Roberta Pt Br

🚀 RoBERTa for Portuguese (pt_BR)

A Portuguese language model fine - tuned for Brazilian Portuguese, capable of filling masked words in text.

🚀 Quick Start

This model can be used to fill masked words in text. Here is a simple example using the transformers library:

from transformers import pipeline
unmasker = pipeline('fill-mask', model='josu/roberta-pt-br')
text = 'Brasilia é a capital do <mask>'


[{'sequence': 'Brasilia é a capital do Brasil',
  'score': 0.24386335909366608,
  'token': 707,
  'token_str': ' Brasil'},
 {'sequence': 'Brasilia é a capital do estado',
  'score': 0.2320091277360916,
  'token': 1031,
  'token_str': ' estado'},
 {'sequence': 'Brasilia é a capital do país',
  'score': 0.0665697380900383,
  'token': 998,
  'token_str': ' país'},
 {'sequence': 'Brasilia é a capital do Rio',
  'score': 0.05980581417679787,
  'token': 993,
  'token_str': ' Rio'},
 {'sequence': 'Brasilia é a capital do capital',
  'score': 0.058453518897295,
  'token': 2027,
  'token_str': ' capital'}]

💻 Usage Examples

Basic Usage

from transformers import pipeline
unmasker = pipeline('fill-mask', model='josu/roberta-pt-br')
text = 'Brasilia é a capital do <mask>'
result = unmasker(text)
print(result)

Advanced Usage

# You can use this model in a more complex NLP pipeline, such as integrating it into a chatbot or a text - generation system.
# For example, you can use a loop to process multiple masked texts.
texts = ['Brasilia é a capital do <mask>', 'O <mask> é um animal doméstico']
for text in texts:
    result = unmasker(text)
    print(result)

📚 Documentation

Tags

• Portuguese
• Brazil
• pt_BR

Widget

The widget can be used to test the model with the following text:

• Brasilia é a capital do <mask>