June 9, 2024 WaveRNN inference can be accelerated by using block-sparse weight matrices combined with specialized block-sparse matrix-vector multiplication kernels. In this series of posts, we're going to go through the WaveRNN neural vocoder for audio waveform synthesis, along with a variety of implementation details and commonly used extensions. For a real implementation, check out the gibiansky/wavernn repository.Posts in the Series:IntroReducing Quantization NoiseAutoregressive InferenceOpti...