The code of paper is developed on a python package for processing NLP tasks based on transformers, pytorch, datasets etc. libraries. Its mission is to reduce the duplicate labors when we set up and NLP models or framework in current popular deep learning framework or methodology cross multiple nodes.
- PPF: used to evaluate methods for PTR.
Source of data: PPF
Clone the repository and install NLPx with the following commands
git clone git@github.com:codesedoc/FDSC.git
cd FDSC
pip install -e .
- Ubuntu (22.04 LTS)
- Docker (>= 23.0.5)
bash sh/docker-1024
Running follow command to build the image of basic environment of NLPx.
docker compose build nlpx-env
To use Nvidia GPU in docker containers, please install the "NVIDIA Container Toolkit" referring to here.
Running follow command to start
bash sh/docker-run
To conduct different variants of method in paper, please define the value of variable "ARGS_FILE" to the path of experiment argument file.
@InProceedings{sheng-etal-2023-decoupling,
author="Xu, Sheng
and Suzuki, Yoshimi
and Li, Jiyi
and Fukumoto, Fumiyo",
editor="Luo, Biao
and Cheng, Long
and Wu, Zheng-Guang
and Li, Hongyi
and Li, Chaojie",
title="Decoupling Style from Contents for Positive Text Reframing",
booktitle="Neural Information Processing",
year="2024",
publisher="Springer Nature Singapore",
address="Singapore",
pages="73--84",
isbn="978-981-99-8178-6"
}