It should be possible to re-use the seeds from a previous classification run and not just from a previous classification step.

Currently it is only possible to use the seeds from the previous run. We should simply be able to say use previous seeds for each step. It is not obvious how the seeds are saved into the Run1 R object, plus they are saved as an object plus name, and not as a serialized object which can be easily loaded and assigned to any variable.

We think the seeds object should be saved into an independent file, one per step (such as seeds_1), and it should saved using the saveRDS function. Furthermore, these files should then be loaded from any location and not just the current output location so the user can repeat sampling using seeds from other runs.