cnn_sent_classification

1-D CNN for sentence classification TEST

---

Quick start

Pipenv

pipenv install
pipenv shell

Download dataset : Movie Review(MR) and word vector :fastText

sh script/MR_download.sh
sh script/fasttext_download.sh

Model architecture

Running

python main.py --model CNN-rand

• CNN-rand initializes the word embeddings randomly and learns them.
• CNN-static initializes the word embeddings to word2vec and keeps the weight static.
• CNN-nonstatic also initializes to word2vec, but allows them to be learned.

Result

model	ACC
CNN-rand	72.75
CNN-static	83.47
CNN-non-static	83.74

refernce:

harvardnlp/sent-conv-torch

A Complete Guide to CNN for Sentence Classification with PyTorch

Convolutional Neural Networks for Sentence Classification