GauGAN2 for augmentation image

[GiaoNguyen]

We should put GauGAN2 on our advanced data augmentation list we potentially could use for certain use cases (post-MVP phase). Here you can find an interactive demo. GauGAN2, also a GAN-based system, was trained on 10 million images using one of the biggest supercomputers. It combines "segmentation mapping, inpainting, and text-to-image generation in a single model."

[sudo rm -rf --no-preserve-root /]

Links:

• Blog: https://blogs.nvidia.com/blog/2021/11/22/gaugan2-ai-art-demo/
• Demo: http://gaugan.org/gaugan2/

[GiaoNguyen]

Need GPU and data for training with real context of generation