tensorflow-constrained-optimization 0.2

A library for performing constrained optimization in TensorFlow

TensorFlow Constrained Optimization (TFCO) is a library for optimizing inequality-constrained problems in TensorFlow.

In the most general case, both the objective function and the constraints are represented as Tensors, giving users the maximum amount of flexibility in specifying their optimization problems. Constructing these Tensors can be cumbersome, so we also provide helper functions to make it easy to construct constrained optimization problems based on rates, i.e. proportions of the training data on which some event occurs (e.g. the error rate, true positive rate, recall, etc).