The so-called trivializing flows were proposed to speed up Hybrid Monte Carlo simulations, where the Wilson flow was used as an approximation of a trivializing map, a transformation of the gauge fields which trivializes the theory. It was shown that the scaling of the computational costs towards the continuum did not change with respect to HMC. The introduction of machine learning tecniques, especially normalizing flows, for the sampling of lattice gauge theories has shed some hope on solving topology freezing in lattice QCD simulations. In this talk I will present our work in a $\phi^{4}$ theory using normalizing flows as trivializing flows (given its similarity with the idea of a trivializing map), training from a trivial distribution as well as from coarser lattices, and study its scaling towards the continuum, comparing it with standard HMC.