🤖 Edge#233: Understanding DALL-E 2

In this issue:

• we explain DALL-E 2;

• we discuss the DALL-E 2 paper; 

• we explore DALL-E Mini (Now Craiyon), the most popular DALL-E implementation on the market.

Enjoy the learning!  

---

💡 ML Concept of the Day: Understanding DALL-E 2 

We have gone long enough in our series about text-to-image synthesis without covering the most famous models of all: OpenAI’s DALL-E and DALL-E 2. These models have been dominating the headlines regarding text-to-image generative models and have certainly accelerated the research in text-to-image models. The first version of DALL-E was released in January 2021 and used a combination of CLIP (Edge#219) and diffusion methods to generate photorealistic images from textual descriptions.  

DALL-E 2 is the second iteration of this OpenAI architecture, providing significant improvements over its predecessor. From a capability standpoint,