Have an idea but don't know how to draw it? Let HokusAI help you!
Explore the docs »
View Demo
·
Report Bug
·
Request Feature
Table of Contents
Named after the famous painter Katsushika Hokusai (葛飾北斎) – think of the Great Wave painting – HokusAI is built to turn your dreams into reality. Here is an example image we generated from the prompt "a fantasy magical cottage in the woods." Built with pixray, our deep learning architecture leverages state-of-the-art models from OpenAI's Contrastive Language–Image Pre-training (CLIP) architecture and University Heidelberg's Vector Quantized Generative Adversarial Network (VQGAN).
CLIP in particular excells at what is called zero-shot learning, which means it performs very well even if it has no background knowledge
Here's how it works in a simple sense
- CLIP takes in text/images and outputs their similarity
- VQGAN works by putting 2 neural networks in a battle to the end
- The first neural network is called a Generator, and its job is to produce fake outputs given some input. You can think of this guy as a counterfeiter
- The second neural network is called a Discriminator, and its job is to make determine whether the its inputs are real or fake. You can think of this guy as the cop
- These two are locked in a zero-sum game, and eventually, the goal is to have the Generator produce outputs indifferentiable from the original dataset you want to mimic
- Essentially, we have VQGAN generate images and CLIP tells us how good they are. Then, we teach it to do better
- React.js
- React Router
- Node.js
- PyTorch
- Firebase
- WebXR
- Pixray
- and much more!
This is an example of how you may give instructions on setting up your project locally. To get a local copy up and running follow these simple example steps.
Make sure you have node (version 14+) installed.
Below is an example of how you can instruct your audience on installing and setting up your app. This template doesn't rely on any external dependencies or services.
- Clone the repo
git clone https://github.com/HokusAI-Art/HokusAI-WebApp
- Install NPM packages
yarn
- Create file called '.env.local' with the Firebase JSON credentials for a web application:
REACT_APP_FIREBASE_API_KEY={"apiKey":"...","authDomain":"...","projectId":"...","storageBucket":"...","messagingSenderId":"...","appId":"...","measurementId":"..."}
Make sure your backend is up and running to generate the images! Run the app with yarn start
Distributed under the MIT License. See LICENSE for more information.
Richard Luo - sites.gatech.edu/richard
Deepak Ramalingam - https://rdeepak2002.github.io/#/home
Lucas Yim - https://github.com/ldyim
Here are some cool resources to check out!