big-sleep VS StyleCLIP

I am currently pursuing my Master's degree in Computer Science and I am interested in working on a deep learning model that can generate images based on text descriptions. I've been interested in the field for a long time (think google deep dream, or a few years back I was very into big-sleep)

Big Sleep

Other AIs that I don't have experience with but have heard good things about are DALL-E 2 and the open source Big Sleep.

Text-to-image, e.g. with Big Sleep

At first I didn't like the reason he claimed LaMDA was conscious because it seemed mostly based on the text output of the models, but listening to that made me realize he watered it down for the mainstream medias. And I actually had my own encounter with the consciousness of Big Sleep after using it a lot one day : it looked like a mass of eyes vaguely shaped like a rabbit, and kept showing me random images. I wouldn't be as convinced they have a consciousness if I didn't see it with my 3rd eye. But then again, I also found evidence that regular non AI programs can develop one too so who know.

and after a few hours got this: https://i.imgur.com/FxdfdmV.png

Not nearly as cool as the real DALL-e, but maybe I'm missing something.

[1] https://github.com/lucidrains/big-sleep

(Added Mar. 23, 2021) Big Sleep - Colaboratory by LtqxWYEG. Uses BigGAN to generate images. Reference.

While CLIP on its own is useful for applications such as zero-shot classification, semantic searches, and unsupervised data exploration, CLIP is also used as a building block in a vast array of multimodal applications, from Stable Diffusion and DALL-E to StyleCLIP and OWL-ViT. For most of these downstream applications, the initial CLIP model is regarded as a “pre-trained” starting point, and the entire model is fine-tuned for its new use case.

I'm currently building a fork for StyleCLIP global directions which allows you to control multiple semantic parameters simoultaneously to generate and edit an image with StyleGAN and CLIP in realtime. I want to showcase its potential as a design tool. Unfortunately, GAN weights are trained on very domain-specific (faces, cars, churches) data. This makes them inferior to modern diffusion models which I can use to generate whatever comes to mind. Although I know we won't have a GAN-based DALL-E counterpart anytime soon, I still would love to use my system with weights that can output a wide variety of things.

(Added Feb. 15, 2021) StyleCLIP - Colaboratory by orpatashnik. Uses StyleGAN to generate images. GitHub. Twitter reference. Reddit post.

I used StyleCLIP. I purchased some Google Colab time to use their GPUs. I'll probably do some more later this week.

Link for Styleclip

So, I watched this Corridor Crew video on generating realistic faces from CG characters, and I wanted to try it out on the RDR2 models. The github link for the original work is here. If you guys are interested I can generate the faces of more characters from RDR2 and RDR1. I can even try some from RD Revolver.

Combining these methods with CLIP allows you to generate images based on text. This one uses a face generator. https://github.com/orpatashnik/StyleCLIP