jiangqn / IAN-pytorch

A pytorch implemention of IAN for aspect based sentiment analysis.

IAN-pytorch

This is a pytorch implementation for Interactive Attention Networks for Aspect-Level Sentiment Classification (Dehong Ma et al, IJCAI 2017).

requirements

pytorch 1.0.0
numpy
spacy

Quick Start

• Download the 300-dimensional pre-trained word vectors from Glove and save it in the 'data' folder as 'data/glove.840B.300d.txt'

• Use the following command to run the code.

python main.py

Results

Dataset	Accuracy
Laptop	70.376
Restaurant	77.679

Reference

During the procedure of implementation, the code in lpq29743/IAN is referenced.