IL-flOw: Imitation Learning from Observation using Normalizing Flows