artee.ai
This repo contains some models and code from artee.ai project - endless AI Generated Tees store.
T-Shirt designs and captions were generated using neural networks. Each time updating the page, a unique design is obtained. Any t-shirt can be ordered by clicking on it. Tees printed in UK using Teemill service and delivered worldwide.
StyleGAN model
A high-quality generative model is the heart of our product. So we started with research in this direction. t is obvious that the quality will depend on two things - algorithm and data.
As an algorithm we decided to use StyleGAN (original NVIDIA's repo link), since it generates high-resolution images (1024 x 1024) and it was proved that it works for different domains (tons of this[soemthing]doesnotexists sites).
As a dataset, we used subset of wikiart styles and The Museum of Modern Art dataset. We have removed low-quality images, crop non-square images and it gave us a final dataset of about 40.000 images.
You can check how our model works using colab or pytorch_stylegan_art.ipynb notebook.
Google Cloud
We use google cloud products as a backend for our store. We don't want to generate designs on the fly, because it is expensive to have GPU instance always up. Instead of this we pre-generate hundreds of thousands of designs with all required meta and store them on google cloud storage. Also, we have a google cloud firestore database with information about products. So we try not to show t-shirts which have already been ordered. f our database runs out we will just generate new designs.
We have two API endpoints (code in gcloud folder):
- get_products - returns the list of 12 t-shirts with all required meta
- get_product_url - generate URL for custom teemill product, so a user can order t-shirt
Both endpoints use google cloud functions, so we have fully serverless and scalable architecture.