Is there a way to use logprob.conditional_logp#? I am thinking maybe I add something like 1000*Potential(conditional_logp(X_t)) to the model where 1000 is the weight to the conditional logp but conditional_logp takes value variable.
Is there a way to use logprob.conditional_logp#? I am thinking maybe I add something like 1000*Potential(conditional_logp(X_t)) to the model where 1000 is the weight to the conditional logp but conditional_logp takes value variable.