I started implementing some functions to train a DBN based on your RBM implementations. It's kind of a hack right now but I think the API can be improved and the functions can be generalized a little bit.
My repository is here in case you want to take a look. I actually have an updated version for which only the first RBM is Gaussian-Bernoulli and all the others are Bernoulli, but I still did not push it to GitHub. If you think this is a nice fit for this library, I will start implementing it in a branch of Boltzmann.jl instead of using my own library.
Hi,
I started implementing some functions to train a DBN based on your RBM implementations. It's kind of a hack right now but I think the API can be improved and the functions can be generalized a little bit.
My repository is here in case you want to take a look. I actually have an updated version for which only the first RBM is Gaussian-Bernoulli and all the others are Bernoulli, but I still did not push it to GitHub. If you think this is a nice fit for this library, I will start implementing it in a branch of Boltzmann.jl instead of using my own library.