TOWE-EACL

Code for EACL 2021 paper : Attention-based Relational Graph Convolutional Network for Target-Oriented Opinion Words Extraction

Models and results can be found at our EACL 2021 paper https://www.aclweb.org/anthology/2021.eacl-main.170.pdf

requirement

• python==3.7.7
• numpy==1.19.4
• pandas==1.1.4
• torch==1.7.0
• torch-cluster==1.5.8
• torch-scatter==2.0.5
• torch-sparse==0.6.8
• torch-spline-conv==1.2.0
• torch-geometric==1.6.1
• tqdm==4.46.0
• fitlog==0.9.13
• spacy==2.3.4
• transformers==4.1.1

pre-requisites

1. prepare spacy model for dependency parse.
   • we prepare en_core_web_sm-2.2.5.tar.gz at ./data/spacy_model, please unzip it.
2. create folders ./models , ./log , ./logs in the root directory of this project.
   • ./models is the folder where model store.
   • ./log is the folder stored log recording train and valid process.
   • ./logs is the folder stored logs which fitlog generate.
3. prepare bert model.
   • please download https://huggingface.co/bert-base-uncased

model

• Target-BiLSTM
• ARGCN

train and evaluate

bash run.sh

result