---
dated: '2022-05-01'
title: 'Generating Videos with VQGAN+CLIP'
blurb: 'A super cool model that you can use to generate images and videos, with a little bit of code.'
image: ./machine-god.webp
---

import Youtube from '@components/Youtube.astro'

If you're in the AI space, specifically generative models, you owe a great deal of respect to [VQGAN](https://compvis.github.io/taming-transformers/) and [CLIP](https://openai.com/research/clip). When combined, they can be used to iteratively refine images to more closely match a prompt.


These are some results from recent work on generating art using machine learning and artificial intelligence. Most of the following are using VQGAN+CLIP, or just CLIP alone.

I gave [a presentation](/posts/vqgan-clip/cade-brown-magma-ai-2021.pdf) at UTK's Innovative Computing Laboratory (ICL) that describes some of these processes.

## Short Square Morph Videos

### mush2dna

<figure>
  <video src="https://i.imgur.com/vQYMpGI.mp4" controls playsinline preload="metadata" aria-label="Fungus morphing into DNA" />
  <figcaption>Fungus morphing into DNA</figcaption>
</figure>

### CEB 0

<figure>
  <video src="https://i.imgur.com/DguqciG.mp4" controls playsinline preload="metadata" aria-label="My face morphing into polygons" />
  <figcaption>My face morphing into polygons</figcaption>
</figure>

### AZH 0

<figure>
  <video src="https://i.imgur.com/TqdlQMl.mp4" controls playsinline preload="metadata" aria-label="My boss undergoing a revelation" />
  <figcaption>My boss undergoing a revelation</figcaption>
</figure>

## Music Videos

### (demo) ghoti - Towards Holier Places

<Youtube src="https://www.youtube.com/embed/LQW_X_mHM_k" title="YouTube video player" />