Open fojoshi opened 1 year ago
@fojoshi It would be tricky for a task to know the reason why it had been stopped: That information is available to the logic that controls execution, in the ClearML case - The optimizer task.
When you go over your optimizer's log - Do you not see which tasks were pruned?
Proposal Summary
Make the optuna pruning experiment an informative message
Motivation
While executing the hyper-parameter task with the optimizer set to Optuna, it would be terminating experiments based on the pruning algorithm. However, the status shows
Aborted
but with the message log showing,User aborted: stopping task (3)
which is incoherent. It would be great to get a detailed message about why the experiment was aborted.Related Discussion
If this continues a slack thread, please provide a link to the original slack thread.