Is it possible to reuse weights from a model with different dimensions?

I’ve played around with copying weights from a pretrained model into a new model in order to change the maximum sequence length, and that has saved me lots of time compared to starting from scratch.

But is it possible to reuse weights between models with different numbers of layers and dimensions (e.g. bart-base and bart-large)? I’m assuming the answer is no, but I just wanted to double-check.