In this project the text of some books are chosen to train an LSTM network and comparison of some different configurations are considered to indicated the most suitable one for this task. Then in the next part, the best network will be train by 4 different books, and at each step, the model will generate sentences which we hope to be aligned with the author's wirting style.