Member-only story
NVIDIA’s Impressive GAN Applications
The tech giant has been at the forefront of the adoption of GAN models in real world solutions.
I recently started an AI-focused educational newsletter, that already has over 100,000 subscribers. TheSequence is a no-BS (meaning no hype, no news etc) ML-oriented newsletter that takes 5 minutes to read. The goal is to keep you up to date with machine learning projects, research papers and concepts. Please give it a try by subscribing below:
NVIDIA has been one of the most influential companies advancing the research in GAN models. In addition to the research contributions, NVIDIA has released several applications that highlight the possibilities of GANs. These applications are not only practical in their own nature but also can guide data science teams about the proper architectures for GAN solutions. Below, I’ve listed some of the most notable GAN-powered projects created by NVIDIA:
GauGAN
GauGAN is one of the most interesting GAN demos ever created. The model is able to generate artistic images from a text input. In addition to the text-to-image generation, GauGAN combines different techniques such as NLP, semantic segmentation, sketch and style in a single GAN model which highlights the richness and possibilities of this type of architecture.
One of the aspects in which GauGAN excels is its extensibility. The framework allows a user to create custom segmentation maps and manipulate the scenes. The model has been trained on over 10 million of high quality images. In addition to the model, NVIDIA released an interactive demo that allow users to play with GauGAN to generate artistic landscape images from text.
GANverse3D
GANs are widely used for synthetic image generation but remain quite unexplored in other…
