cuqi.sampler.NUTS.tune# NUTS.tune(skip_len, update_count)[source]# adapt epsilon during burn-in using dual averaging
