README.md

metadata

language:
  - en
  - de

Our bibert-ende is a bilingual English-German Language Model. Please check out our EMNLP 2021 paper "BERT, mBERT, or BiBERT? A Study on Contextualized Embeddings for Neural Machine Translation" for more details.

@misc{xu2021bert,
      title={BERT, mBERT, or BiBERT? A Study on Contextualized Embeddings for Neural Machine Translation}, 
      author={Haoran Xu and Benjamin Van Durme and Kenton Murray},
      year={2021},
      eprint={2109.04588},
      archivePrefix={arXiv},
      primaryClass={cs.CL}
}

Download

from transformers import BertTokenizer, AutoModel
tokenizer = BertTokenizer.from_pretrained("jhu-clsp/bibert-ende")
model = AutoModel.from_pretrained("jhu-clsp/bibert-ende")