Skip to content

Implement reinforcement learning

Nicola Gatto requested to merge implement-reinforcement-learning into master

Implements the reinforcement parameters that came with monticore/EmbeddedMontiArc/languages/CNNTrainLang!15 (merged). This means:

  • Update CNNTrainLang version to 0.3.0-SNAPSHOT
  • Add new python templates that implement Reinforcement Learning (more precisely DQN-learning)
  • Add template controller methods that allow access to the mentioned parameters (see src/main/java/de/monticore/lang/monticar/cnnarch/gluongenerator/ReinforcementConfigurationData.java)
  • Fix the input field of broken C++ predictor

Merge request reports