This repository is the official a implementation of IADB (Iterative α-(de)Blending: a Minimalist Deterministic Diffusion Model), published at Siggraph 2023 based on the official implementation.
Modified for other dataset and to test memory optimizations on different FP config.
For a simple and intuitive explanation of our method, you can read our blog post and check our 2D tutorial.
If you want to setup a new conda environment, download a dataset (celeba) and launch a training, you can follow this:
conda env create -f environment.yml
conda activate iadb
python3 iadb.py
python3 iadb_cifar10_default.py
Python 3 dependencies:
This code has been tested with Python 3.8 on Ubuntu 22.04. We recommend setting up a dedicated Conda environment using Python 3.8 and Pytorch 2.0.1.
The iadb.py contains a simple training loop.
It demonstrates how to train a new IADB model and how to generate results (using the provided sample_iadb function).