Skip to content

Latest commit

 

History

History
22 lines (15 loc) · 1.38 KB

README.md

File metadata and controls

22 lines (15 loc) · 1.38 KB
Raw

DeepDixit

This is a game that leverages the weirdness of the images generated by CLIP.

The idea is that since the images generated are quite abstract and hard to decipher, it makes a fun game to try and guess the prompt that originally led the neural network to generate the image.

Running locally

  1. Add an environment variable DATABASE_URI with the connection string from the Heroku Postgres Database. Make sure to change the postgres://... at the beginning to postgresql://...
  2. Run pip install -r requirements.txt to install all dependencies.
  3. Run python wsgi.py from the root of this repository to start the Flask server.

Instructions for TA

  • For the web scraping scripts, look here
  • For the website components, see the app directory
  • For notebooks that perform image generations, check out the original notebooks by Ryan Murdock. (we use two modules that build around the models from those notebooks)
  • To see the code and parameters we are using to perform image generation, see gendata.py
  • Scripts used to load data into our database can be found in the db directoy