AppArel

Thumbnail
Artboard
LookBook
Landing Page

Inspiration

There are new and exciting fashion designs being introduced into the market, and we wish to bolster this phenomenon by harnessing the power of Deep Learning. Also, there hasn't been an easy way to store designs for future reference.

Until now.

What it does

AppArel has some main functionalities:

An API-driven solution that lets you generate lists of designs you like
A GAN-driven solution that suggests new items of clothing for design inspiration (Shirts, pants, dresses)
An artboard feature that lets you import the generated images and work your sketches onto them.

It is a one stop tool to find inspiration for new designs, set up your own collection and sketch new ideas for your next look.

How we built it

The frontend was built using Figma and Angular
The backend was built using Flask and has been deployed to Replit.
The network was generated by training NIVIDIA's StyleGAN2ADA network on custom data

Challenges we ran into

We faced many problems configuring the development environment to train the model. Also, the models took a long time to train
Integrating the front-end with the backend was also quite tricky.

Accomplishments that we're proud of

All three of us learnt something new through this hackathon and we are proud to be able to develop and host an end-to-end product like this in such a short time. Integrating all three modules was a task and so was time management. We learnt more about Angular - how we can structure our project better, how GAN essentially works and how we could leverage different technologies to make something new. Integrating the backend with frontend was interesting too.