Problem when generator for tf.keras.fit returns more than one element

Question

Problem Description

I\'ve written an generator which tokenizes and pads text sequences and want to use it in Model.fit(). This works fine if i instruct t