multinomial_allocation# bayesflow.utils.keras_utils.multinomial_allocation(weights: Mapping[str, float], num_samples: int, seed=None) → dict[str, int][source]# Allocate num_samples draws across weights via multinomial sampling.
