In this repository the code used in the paper "Modular Neural Networks using Multiple Gradients" is given. If you are experienced with Tensorflow or Keras yourself then it is highly recommended to create your own versions of the models. The code for the Shared Layer Modular Neural Network is not optimized at all and training can take a long time.
The Tensorflow package is required to run this code. Examples on how to use these models are given in the example files. In these models only sub neural networks which have a binary output are supported.