Inspiration
We as a diverse team have always found it hard to see ourselves represented in a meaningful way in all sorts of things, ranging from shows to movies. Making this project was very important to us as it solves an issues that many other kids, like us, face - creating an application that allows them to see themselves on the screen felt very important. Also, we wanted to make sure the application we built was for an industry which connected and impacted the current younger generation.
What it does
Avatar X is revolutionizing gaming with an AI-driven solution that creates personalized avatars and props from user photos, addressing the lack of diversity and customization in video game characters.
How we built it
We started with an open source - shap-e and point-e. Then we tested a couple of models on Google Colab. After that we figured out a diffuser from huggingpeace. We made backend with python integrated with the Gemini and the open source model, and the frontend authentication was handled by Clerk.
Challenges we ran into
Figuring out the right open source model to use, as well as how to create the vision we had in a way that fit in our limited parameters (eg GPU).
Accomplishments that we're proud of
We're really proud we managed to get the project done in time to present! We were worried we wouldn't be able to make such a complex tool so quickly. We're also really proud of the fact that we came together despite having never met each other before and worked so seamlessly to create something we all liked.
What we learned
We spent hours just researching and gathering data, and we came across lots of interesting information that we hadn't known before, especially regarding open source models.
What's next for Avatar X
We want to improve our work and create something that will allow many people across the globe to experience the joy of being seen, even in something as simple as their favourite games.
Log in or sign up for Devpost to join the conversation.