Image generation model that you can understand by implementing it - From CNN basics to Diffusion

Lecture Topics 📖

• Understanding while implementing an image generation model
• Implemented with AUTO ENCODER / VAE / GAN / DIFFUSION

Course Target Audience/Course Purpose 🙆‍♀️

Types of students considered by knowledge sharers

• For those curious about image generation models
• I know the general flow, but I find it difficult to actually implement it.
• For those who want to understand the meaning through a simple implementation rather than a difficult implementation

Students' Concerns & Solutions

• Don't you want to learn by implementing your own image generation model?
• Aren't the codes you find on the Internet too complicated to understand?
• Let's implement only the core with a simple implementation.

Changes in students after attending the lecture

• Understanding Image Generation Models

Lecture Features ✨

I've been studying deep learning for over five years, even though I'm not a major in the field. Image generation models are fascinating, and I'm studying them. However, every resource I find presents such complex code implementations that I find analyzing the code more challenging than studying the actual model. To avoid the same struggles as I did, I'll focus on the core concepts and implementations, along with a simple, easy-to-follow guide.

• We will walk you through the entire coding process together.
• Rather than detailed formulas, I will explain with an easy-to-understand example.
• Let's run it in Colab.
• You should be able to handle Tensorflow simply.

What you'll learn 📚

_{Understanding CNNs from a Different Perspective - Extracting Outlines with CNNs}

_{Mapping handwritten data to latent space using AUTO ENCODER/VAE}

_{Restored from potential space}

_{Generating handwriting data with GAN}

_{Generating handwriting data with DIFFUSION}

Expected Questions Q&A 💬

Q. Can you explain the formula?
A. Rather than giving detailed formulas, I will explain the concept through examples.

Q. Is it difficult to implement the code?
A. I will implement the concept in the simplest way possible.

Things to note before taking the course 📢

Practice environment

• Written in Google Colab.
• Basic knowledge of Tensorflow is required.

Learning Materials

• Source code provided

Player Knowledge and Precautions

• Simple TensorFlow Usage / A Simple Understanding of the Normal Distribution

Introducing the Knowledge Sharer ✒️

Hello! My name is [Awesome]. I'm a non-major and have been studying deep learning for over five years. Deep learning is so much fun. I'm particularly fascinated by image generation models. While researching, I often found the concepts or code to be too complicated. I'd like to share what I've learned. As a non-major, I've often questioned whether it's appropriate to share this lecture. However, I believe there are definitely people in a similar situation, so I've gathered my courage and decided to share it.