README.md

metadata

license: mit
language:
  - pt
metrics:
  accuracy:
    Neutral: 0.99
    Positive: 0.97
    Negative: 0.98
base_model: neuralmind/bert-base-portuguese-cased
library_name: transformers
tags:
  - sentiment analysis
  - nlp
  - glassdoor
pipeline_tag: text-classification

BERTimbau for Sentiment Analysis of Glassdoor Reviews

Introduction

This model fine-tunes neuralmind/bert-base-portuguese-cased for sentiment analysis of Glassdoor reviews about IT companies in Cuiabá.

The dataset used to train the model consists of 2,532 reviews sourced from Glassdoor.

For more detail about the project, follow my GitHub.

Example Usage

from transformers import pipeline

pipe = pipeline("text-classification", model="stevillis/bertimbau-finetuned-glassdoor-reviews")

result = pipe("Empresa boa para trabalhar")

print(result) # Expected output: [{'label': 'positive', 'score': 0.9993522763252258}]