Text this: A deep learning model with features embedding for style-conditioned music generation /