#132 · Primary category: Image Generation

big-sleep

artificial-intelligence deep-learning generative-adversarial-networks multimodality text-to-image

A simple command line tool for text to image generation, using OpenAI's CLIP and a BigGAN. Technique was originally created by https://twitter.com/advadnoun

Project last updated:02/06/22

GitHub Stars

2.6K

Forks

300

Contributors

14

License

MIT

Why we included this project

Big Sleep turns a text prompt into a generated image with a single terminal command, combining OpenAI's CLIP with a BigGAN generator. It packages Ryan Murdock's original deep-dreaming approach into something anyone with a GPU can run without assembling notebook cells by hand. The codebase is small enough to read end to end, which makes it a handy study tool if you want to see how text can steer a generative adversarial network. Bundled Colab notebooks let you try it even without local hardware. Think of it as a reference implementation and creative toy rather than a production pipeline, useful for understanding prompt-conditioned synthesis before you move on to newer model families.

Articles for this project

No articles for this project yet.

To suggest a topic or contribute an article, contact us.

Related projects in this category