Hi,
After my model's iterations ended and task turned to Completed I wanted to still train the model further for more iterations. I found there is continue_last_task argument which allows to do that, so I've just run the new task with this argument equal to the ID of the task I wanted to continue. However I think I misunderstood or misused this argument since my training loss suddenly jumped after the continuation (first task iterations ended at 39). Could you please advise how to achieve continuation of the task, so the behavior is just as if I have had put bigger amount of iterations since the beginning?