语言:

• 英语

• 葡萄牙语

数据集:

• EMEA

• ParaCrawl 99k

• CAPES

• Scielo

• JRC-Acquis

• 生物医学领域语料库

标签:

• 翻译

评估指标:

• BLEU

简介

本仓库提供了一个针对英葡翻译任务的T5模型实现，采用轻量级硬件配置。我们通过改进分词器和后处理流程提升了翻译效果，并使用了葡萄牙语预训练模型进行翻译。更多信息请访问我们的代码库。另请参阅我们的论文！

使用说明

只需按照"在Transformers中使用"的指引操作。需要注意的是，使用前需为T5模型添加任务前缀词。

也可创建翻译管道，以例句"I like to eat rice"为例：

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

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

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

enpt_pipeline("translate English to Portuguese: I like to eat rice.")

引用文献

@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",
}