In order to track the generality of the RBM during training, and to potentially perform cross-validation for on-line parameter updates, it would be nice to be able to pass a validation dataset to the fit() function as an optional keyword.
If the argument is passed, then the fit function should perform, potentially, per-epoch checks on the validation set (e.g. pseudo-likelihood, reconstruction error).
In order to track the generality of the RBM during training, and to potentially perform cross-validation for on-line parameter updates, it would be nice to be able to pass a validation dataset to the
fit()
function as an optional keyword.If the argument is passed, then the fit function should perform, potentially, per-epoch checks on the validation set (e.g. pseudo-likelihood, reconstruction error).