🪄🖼 Edge#241: Emerging Capabilities of Text-to-Image Synthesis Models

+NVIDIA's textual inversion approach; +Outpainting interfaces

In this issue:

• we conclude our text-to-image series discussing the emerging capabilities of text-to-image synthesis models; 

• we explain NVIDIA’s textual inversion approach to improving text-to-image synthesis;  

• we explore DALL-E and Stable Diffusion Outpainting Interfaces. 

Enjoy the learning!  

---

💡 ML Concept of the Day: Emerging Capabilities of Text-to-Image Synthesis Models 

To conclude our series about text-to-image synthesis models, today I would like to discuss some of the new areas of research that are powering new capabilities in this type of model. The first efforts in large-scale text-to-image generation have been focused on the creation of high-fidelity outputs. As that problem looks more and more solved, complementary capabilities are commanding a lot of research in the text-to-image space. Specifically, we think the following capabilities are going to become prominent components of the next generation of text-to-image generation models: