This repository contains codes and configs for training models presented in:
Chinese Opinion Role Labeling with Corpus Translation: A Pivot Study [EMNLP2021]
To train our ChineseORL-with-Corpus-Trainslation models, you should get into one folder what you want, and set your corpus in expdata/opinion.cfg
set your GPU device in runme.sh:
export CUDA_VISIBLE_DEVICES=[gpu_num]
run:
sh runme.sh
During the training process, test will conduct every n step, set n in expdata/opinion.cfg:
validate_every = [n]
In our training process, we test every half epoch.
If you want to abtain the Chinese MPQA corpus, please email zenrran@gmail.com with your information (name, organization, and the purpose).