Some questions about m_V

[xiaobaixiaobaiaa]

When I read your code, I have some doubts, I hope you can answer. Thank you vary much! 1、I noticed that you initialize m_V as the mean of latent variables. 2、While training on m_V, I noticed that you performed some operations on m_V and m_theta What is the connection between m_V and m_theta? What is your purpose for these operations? Can you answer in detail? And, if I don't use a variational encoder, then I generate z directly. Can I use this method to train when training m_V?