//=time() ?>
Text-to-image with OpenAI's CLIP is getting more and more surreal. Here is "A banquet hall"
Amazing results on text-to-image generation by DALL-E. Seems like language interface is the future of HCI. https://t.co/hv900VtDvG
These are from a StyleGAN-encoder trained on my WikiArt paintings model.
Work-in-progress notebook in Eden: https://t.co/mk7XEI5tQ3
Here’s a few more recognizable images regenerated by the model
another Kandinsky texture loop, but this one using Perlin noise distortion. it's a little clumpy, but i like it
hi-res texture synthesis, i.e. contentless style transfer. here is one based on Kandinsky's Composition 8 (1600x1260 version at https://t.co/QjaHr3eVa7)
some more neural synthesis/deepdream experiments... masks for new channels radiating outward from the center