Eye Of Horus

• 

  kitten-box connected with joystick hardware

• 

  Application to select images to display on the KittenBot

Inspiration

Your childhood becomes an integral part of your life and video games have been an integral part of our childhoods. Games like Minecraft have affected and helped an entire generation and continue to do so with their latest innovation using Oasis AI to 3d render Minecraft. If one of the biggest open world games in the world can be AI generated then why can't the real world be as well.

What it does

We used a Kitten-box to display an environment rendered from AI generated images with the ability to navigate using a joystick using Streamlit, Kittencode-online & Databricks

How we built it

We created the hardware using Kitten-box , wires, joystick and other electrical components. we coded the analog instructions (used for navigation) and the image displaying instructions on KittenCode online. we created an image format for the display handling a large variety of problems, and a usage of database to generate images based on the user prompt using databricks for AI, we created a frontend using streamlit. The Databricks is the backbone for this, alongside openAI DALL-E models, where we were able to use their LLM and LangChain (from DataBricks) to devise a continuous text input device.

Challenges we ran into

It wasn't an easy project with hurdles awaiting at each step, we came across challenges such as displaying images on the Kitten-box, generating AI images while configuring the use of joystick analog to navigate through the world. The primary issues at hand were the variety of compability issues re the kittenbot due to the limited firmware control, and no direct ability to display images requiring a manual drawing approach and custom image format.

Accomplishments that we're proud of

We are proud to present our Kitten-box hardware that can use ai generated images to render an open world simulation from user input prompts. We are also proud of how we came together as a team and were able to successfully complete this project within the 24 hour time limit during which we not only learned new things but improved on our knowledge as well as working within a team.

What we learned

Upon completion of our project, we all have learned skills encompassing

• mastery over Kitten-box and its applications
• frontend development using Streamlit -backend development integrated with AI using Databricks
• teamwork and coordination with efficient strategies and decision making ## What's next for Eye Of Horus after completion of Eye Of Horus, we plan on working upon it to improve the quality of the project and hope to add additional features to it to make it more user friendly

We've also been working on a particular feature that would use Databricks for it's retrieval augmented generation capacity, however due to time constraints we prioritised a working system.

Built With

• dalle
• flask
• kittenbot
• langchain
• python
• streamlit