Closed petergro-hub closed 5 months ago
Hi petergro-hub, logs have been sent to your email.
Thank you for the logs, is it possible that there is a change in server structure? I have submitted a version where out of the 850 images processed in time, 1/3 were done with a large method & 2/3 with another (shorter). Now the model processed 800 images, 1/20th were done with the large method and 19/20 with the shorter version. The logs show no trace of error or timeout (which is built-in to print to log should anything happen). So I don't quite understand how this is possible. (the short version alone taking only an hour for all samples on your server)
In an effort to circumvent this, is it possible, as stated in #89 , to move up one of the moved submission deadlines ? I'd just submit a final shorter method.
Thank you for the logs, is it possible that there is a change in server structure?
We have checked our servers and have not found any systemic bias -- your submissions have been executed across the servers, and there is overlap of both Time Limit Exceeded as well as successful submissions for any particular server. We have not changed the server structure in any other way.
In an effort to circumvent this, is it possible, as stated in https://github.com/AISG-Technology-Team/AISG-Online-Safety-Challenge-Submission-Guide/issues/89 , to move up one of the moved submission deadlines ? I'd just submit a final shorter method.
Thank you, we have responded in #89.
Thank you for checking
could I please request the logs for 07c4c150-3e3a-4a52-9904-2692c1c9c1cf , I don't quite understand how this could timeout unless the server conditions had changed.
Thank you in advance