genetically_evolving_neural_network 0.3.0

This package mimicks the process of Genetic Evolution on Entities, each comprised of a Neural Network, through cross-breeding parents and genetic mutations.

0.3.0

• Implemented to and from JSON on models
• Implemented file parsing for [GENNGeneration] objects

0.2.0

• Refactored mutations logic to be more efficient
• Opened entityService and populationService up to be visible for GENN constructor
• Added additional tests and documentation
• Updated example directory
  • Added Image Number Classifier complex example
  • Made example files much more reusable for future examples
  • Added Positive or Negative Number Classifier simple example

0.1.1+hotfix

• Fixed bug around too many perceptrons existing in output layer

0.1.1

• Updated package documentation

0.1.0

• Initial release
• Working implementation of a Neural Network changing over time through Genetic Evolution
• Perceptrons and Perceptron Layers will mutate, duplicate, and remove themselves randomly over time