Open ParadaCarleton opened 2 years ago
Rejected samples are often very useful for Rao-Blackwellization, or even just for identifying where the sampler is having trouble. Maybe we should have some way to return and store rejected samples?
Rejected samples are often very useful for Rao-Blackwellization, or even just for identifying where the sampler is having trouble. Maybe we should have some way to return and store rejected samples?