pymc - Re-running a model with new data in pymc3 -
i wondering if there mechanism in pymc3 re-run model new data. after setting model , before sampling, assume pymc3 optimization (and compilation?) of model takes quite time. setup model once, , run long sequence different (independent) data sets through it.
i tried setting model outside loop (defining priors, etc.) , updating likelihood new measurements inside loop (and run sampling inside). estimates, however, not change changing data. hence think model using data provided first.
many , best regards
jan
Comments
Post a Comment