Welcome to the official repository for “Generative Foundation Models: A Comprehensive Beginner’s Handbook.”
This handbook distills the theory and practice behind the most influential generative model families released in the last five years—Transformers, Diffusion models, latent-space approaches, and recent linear-time SSMs such as Mamba and Retentive Networks.
Why star & fork?
- Stars keep the project visible and motivate continuous improvements.
- Forks let you annotate, localize, or extend the book for your own courses or research.
- Every new star signals that the next chapter or tutorial is worth shipping sooner 🚀
| Chapter | Title | Key Topics |
|---|---|---|
| 1 | Introduction | Foundation-model landscape, notation |
| 2 | Transformers for Sequence Modeling | Self-attention, scaling laws |
| 3 | Vision Transformer (ViT) | Patch tokenization, image generation variants |
| 4 | Mamba | Selective State-Space Models, linear-time inference |
| 5 | U-Net Architecture | Encoder-decoder with skip connections |
| 6 | Denoising Diffusion Probabilistic Models (DDPM) | Forward/reverse processes, training objectives |
| 7 | Diffusion Transformer (DiT) | Pure-transformer denoisers, AdaLN conditioning |
| 8 | Retentive Networks (RetNet) | Multi-scale retention, recurrent decoding |
| 9 | Latent Diffusion Models (LDM) | VAE compression, Stable Diffusion paradigm |
| 10 | Text-to-3D Generation | DreamFusion, Magic3D pipelines |
| 11 | Conclusions & Outlook | Research frontiers, open problems |
Lian, J. J. (2025). Generative foundation models: A comprehensive beginner’s handbook. SSRN. http://dx.doi.org/10.2139/ssrn.5259947
@misc{lian2025generative,
author = {Junbo Jacob Lian},
title = {Generative Foundation Models: A Comprehensive Beginner’s Handbook},
year = {2025},
month = {April},
note = {Available at SSRN: \url{https://ssrn.com/abstract=5259947} or \url{http://dx.doi.org/10.2139/ssrn.5259947}},
howpublished = {SSRN},
url = {https://ssrn.com/abstract=5259947},
}# 1 Clone
git clone https://github.com/<user>/generative-foundation-models.git
cd generative-foundation-models
# 2 Browse the handbook (LaTeX sources)
code .
# 3 Compile (requires LaTeX + make)
make pdf
# 4 Enjoy the latest PDF
open dist/Generative_Foundation_Models.pdf