MinMaxScaler for constant values

Transformations based on the MinMaxScaler (e.g. ZeroToOne) produce NaN values (division by zero) if the variable has constant values (min==max).

In CAE, this only throws a warning during scaling, but later producess invalid losses (loss=nan.0) during training.

CVAE fails altogether:

---> 16     cae.fit(datamodule, name_run='', max_epochs=5, accelerator='cpu', flag_wandb=False)

File ~\CODE\aixd\src\aixd\mlmodel\architecture\cond_ae_model.py:675, in CondAEModel.fit(self, datamodule, name_run, max_epochs, callbacks, loggers, accelerator, flag_early_stop, criteria, flag_wandb, wandb_entity, **kwargs)
    666 # Create a model_trainer object and fit the model
    667 self.model_trainer = pl.Trainer(
    668     accelerator=accelerator,
    669     max_epochs=max_epochs,
   (...)
    673     **kwargs,
    674 )
--> 675 self.model_trainer.fit(self, datamodule=datamodule)
    677 if flag_wandb:
    678     import wandb

File ..\Anaconda3\envs\aixd-dev\lib\site-packages\pytorch_lightning\trainer\trainer.py:538, in Trainer.fit(self, model, train_dataloaders, val_dataloaders, datamodule, ckpt_path)
    536 self.state.status = TrainerStatus.RUNNING
    537 self.training = True
--> 538 call._call_and_handle_interrupt(
    539     self, self._fit_impl, model, train_dataloaders, val_dataloaders, datamodule, ckpt_path
    540 )
...
        [nan, nan, nan],
        [nan, nan, nan],
        [nan, nan, nan],
        [nan, nan, nan],
        [nan, nan, nan]], grad_fn=<AddmmBackward0>)

Related to #161 (closed)