Bi-LSTM deep reinforcement learning network-based drum music generation method

The invention discloses a folk drumbeat generation method based on a Bi-LSTM deep reinforcement learning network. The method comprises the following steps: S100, representing folk ancient music as a note sequence; s200, converting the generated note sequence into a note vector set through a coding technology; s300, inputting the note vector set into a Bi-LSTM neural network for training, and generating a Bi-LSTM drumbeat generation network based on a character level; s400, the Bi-LSTM drumbeat generation network based on the character level is trained based on a reinforcement learning Actor-Critic algorithm, and a deep reinforcement learning network based on Bi-LSTM is obtained; and S500, inputting the input note sequence into the Bi-LSTM-based deep reinforcement learning network, and autonomously generating drum music works. According to the method, the generation of high-quality and artistic Xian drum music can be basically realized, and the problem of lack of Xian drum music tracks is relieved. 一种基于Bi-LSTM