Diffusion Model Meetup & Paper Reading (Online + SF In-Person) — DDIM Paper Walk Through + Implementation from Scratch
TLDR: This is one of our first sessions for our new online Diffusion Model Paper Reading Group that is open to everyone. We are mainly a group of engineers and PM friends who work in the AI field, based in NY, SF, Toronto, and Boston. We will explain one of the most popular Diffusion Model Architectures from scratch and help you to be able to code it out!
Requirements: You should be able to follow through as long as you have basic PyTorch & a rough understanding of how Diffusion Model work & Python programming experience.
This will be a hybrid workshop in person (from SF) and online. We will release the SF location here soon!
🗓 Agenda
Hour 1: Paper deep dive — Denoising Diffusion Implicit Models (DDIM)
Hour 2: Speed Friending & Mingle
📚 Pre-Class Learning
Review these to get the most out of our session:
🧾 DDIM Paper: Denoising Diffusion Implicit Models (Song et al., 2020): https://arxiv.org/abs/2010.02502
🧾 Paper Summary with Code: LabML DDIM Notes
🎥 Optional Video to watch: https://www.youtube.com/watch?v=SKFcQ-31ZyM
🎥 Optional Video to watch: https://www.youtube.com/watch?v=JRqJ20wb-1M
✨ Why Learn DDIM Now
DDIM introduced non-Markovian sampling and helped cut generation steps from hundreds to just 25–50 — making real-time GenAI practical. It’s a crucial step in the evolution of diffusion-based generation, used in models like Stable Diffusion.
🧠 About the Diffusion Model Bootcamp
This session is part of our 5-month, peer-led learning journey — designed for curious engineers, PMs, researchers, and technical artists.
🔹 No ML background needed — just interest and 2–4 hours/week
🔹 Learn key architectures in the diffusion model landscape
🔹 Build your own GenAI app, train a model, and explore ComfyUI
🔹 Monthly paper readings, final project, and supportive peer group
📅 Next cohort begins October 2025
📣 Spots are limited — join this session to explore if it’s a fit!
👉 Questions? DM the organizer!