This repo includes everything you need to build a web app for Peltarion's tutorial Author style predictor. In this tutorial, we’ll show you how to use our platform to build a deep learning model on your own and figure out which Nobel laureate you were in your past life!
The tutorial will teach you how to:
- Build, train, and deploy a BERT model in the Peltarion Platform.
- Create a web app that can decide which author a piece of text could have been written by.
When you have built, trained, and deployed a model with the Peltarion platform, you can use the content in this repo to create a web app that uses the model.
The idea is that the app will display a simple web page with a text area and a submit button. The aspiring writer submits some text, clicks the submit button, and gets a response page stating which author could have written the text. Simple but effective!
If you want to try out our version of the app - get into the mood with hot chocolate and a candle... then open the app author-style.demo.peltarion.com and write a few sentences right from your mind or heart. (ok, if creativity is zero it's perfectly fine to copy something from Google). Then - push the button and check your writing... Surprised? Honored? Terrified?
If you don't have an account on the Peltarion Platform yet, here's where you sign up. It's all free and you get instant access
NOTE: These steps are primarily inteded for quick reference / summary. We strongly recommend you to take the full tutorial Author style predictor
- Deployed model on Peltarion's platform (see above)
- Nodejs - v10+ (tested with v10 and v12)
- Linux / mac environment (not tested on windows, Dockerfile is provided)
- Create a new folder called
config
in the root folder of this project - Copy sample-config.json to
config/app-config.json
- Update
app-config.json
with your data (see tutorial)
Make sure you have created a config file (see the tutorial: Author style predictor)
npm install
npm start
Point a browser to 127.0.0.1:3000
Make sure you have created a config file (see above)
npm run docker-build
npm run docker-up
Point a browser to 127.0.0.1:3001