Ultimate Guide to Diffusion Models | ML Coding Series | Denoising Diffusion Probabilistic Models

❤️ Become The AI Epiphany Patreon ❤️ 👨‍👩‍👧‍👦 Join our Discord community 👨‍👩‍👧‍👦 In this 3rd video of my ML coding series, we do a deep dive into diffusion models! Diffusion is the powerhouse behind recent text-to-image generation models such as OpenAI’s DALL-E 2, Google’s Imagen, etc. I first give you some context by going over 2 seminal diffusion papers: * Denoising Diffusion Probabilistic Models * Improved Denoising Diffusion Probabilistic Models And then we dig deep into the actual code analysis comparing mathematical formulas behind diffusion with actual code implementation. ▬▬▬▬▬▬▬▬▬▬▬▬▬▬▬▬▬▬▬▬▬▬▬▬ ✅ DDPM paper: ✅ Improved DDPM paper: ✅ Improved DDPM code: ▬▬▬▬▬▬▬▬▬▬▬▬▬▬▬▬▬▬▬▬▬▬▬▬ ⌚️ Timetable: 00:00:00 (Paper) Denoising Diffusion Probabilistic Models 00:16:00 (Paper) Improved DDPMs 00:23:10 (Coding starts) Training DDPMs 00:24:50 UNet model creation walk-through 00:35:05 Gaussian Diffusion model creation walk-through 00:43:50 Training loop 00:56:08 Computing noise and variance (forward prop through UNet) 01:04:00 Variational lower bound loss 01:17:25 MSE loss 01:19:23 Sampling from diffusion models 01:26:50 Sampling an actual image 01:28:03 Outro ▬▬▬▬▬▬▬▬▬▬▬▬▬▬▬▬▬▬▬▬▬▬▬▬ 💰 BECOME A PATREON OF THE AI EPIPHANY ❤️ If these videos, GitHub projects, and blogs help you, consider helping me out by supporting me on Patreon! The AI Epiphany - One-time donation - Huge thank you to these AI Epiphany patreons: Eli Mahler Petar Veličković ▬▬▬▬▬▬▬▬▬▬▬▬▬▬▬▬▬▬▬▬▬▬▬▬ 💼 LinkedIn - 🐦 Twitter - 👨‍👩‍👧‍👦 Discord - 📺 YouTube - 📚 Medium - 💻 GitHub - 📢 AI Newsletter - ▬▬▬▬▬▬▬▬▬▬▬▬▬▬▬▬▬▬▬▬▬▬▬▬ #diffusion #ddpm #coding