A Voice Cloning Method Based on the Improved HiFi-GAN Model

Figure 4

Structure diagram of the generator (K represents the size of the convolution kernel).