Drastically increased memory load

[leoniewgnr]

Describe the bug
When running on Azure ML, the required RAM increases drastically when conditional seasonalities are added to the model. Not tested on local machine

To Reproduce
build basic model and then add something like this

# add conditional seasonality
m.add_seasonality(name='winter', period=1, fourier_order=6, condition_name='winter')
m.add_seasonality(name='summer', period=1, fourier_order=6, condition_name='summer')

Expected behavior
no drastically increase

Candidates for increased memory

neural_prophet/neuralprophet/df_utils.py

Line 46 in f08dcf8

df_copy = df.copy(deep=True)

OR
the for loop in time dataset takes very long and there could also be a lot of memory consumed:

neural_prophet/neuralprophet/time_dataset.py

Line 31 in f08dcf8

self.combined_timedataset = [item for timedataset in timedatasets for item in timedataset]

Update: I am pretty sure that it is line 30 or line 31 because when profiling the memory consumption, memory increases by 20x in those two lines of code
Wihtin this function the main contributor is this:

neural_prophet/neuralprophet/time_dataset.py

Line 263 in f08dcf8

def tabularize_univariate_datetime(