Complete Tutorial: 4K Generative Art Animation PYTTI (Python Text to Image) and After Effects (2022

May 2022 Update: some of the steps have changed since this initial video was posted 3 months ago, and a new video is in the works. That said, the comments below from other viewers have been extremely helpful, so if you run into any problems please read the comments and/or add your own comments for suggestions to cover in the next version.

Here's a very simple to follow tutorial on how to make your own AI Generative Art Animation.

Final animation created in this tutorial: https://www.youtube.com/watch?v=NP8x34TlNXA

Alternate version using VQGAN instead of limited colors: https://www.youtube.com/watch?v=BSsKff6d1sg

Timestamps:
00:00 Complete Tutorial: 4K Generative Art Animation using PYTTI (Python Text to Image) and After Effects
00:50 sportsracer48 PYTTI Colab Notebook
02:52 Text and Image Prompts
09:00 Image Settings (resolution, motion, framerate)
12:15 Generate JSON settings string
13:43 First frame previews
15:10 Frames previews after 2.5 hours
16:55 Stopping / Restarting / Resuming
22:46 Finished after 32 hours
23:20 Downloading frames and importing to After Effects

Requirements:
- $5 patreon support for access to sportsracer48 colab notebook: This follows the PYTTI Colab Notebook by sportsracer48 - which is only available via their Patreon.
https://www.patreon.com/sportsracer48

$9.99 google colab pro: This will get you a GPU to run the script on. It's possible to use the free tier, but expect much longer runtimes and possible memory issues.

I also use Adobe After Effects to upscale from 1080 to 4k along with stretching the video from 40 to 60 seconds using pixel motion frame blending.

You can get even better results using more specialized scripts/software such as Topaz Gigapixel for upsampling resolution and twixtor pro and FlowFrames but it requires a bit more steps, and I wanted to keep his tutorial as simple as possible.

Music: https://soundcloud.com/soundscientists/sacred-bells