Not a painter? No problem because all you need to do is say a few words and the latest version of GauGAN2 can produce a picture.
Building on the power of artificial intelligence (AI), the deep learning model lets anyone turn their imagination into masterpieces with just a few words. For instance, say “wave over rocks” and the AI creates the picture in real time. Add an additional word and the model modifies the picture immediately.
This new text-to-image feature can be experienced on NVIDIA AI Demos, where users can create and customise scenes quicker and with finer control.
GauGAN2 combines segmentation mapping, inpainting and text-to-image generation in a single model, making it a powerful tool to create photorealistic art with a mix of words and drawings.
The deep learning model was trained on 10 million high-quality landscape images using the NVIDIA Selene supercomputer. Researchers used a neural network that learns the connection between words and the visuals they correspond to like “winter,” “foggy” or “rainbow.”
The research demo illustrates the possibilities for powerful image-generation tools for artists, or for that matter, anyone.