Pyro is a probabilistic programming library built on PyTorch. It allows for flexible and expressive modeling of probabilistic systems and supports a wide range of applications, from Bayesian modeling to deep probabilistic models.

Stan is a probabilistic programming language that excels in Bayesian modeling and statistical analysis. It provides a declarative syntax for defining probabilistic models and employs Hamiltonian Monte Carlo for efficient sampling. PyStan is Python implementation of Stan.

Edward2 (TensorFlow Probability) is part of the TensorFlow Probability library and is designed for probabilistic modeling using TensorFlow. It extends TensorFlow to include probabilistic programming constructs, enabling the definition of Bayesian models within TensorFlow.

PyMC3 is a probabilistic programming library that focuses on Bayesian statistical modeling. It offers a high-level, intuitive syntax for specifying probabilistic models and uses advanced sampling methods for inference.