abstract |
The invention discloses a method for generating melody MIDI accompaniment based on a deep neural network, and belongs to the technical fields of artificial intelligence and music technology. The melody MIDI accompaniment generation method includes collecting a MIDI data set with genre tags, processing the MIDI data set to obtain a data set, encoding it through MuMIDI, obtaining lexical entries, and inputting the lexical entries into a GC‑Transformer model for training until the total loss The function converges, and the training of the GC-Transformer model is completed. Finally, the MIDI clips containing only the melody are encoded and input into the trained GC-Transformer model, and finally the MIDI clips with good accompaniment are output. The melody MIDI accompaniment generation method of the present invention has many advantages, such as short generation time, high generation quality, less hardware resource consumption, less training data, and the like. |