search

gle-bellier / discrete-fm

Educational implementation of the Discrete Flow Matching paper
48 stars 3 forks source link
flow-matching generative-ai generative-model machine-learning pytorch pytorch-lightning

readme

Discrete Flow Matching

Open in Colab

notebook by Georges Le Bellier - Twitter, Website

This is an unofficial implementation of the paper made Discrete Flow Matching by Itai Gat et al [1]. This notebook (notebook/discretefm.ipynb) has an educational purpose and explores the case of discrete image generation. We strongly encourage the readers to check the whole framework in the original paper and we report the formula with their indexes in the paper.

References:

📄 [1] Discrete Flow Matching by Itai Gat, Tal Remez, Neta Shaul, Felix Kreuk, Ricky T. Q. Chen, Gabriel Synnaeve, Yossi Adi, Yaron Lipman - Article

📄 [2] Generative Flows on Discrete State-Spaces: Enabling Multimodal Flows with Applications to Protein Co-Design by Andrew Campbell, Jason Yim, Regina Barzilay, Tom Rainforth, Tommi Jaakkola Article

📄 [3] Flow Matching for Generative Modeling by Yaron Lipman, Ricky T. Q. Chen, Heli Ben-Hamu, Maximilian Nickel, Matt Le - Article

🐍 [4] Introduction to Flow Matching by Georges Le Bellier - Notebook

digits