An example of a modern Transformer that uses the encoder decoder structure is the T$5$ model from Google. This model reframes a range of tasks into a text$-$to$-$text framework, including translation, linguistic acceptability, sentence similarity, and document summarization. For the project submission, submit the following:

$1.$ See T$5$ transformer model description via image and examples from the book and explain its major components and its workings.

$2.$ Create and explain your own example of cross$-$referential attention in style that would T$5$ use.

$3.$ Based on working of T$5,$ how would you design a generative model that is able to create a short video given only a text prompt as input.