Properties of time series: stationary and non-stationary

I wonder that can the model such as rnn and lstm handle any kind of time series? I mean can it handle both stationary data and non-stationary data or do we need to do something before feed into the rnn and lstm model?

You can build a time-series model with any layers you want. Please read this topic as well.

In the assignments, you’ll explore recurrent and convolutional methods of building models for time series data.