November 2019
Intermediate to advanced
346 pages
9h 36m
English
We begin by importing PyTorch, the neural network model, and math for some basic computations (step 1). More interestingly, in step 2, we specify content and style audio. In the content file, you can utter whatever phrase you wish, for example, you can't do cybersecurity without machine learning. Then, in the style file, you select a recording of someone's voice, for example, a recording of a famous individual such as Elon Musk. The final result of the voice impersonation is that Elon Musk says that you can't do cybersecurity without machine learning. Steps 3, 4, and 5 involve some legwork to prepare our data to be fed into our model and then instantiate a Random CNN model and its optimizer. The main feature of the model is ...