This project focuses mainly on generating new music notes with the help of two deep learning models viz., LSTM and GAN. Humans can recognize music patterns just by hearing and seeing. They need not be experts in the musical field to do this. Hence, they can generate new music notes using the patterns they recognized. This project targets to do the same. Two models – LSTM with attention and GAN were used to recognize and learn a bunch of musical sets (instruments music). With the patterns they learned, the model was trained to produce a fresh piece of music that was not the same as the ones they got trained on. The LSTM model was designed to process the music with chords like the ones that are produced from violin, guitar, or cello. The GAN was designed to process any type of music except the ones with a lyrical component. Both of these models were able to produce fresh pieces of music which the human ears can consider as a piece of good music. A large musical company or any music composer might find this project intriguing to compose new variations of other compositions.