Token classification task processing

Use Cases

Legal Document Processing

Legal Contract Entity Extraction

Automatically identifies key entities such as parties, dates, and amounts in contracts

Accuracy exceeds 98%

Judicial Document Analysis

Extracts case-related entity information from court documents

F1 score reaches 98.74%

Information Extraction Systems

Portuguese News Entity Recognition

Extracts person names, organization names, and location information from news articles

🚀 bertimbau-base-lener_br

This model is a fine - tuned version of neuralmind/bert-base-portuguese-cased on the lener_br dataset. It can effectively perform token classification tasks, achieving high precision, recall, F1 score, and accuracy on the evaluation set.

📚 Documentation

Model Information

This model is a fine - tuned version of neuralmind/bert-base-portuguese-cased on the lener_br dataset. It achieves the following results on the evaluation set:

Loss: 0.2298
Precision: 0.8501
Recall: 0.9138
F1: 0.8808
Accuracy: 0.9693

Model Details

Property	Details
Model Type	Fine - tuned version of neuralmind/bert-base-portuguese-cased on lener_br dataset
Training Data	lener_br dataset

Intended uses & limitations

More information needed

Training and evaluation data

More information needed

Training procedure

Training hyperparameters

The following hyperparameters were used during training:

learning_rate: 2e - 05
train_batch_size: 4
eval_batch_size: 4
seed: 42
optimizer: Adam with betas=(0.9,0.999) and epsilon = 1e - 08
lr_scheduler_type: linear
num_epochs: 15

Training results

Training Loss	Epoch	Step	Validation Loss	Precision	Recall	F1	Accuracy
0.0686	1.0	1957	0.1399	0.7759	0.8669	0.8189	0.9641
0.0437	2.0	3914	0.1457	0.7997	0.8938	0.8441	0.9623
0.0313	3.0	5871	0.1675	0.8466	0.8744	0.8603	0.9651
0.0201	4.0	7828	0.1621	0.8713	0.8839	0.8775	0.9718
0.0137	5.0	9785	0.1811	0.7783	0.9159	0.8415	0.9645
0.0105	6.0	11742	0.1836	0.8568	0.9009	0.8783	0.9692
0.0105	7.0	13699	0.1649	0.8339	0.9125	0.8714	0.9725
0.0059	8.0	15656	0.2298	0.8501	0.9138	0.8808	0.9693
0.0051	9.0	17613	0.2210	0.8437	0.9045	0.8731	0.9693
0.0061	10.0	19570	0.2499	0.8627	0.8946	0.8784	0.9681
0.0041	11.0	21527	0.1985	0.8560	0.9052	0.8799	0.9720
0.003	12.0	23484	0.2204	0.8498	0.9065	0.8772	0.9699
0.0014	13.0	25441	0.2152	0.8425	0.9067	0.8734	0.9709
0.0005	14.0	27398	0.2317	0.8553	0.8987	0.8765	0.9705
0.0015	15.0	29355	0.2436	0.8543	0.8989	0.8760	0.9700

Framework versions

Transformers 4.8.2
Pytorch 1.9.0+cu102
Datasets 1.9.0
Tokenizers 0.10.3

📄 License

This model is released under the MIT license.

Featured Recommended AI Models

Empowering the Future, Your AI Solution Knowledge Base

English 简体中文繁體中文にほんご