v0.5.0

30.05.2023 - Version 0.5.0

Maintenance

• Adding a check in the TargetManager.py class that all provided splits are not empty.
  This avoids getting an error after costly training if the test set was empty. (Closes #79)
• Adding double-check if the cuda device from the out.yml file is available for the Inferencer
  module in cuda_device.py (Closes #73)
• Simplifying the predict example.
  Manual path correction is no longer necessary. Also added fix for mapped_predictions to show up correctly
• Minor dependency updates