Stable Diffusion from Scratch in PyTorch | Conditional Latent Diffusion Models

In this video, we'll cover all the different types of conditioning in latent diffusion and finish stable diffusion implementation in PyTorch and after this you would be able to build and train Stable Diffusion from scratch. This is Part II of the tutorial where I get into conditioning in latent diffusion models. We dive deep into class conditioning in latent diffusion models, implementing class conditional latent diffusion model and see results of training latent diffusion model to conditionally generate mnist digits. Then we get into semantic synthesis in latent diffusion models, and implement that. We also understand super resolution in latent diffusion and see how inpainting with latent diffusion model can be done without training it for that. We then see latent diffusion model is fine tuned for inpainting. Finally we get into cross attention and see how stable diffusion uses cross attention to do text conditioning. We will also implement cross attention for latent diffusion models and see results of text to image latent diffusion models We then see how to move from latent diffusion to stable diffusion model and also talk about clip a bit. By the end of this video, you will be able to code conditional stable diffusion in PyTorch by yourself. ⏱️ Timestamps 00:00 Introduction 00:54 Recap of Unconditional latent Diffusion Models 02:15 Class Conditioning in Latent Diffusion Models 07:05 Recap of Implementation of Latent Diffusion Models 11:10 Class Conditioning Implementation in Latent Diffusion Models 16:40 Results of Class Conditioning 17:37 Spatial Image Conditioning in Latent Diffusion Models 18:47 Semantic Synthesis in Latent Diffusion Models 20:25 Semantic Synthesis Implementation in LDM 24:55 Results of Semantic Synthesis 25:49 Super Resolution using Latent Diffusion Models 26:42 Inpainting with Latent Diffusion Models 31:25 Text Conditioning Introduction 31:33 Self Attention Explained 34:58 Cross Attention Explained 38:28 Image Conditioning using Cross Attention 39:14 Text Conditioning Implementation using Cross Attention 46:26 Text Conditioning Results 47:21 Conditional Latent Diffusion Models to Stable Diffusion 51:12 Outro Paper - http://tinyurl.com/exai-latent-diffusion-paper Implementation - http://tinyurl.com/exai-stable-diffusion-repo 🔔 Subscribe https://tinyurl.com/exai-channel-link 📌 Keywords: #stablediffusion