Sentiment Classification, where you have comments written below a You Tube video as the input to you network and the RNN would need to decide if the comment is positive or negative.

Your input to the RNN model is a series of musical notes, and the output is which notes' frequency is $532$Hertz.

Music generation, where you select a genre and the output of RNN is a music notes generated.

You are giving different sentences to your model. The model is supposed to recognize the verbs among all the words.

You have a text in German language and your RNN model would translate it into French language.

You have a series of images as the input to the RNN model, and the model should generated a text representing what the action is$.$