Consider this scenario: After defining a learner and starting learn.fit_one_cycle for 10 epochs, I decide to stop fitting in the middle of epoch 5.
Is there a way to use this 4 and 1/2 epochs that had my learner fitting? If the last unfinished epoch is the problem of recycling this fitting is there a way to remove this incomplete epoch and use the 4 complete epochs in a future fit?
Thanks