This is the official code for the paper Learning with Restricted Boltzmann Machines: Asymptotics of AMP and GD in High Dimensions.

We provide the implementation of AMP-RBM, GD together with their SE. We also provide the implementation of contrast divergence (CD) on spiked covariance model. In the Fashion MNIST folder, we implement GD and CD to train RBMs. The implementation of CD is adapted from https://github.com/mr-easy/Restricted-Boltzmann-Machine/.