GrievingTurkey78 , I'm not sure. Let me check.
Do you have cpu/gpu tracking through both pytorch lightning AND ClearML reported in your task?
GrievingTurkey78 , can it be a heavy calculation that takes time? ClearML has a fallback to time instead of iterations if a certain timeout has passed. You can configure it with
CostlyOstrich36 Pytorch lightning exposes the
current_epoch in the trainer, not sure if that is what you mean.
GrievingTurkey78 , do you have iterations stated explicitly somewhere in the script?
I set it to
200000 ! But the problem stems from when the first plot is the clearml cpu and gpu monitoring, were you able to reproduce it? Even if I set the number fairly large when the monitoring plot was reported the message appeared.
GrievingTurkey78 , what timeout did you set? Please note that it's in seconds so it needs to be a fairly large number
GrievingTurkey78 , let me take a look into it 🙂
Hey CostlyOstrich36 I am doing a lot of things before the first plot is reported! Is the
seconds_from_start parameter unbounded? What should I do if it takes a lot of time to report the first plot?
GrievingTurkey78 , the default is 3 minutes. You can try setting to a long enough time to make sure it doesn't skip the epoch 🙂
I set the number to a crazy value and it fails around the same iteration
GrievingTurkey78 , can you try disabling the cpu/gpu detection?
I'll give that a try! Thanks CostlyOstrich36
GrievingTurkey78 , did you try calling
task.set_resource_monitor_iteration_timeout after the task init?
GrievingTurkey78 , please try
Sure! Could you point me out how its done
Oh I think I am wrong! Then it must be the clearml monitoring. Still it fails way before the timer ends.
Last question CostlyOstrich36 sorry to poke you! Seems even though if I set an extremely long time it will still fail when the first plots are reported. The first plots are generated automatically by pytorch lightning and track the cpu and gpu usage. Do you think this could be the cause? or should it also detect the iteration.
CostlyOstrich36 That seemed to do the job! No message after the first epoch, with the caveat of losing resource monitoring. Any idea of what could be causing this? If the resource monitor is the first plot then the iteration detection will fail? Are there any hacks to keep the resource monitoring? Thanks a lot! 🙌