I'm working on time series analysis, interested in using gluonts. I'm following the tutorial provided at the link below. I'm trying to understand what is going on under the hood so I can understand how or if it is preprocessing & normalizing the data and what the model architecture is so I can compare it with other similar time series models. In the tutorial, I see that the data they are using ranges from 400 to 900, so I'm wondering if somewhere inside the model the data are being normalized before training, or is it just training on the raw numbers? Does anyone know if there is a reference somewhere for the general architecture of this model as well as the other gluonts models?
https://ts.gluon.ai/stable/tutorials/forecasting/quick_start_tutorial.html