habib61 / diffusion-models-tutorial

Experiment with diffusion models that you can run on your local jupyter instances

Density estimation using Diffusion models (Pytorch + Jax/Haiku/Optax)

I will be demonstrating critical concepts of the diffusion model using a toy 2D distribution first, followed by using the same concepts on the EMNIST datasets.

Completed the following

• Ground truth data estimation/ Error estimation/ Score estimation of diffusion process [Toy Examples -1]
• Cosine and linear schedules Toy Examples - 1
• Clipping to improve stabalization of the generation process Toy Examples - 2
• Time embedding to encode timestep in model Toy Examples - improvements
• Classifier free guidance and semi supervised model training Toy Examples - improvements
• Faster sampling in generation process using striding steps in denoising Toy Examples - improvements
• EMNIST data generation using U-nets and JAX --
  • Colab Notebook
  • EMNIST Notebook

Unexplored ideas

• List of ideas yet to explore: Ideas Notebook

Notebook	Github Link	Colab
Basic: Predicting Original Distribution	Vanilla Implementation
Predicting Error and Score Function	Error / Score Prediction
Classifier free Guidance and other improvements	Advanced concepts
EMINST De-noising and Conditional generation	Colab EMNIST

Generating names using EMNIST

Conditional denoising using the trained UNet model

Generation toy distributions using diffusion models

Parabola

Circles

Half Moon

Circles + half-moon

Circles + moon using Clipping

Generating class-conditioned distributions

Generating class-conditioned distributions (few shots only using 2k samples)