zuko and PyTorchDiscreteFlows
The tools are competitors, with Zuko being a more mature and widely adopted general-purpose library for continuous normalizing flows in PyTorch, while PyTorchDiscreteFlows specifically targets and implements discrete normalizing flows.
About zuko
probabilists/zuko
Normalizing flows in PyTorch
This project helps machine learning engineers and researchers build advanced probabilistic models. It takes in structured data and outputs flexible, high-dimensional probability distributions that can be easily trained and sampled. It is ideal for those working on complex density estimation or generative modeling tasks.
About PyTorchDiscreteFlows
TrentBrick/PyTorchDiscreteFlows
Discrete Normalizing Flows implemented in PyTorch
This is a tool for machine learning researchers and practitioners who are working with discrete data distributions. It helps you model complex, discrete data by transforming simple distributions into more intricate ones. You provide your discrete data, and it outputs a model that can generate similar discrete data or estimate the likelihood of existing data points. It's for those exploring advanced generative models in PyTorch.
Related comparisons
Scores updated daily from GitHub, PyPI, and npm data. How scores work