Skip to content

Latest commit

 

History

History
117 lines (91 loc) · 4.78 KB

README.md

File metadata and controls

117 lines (91 loc) · 4.78 KB
Raw

NLU Task Fine-tune On VECO

Download | Fine-tune | Result |

This repository contains fine-tune code of NLU tasks on VECO, and the code base is built on top of XTREME.

Pre-trained models

Model Description # Params Download
VECO_large VECO model with 24 encoder layers trained on 50 languages' monolingual and bilingual corpus 550M veco-large.tar.gz

Download the data

In order to run this code, the first step is to download the dependencies. We assume you have installed anaconda and use Python 3.7+. The additional requirements including transformers, seqeval (for sequence labelling evaluation), tensorboardx, jieba, kytea, and pythainlp (for text segmentation in Chinese, Japanese, and Thai), and sacremoses can be installed by running the following script:

bash install_tools.sh

The next step is to download the data. To this end, first create a download folder with mkdir -p download in the root of this project, and then run the following command to download the datasets:

bash scripts/download_data.sh

Run fine-tune

The evaluation setting in XTREME is zero-shot cross-lingual transfer from English. We fine-tune VECO models on the labelled data of XNLI and XQuAD task in English. Each fine-tuned model is then applied to the test data of the same task in other languages to obtain predictions.

Sentence Classification Task (XNLI)

The second sentence classification dataset is the Cross-lingual Natural Language Inference (XNLI) dataset. You can fine-tune a pre-trained multilingual model on the English MNLI data with the following command:

bash scripts/train_xnli.sh [MODEL_PATH] [GPU] [LR] [EPOCH] [BATCH_SIZE]

Question Answering Task (XQuAD)

For XQuAD, the model should be trained on the English SQuAD training set. Using the following command, you can first fine-tune a pre-trained multilingual model on SQuAD English training data, and then you can obtain predictions on the test data of XQuAD task.

bash scripts/train_squad.sh [MODEL_PATH] [GPU] [LR] [EPOCH] [BATCH_SIZE]

Sentence Retrieval Task (Tatoeba)

The second cross-lingual sentence retrieval dataset we use is the Tatoeba dataset. You can directly apply the model to obtain predictions on the test data of the task:

bash scripts/run_tatoeba.sh [MODEL_PATH] [GPU] [LAYER]

Result

The results of XNLI, XQuAD and Tatoeba task can be reproduced using the default hyperparameters listed in .sh file.

Model XNLI
(Acc)		 XQuAD
(F1/EM)		 Tatoeba
(Acc)
VECO_large 79.9 77.5/61.9 75.1

Paper

If you use this benchmark or the code in this repo, please cite the following papers \cite{Luo2020VECO} and \cite{hu2020xtreme}.

@article{Luo2020VECO
  author    = {Fuli Luo and Wei Wang and Jiahao Liu and Yijia Liu and Bin Bi and Songfang Huang and Fei Huang and Luo Si},
  title     = {{VECO:} Variable Encoder-decoder Pre-training for Cross-lingual Understanding and Generation},
  journal   = {CoRR},
  volume    = {abs/2010.16046},
  year      = {2020},
  url       = {https://arxiv.org/abs/2010.16046},
  archivePrefix = {arXiv},
  eprint    = {2010.16046},
}

@article{hu2020xtreme,
      author    = {Junjie Hu and Sebastian Ruder and Aditya Siddhant and Graham Neubig and Orhan Firat and Melvin Johnson},
      title     = {XTREME: A Massively Multilingual Multi-task Benchmark for Evaluating Cross-lingual Generalization},
      journal   = {CoRR},
      volume    = {abs/2003.11080},
      year      = {2020},
      archivePrefix = {arXiv},
      eprint    = {2003.11080}
}

Please consider including a note similar to the one below to make sure to cite all the individual datasets in your paper.

@inproceedings{Conneau2018xnli,
    title = "{XNLI}: Evaluating Cross-lingual Sentence Representations",
    author = "Conneau, Alexis  and
      Rinott, Ruty  and
      Lample, Guillaume  and
      Williams, Adina  and
      Bowman, Samuel  and
      Schwenk, Holger  and
      Stoyanov, Veselin",
    booktitle = "Proceedings of EMNLP 2018",
    year = "2018",
    pages = "2475--2485",
}

@inproceedings{artetxe2020cross,
author = {Artetxe, Mikel and Ruder, Sebastian and Yogatama, Dani},
booktitle = {Proceedings of ACL 2020},
title = {{On the Cross-lingual Transferability of Monolingual Representations}},
year = {2020}
}

@article{Artetxe2019massively,
author = {Artetxe, Mikel and Schwenk, Holger},
journal = {Transactions of the ACL 2019},
title = {{Massively Multilingual Sentence Embeddings for Zero-Shot Cross-Lingual Transfer and Beyond}},
year = {2019}
}