Create videos with Stable Diffusion by exploring the latent space

AI image generator that was just released to public 2 weeks ago. The entire model as well as the source code.

Dall-E 2 and then Imagen took the cake with AI art tools, but they were closed APIs. Things were actually good enough to be generally interesting. Before this, img2img / txt2img / etc weren't very interesting.

Stable Diffusion is the first open source tool, and it doesn't require a zillion GB of RAM. so anyone can download it to their computer, and it's getting hacked on and improved on at a faster rate. IIRC there's a good version of the model down to 4 GiB and it runs on M1 CPUs now, making it accessible to an actually pretty large group of people.

Ask Apple. The community already does way too many things that are actually Apple's job.

I'm personally tired of doing Apple's support work and not getting paid for it.

Hey there, author of the project here. You can play with this project in Google Colab, so no need for it to run on your M1 :)

Here's the link: https://colab.research.google.com/github/nateraw/stable-diff...

Enjoy