C4 Wk 3, Lab 1: no tuples (features, labels) are created!?

Course Q&A TensorFlow Developer Professional Sequences, Time Series and Prediction
1

In the function “model_forecast()” we don’t create tuples with features and labels like we did in the function “windowed_dataset()”. Why?

# Create tuples with features and labels 
dataset = dataset.map(lambda window: (window[:-1], window[-1]))
2

The function model_forecast is used to prepare input to the model to make the prediction. So, there’s no outputs here. On the other hand, function windowed_dataset is used to prepare (inputs, outputs) pairs for training the model.

3

Thanks, Balaji! I see now that the function model_forecast() is calling the “model” , which means that we are using the weights that had already been trained in the previous steps.