WARP (World Animation Responsive Participation)

Inspiration

Advances in image generation and llms to generate naratives

What it does

we take adjectives from the user based on how they view the world today and how they imagine the world to be in the year 2100 then we feed these adjectives into an LLM to generate a prompt for stable diffusion (one prompt for current one for future) then we generate a seed image using the current prompt we then use img2img stable diffusion with the future prompt to morph the scene we combine these keyframes into a video using googles FILM network we also generate a story with all adjectives using an LLM we narrate this story using google text to speech finally we put everything in an immersive experience

How we built it

using most advanced apis available along with open source models (Chat GPT, GPT3 , Stable Diffusion (+fine tuning for texual inversion), FILM, google text-to-speech, poseNET)

Challenges we ran into

projector mapping on ubuntu

Accomplishments that we're proud of

First time art installation for both of us

What we learned

What's next for WARP (World Animation Responsive Participation)

January 13th Exebition in Gray Area

Built With

• chat-gpt
• film
• google-texttospeech
• gpt3
• javascript
• p5js
• python
• stable-diffusion