Is it possible to set a separate folding factor for different rounds?

Motivation: for each round, the prover needs to open 1 << folding_factor evaluations for each shift query. For a folding factor of 4, that means 16 evaluations for each query. This doesn't incur a large proof size for the intermediate rounds. However, for the first round, the leaf size can easily become large when multiple polynomials are opened in batch. In that case, it may be beneficial to start with a small folding factor in the first round, and switch to a larger one later.