Research on Music Generation Based on Transformer

Music generation refers to the use of computers to create music through certain algorithms or processes with minimal human intervention, and is an important research direction in the field of artificial intelligence. However, music generation is an unsolved problem facing many challenges, and how to generate high-quality music samples has always been a research hotspot, for this reason, this paper proposes a music generation model based on improved Transformer. Specifically, this paper firstly preprocesses the music data and converts the music data into sequences for the model to process and learn. Secondly, a gated recurrent unit (GRU) is introduced to improve the Transformer, and the feature extraction capability of the Transformer to capture global dependencies and the time series modelling capability of the GRU are used to better understand the contextual relationships of the sequence data, and to generate high-quality music that is closer to real music. Finally, experimental verification of the music generated by the proposed method is carried out in this paper, and the results prove the effectiveness of the method.