Pytorch and TensorFlow for audio processing