This is the DALL-E-Bot: a fully autonomous robot with web-scale diffusion models that can rearrange objects in a scene by inferring a text description of them and generating an image representing human-like arrangement of them. This is done with DALL.E without needing any training or data collection.
Must see robots & drones:
DALL-E-Bot: Introducing Web-Scale Diffusion Models to Robotics
The robot prompts DALL.E with a list of objects it detects and get an image in return that contains human-like object arrangement.
[HT] [official site]
*Our articles may contain aff links. As an Amazon Associate we earn from qualifying purchases. Please read our disclaimer on how we fund this site.
