Auto Quantum Machine Learning for Multisource Classification

• if using OSX, be aware that pennylane hates macs. You'll need to use cpu for computations with torch. Possibly even modify some of the libraries or environmental variables, just so that some internal module doesn't use it.

SD:

• Take care of naming the classes, modules and variables properly. Examples include:
  • Distinctive names for devices in hybrid models using torch. "device" is commonly used by both pennylane and torch
  • mlflow_utils has much more than only mlflow handling
  • naming training algorithm a classifier (see experiments) is confusing at best
• Typehint the module! I had to look for the information about the intended type of the variables in more places than in should need to!
• Parallelizing