GL2U: A Python package implementing the generic loopy 2U algorithm for inference on non-separately specified, non-binary credal networks.
PyBrain: I also participate in the PyBrain project, which is a modular Machine Learning Library for Python. It's goal is to offer flexible, easy-to-use yet still powerful algorithms for Machine Learning Tasks and a variety of predefined environments to test and compare your algorithms.
GL2U-II: A Python/C++ package implementing the improved generic loopy 2U algorithm for inference on non-separately specified, non-binary credal networks. The performance critical part is implemented in C++, together with improved algorithm, GL2U-II is orders of magnitudes faster than the previous version of GL2U.
eNES: Efficient natural evolution strategy. A clean stochastic search algorithm (or evolution strategy, whatever you want to call it) for blackbox optimization.กก