Implemented BCEWithLogitsLoss reweighting function
Removed save_hyperparameters from model
Apply augmentation transforms on singular images
Fixed model summary

Relevance analysis has been moved out of predict.py and into a separate
script. It is not currently functional.

As the criterion is not part of the model but instead a hyperparameter
due to the use of configuration files, it is not moved to the GPU if
selected as a device. We therefore manually move the criterion to the proper
device, which is bad practice when using lightning but works.

The obects.inv file from lightning's documentation is missing a lot of
methods, raising errors when building with sphinx. Problematic methods
have been added to the extras.txt and encoded into extras.inv

Some DensenetRS-specific code has been removed from predictor.py, it
will have to be re-added directly inside the forward of the DensenetRS model.