Test optimization cycle with spiking layers in model

[PeterCalifano]

Spiking NNs libraries snnTorch and Norse should be somewhat integrated with PyTorch in this regard. To do:

1. Assess functionality of ModelTrainingManager "out-of-the-box" (including configuration class).
2. Evaluate missing attributes specifically required by spiking networks.
3. Determine any delta-work/improvement to fully support SNNs optimization.