Can we change (increase) amount of training data in between sessions?

[mikemech83]

As per the title. Just wondering if I can add new samples to a model that's got a few thousand checkpoints into it already. Or if this will break things