Translation Pt En T5

🚀 T5 Portuguese-English Translation Implementation

This project implements T5 for Portuguese-English translation tasks with a modest hardware setup, introducing improvements in tokenization and post - processing.

🚀 Quick Start

This repository provides an implementation of T5 for Portuguese - English translation tasks using a modest hardware setup. We've made some changes to the tokenizer and post - processing to enhance the results and employed a Portuguese pre - trained model for translation. You can find more information in our repository. Also, check out our paper!

✨ Features

• Languages Supported:
  • English
  • Portuguese
• Datasets Used:
  • EMEA
  • ParaCrawl 99k
  • CAPES
  • Scielo
  • JRC - Acquis
  • Biomedical Domain Corpora
• Tags:
  • translation
• Metrics:
  • bleu

💻 Usage Examples

Basic Usage

Just follow the "Use in Transformers" instructions. It's necessary to add a few words before to define the task for T5. You can also create a pipeline for it. Here is an example with the phrase "Eu gosto de comer arroz":

from transformers import AutoTokenizer, AutoModelForSeq2SeqLM, pipeline
  
tokenizer = AutoTokenizer.from_pretrained("unicamp-dl/translation-pt-en-t5")

model = AutoModelForSeq2SeqLM.from_pretrained("unicamp-dl/translation-pt-en-t5")

pten_pipeline = pipeline('text2text-generation', model=model, tokenizer=tokenizer)

pten_pipeline("translate Portuguese to English: Eu gosto de comer arroz.")

📄 License

Citation

@inproceedings{lopes-etal-2020-lite,
    title = "Lite Training Strategies for {P}ortuguese-{E}nglish and {E}nglish-{P}ortuguese Translation",
    author = "Lopes, Alexandre  and
      Nogueira, Rodrigo  and
      Lotufo, Roberto  and
      Pedrini, Helio",
    booktitle = "Proceedings of the Fifth Conference on Machine Translation",
    month = nov,
    year = "2020",
    address = "Online",
    publisher = "Association for Computational Linguistics",
    url = "https://www.aclweb.org/anthology/2020.wmt-1.90",
    pages = "833--840",
}