Course Information
Registration period
Year-round Recruitment
Course Level
Study Mode
Duration
9 Hour(s) 49 Minute(s)
Language
English
Taught by
Maxime Vandegar
Rating
2 views
Course Overview
Diffusion Models from scratch using PyToch | In depth break down of Stable Diffusion and DALL·E 2
Welcome to this course on Diffusion Models!
This course delves into the fascinating world of diffusion models, starting from the initial research paper and advancing to cutting-edge applications such as image generation, inpainting, animations, and more. By combining a theoretical approach, and hands-on implementation using PyTorch, this course will equip you with the knowledge and expertise needed to excel in this exciting field of Generative AI.
Why choose this Diffusion Models Course?
From Theory to Practice: This course begins by dissecting the initial research paper on diffusion models, explaining the concepts and techniques from scratch. Once you have gained a deep understanding of the underlying principles, we will reproduce results from the initial diffusion model paper, from scratch, using PyTorch.
Advanced Image Generation: Building upon the foundational knowledge, we will dive into advanced techniques for image generation using diffusion models.
Inpainting and DALL-E-like Applications: Discover how diffusion models can be used for inpainting, enabling you to fill in missing or damaged parts of images with stunning accuracy. After this session, you will have a deep understanding of how inpainting works with models such as Stable Diffusion or DALL-E, and you will have the knowledge needed to modify it to your needs.
Animation Mastery: Unleash your creativity and learn how to create captivating animations using diffusion models.
Dive into Stable Diffusion: Gain an in-depth understanding of Stable Diffusion and its inner workings by reviewing and analyzing the source code. This will empower you to utilize Stable Diffusion effectively in your own industrial and research projects, beyond just using the API.
Stay Informed with Impactful Research: Stay up to date with the latest advancements in diffusion models by reviewing impactful research papers. Gain insights into the cutting-edge techniques and applications driving the field forward, and expand your knowledge to stay ahead of the curve. Register now to access our comprehensive online course on Diffusion Models and learn how this technology can enhance your projects. Don’t miss this opportunity to learn about the latest advances in Generative AI with Diffusion Models!
Register now to access our comprehensive online course on Diffusion Models and learn how this technology can enhance your projects.
Don’t miss this opportunity to learn about the latest advances in Generative AI with Diffusion Models!
Course Content
- 7 section(s)
- 57 lecture(s)
- Section 1 Introduction
- Section 2 Initial paper on Diffusion Models
- Section 3 Denoising Diffusion Probabilistic Models
- Section 4 Inpainting
- Section 5 Animating Diffusion Models
- Section 6 Stable Diffusion
- Section 7 Paper Review
What You’ll Learn
- How Diffusion Models work
- Implementation of Diffusion Models from scratch using PyTorch
- In depth understanding of inpainting with Diffusion Models
- Deep analysis of Stable Diffusion: opening the black box
- Making great animations with Diffusion Models
- Review of impactful research papers
Skills covered in this course
Reviews
-
SShashank Rao Kadapanatham
The course was clearly outline and I learnt alot. A few topics to be improved: 1. The audio volume was too low. 2. While the code explaination were clear and interesting, the theory was missed out.
-
BBenoit Courbon
The overall content of this course is very worthwile, as learning to train diffusion model from scratch and sample new data from it is super valuable. The format of this course (live coding / live review of papers) could be improved IMO. It has the benefit to help us follow the thought process of the teacher, but on the other hand it results in unclear explanations and loss of time due to code debugging / writing "boring" sections of code (ie plotting). It could be improved by bringing an external set of slides with clean content, and also more preparation for the live coding
-
SSubhasish Bandyopadhyay
I do not understand what was the motivation for the DDPM paper implementation? was it just to know how to convert code from tf to pytorch? extremely disappointed. I hoped to see the relevance of the architecture to the theory of DDPM
-
KKeneilwe Mokoka
Very insightful course. Exactly what I needed. The audio could have been better. But I enjoyed going through the various papers and seeing how that translates into code.
