Scaling Autoregressive Models for Content-Rich Text-to-Image Generation
Offered By: Yannic Kilcher via YouTube
Course Description
Overview
Explore an in-depth analysis of Parti, a groundbreaking autoregressive text-to-image model that demonstrates the power of scaling in AI. Delve into the model's architecture, impressive outputs, and its ability to generate crisp, accurate, and realistic images from complex prompts. Examine the datasets used, including PartiPrompts, and review experimental results that showcase Parti's capabilities. Learn about the model's strengths in combining arbitrary styles and concepts, as well as its potential limitations through failure case studies. Gain insights into the future of AI-generated art and content creation through this comprehensive examination of Parti's innovative approach to text-to-image generation.
Syllabus
- Introduction
- Example Outputs
- Model Architecture
- Datasets incl. PartiPrompts
- Experimental Results
- Picking a cherry tree
- Failure cases
- Final comments
Taught by
Yannic Kilcher
Related Courses
Forensic Engineering: Learning from FailuresDelft University of Technology via edX Failure Analysis And Prevention
Indian Institute of Technology Roorkee via Swayam Dynamic Behaviour of Materials
Indian Institute of Technology Guwahati via Swayam Principles of Metal Forming Technology
Indian Institute of Technology Roorkee via Swayam Success and Failure in Entrepreneurship: Discover the Key to Business Success
Coventry University via FutureLearn