An Indonesian word embedding demo
Clone or download
Latest commit 0a48b4f Jan 31, 2018
Type Name Latest commit message Commit time
Failed to load latest commit information.
backend Initial commit Jan 30, 2018
dist Initial commit Jan 30, 2018
frontend Initial commit Jan 30, 2018
id Initial commit Jan 30, 2018
.gitignore Initial commit Jan 30, 2018
LICENSE Initial commit Jan 30, 2018
Procfile Initial commit Jan 30, 2018 Update Jan 31, 2018
requirements.txt Initial commit Jan 30, 2018 Initial commit Jan 30, 2018

Indonesian Word Embedding

A web application that demonstrates Indonesian word embedding, inspired by Word embedding demo.

See the demo: (might take a few seconds if the app is "asleep"--it needs to be booted and loading the vectors might take a few secs)

Right now it serves pre-trained word vectors from here. The word vectors are trained on the Word2vec model from Wikipedia texts using the CBOW training algorithm.

💡 future improvements

  • Refactoring. This is my first time using Vue so I'm sure there are a lot of things to be refactored that I haven't known yet!
  • Serves multiple models. Not only Word2vec using CBOW but also Word2vec using skip-gram or even FastText.
  • Add more in-depth explanations, possibly for each section.
  • As more explanations are added, consider using Vuex.
  • Indonesian translation.

💭 contribute

If you're interested to work in improving the points above or if you have any idea, feel free to make a pull request! 🙏

⚙ setup

# install front-end
cd frontend
npm install

# serve with hot reload at localhost:8080
npm run dev

# build for production/Flask with minification
npm run build

# install back-end
cd ../backend
virtualenv -p python3 venv
source venv/bin/activate
pip install -r requirements.txt
cd ..

# serve back-end at localhost:5000 flask run