DAY 17 - Text to Image Generation Using OpenAI's CLIP and BigGAN - IT Consultant - SAP, Artificial Intelligence and Machine Learning

We have in our previous post “Tiger Roaring In Forest – Generate Text to Image Using OpenAI CLIP and Siren (Implicit neural representation network)“.

Today in Artificial Intelligence and Deep Learning for natural language processing creating history in many areas.

I am going to show you another library called “Big Sleep”, how to generate an image from text. You can type any combination of words in the program and it will render an abstract Image after many iterations.

Before giving the final output it creates hundreds of images and will proceed towards giving the final abstract image.

Open AI Library – Big Sleep:

Simple command-line tool for text to image generation using OpenAI’s CLIP and BigGAN.

This library can generate images (only images) using natural language.

What is CLIP?

CLIP (Contrastive Language–Image Pre-training) builds on a large body of work on zero-shot transfer, natural language supervision, and multimodal learning. The idea of zero-data learning dates back over a decade but until recently was mostly studied in computer vision as a way of generalizing to unseen object categories.

Source: CLIP

I have tried a couple of examples e-g “Thunderstorm from Sky” and “Ship on the ocean” and below are the final outcome of abstract images.

With more memory in your system, you can get better-quality images.

Run the below program and enter your own image description or you can go to Big Sleep GitHub and click Google Colab to load the code in its IDE.

After many iterations, it was rendering different abstract images of ‘Thunderstorm from Sky”.