We noticed that few nodes are getting provisioned with slower processing since morning, causing few subparts to end up with timeout. Ex, 1-3 in your submission.
We will be manually re-running 1-3 subsets as soon as your submission finishes for now.
Meanwhile, we are checking if anything changed in the provision scripts.
Sorry for the trouble.
ok，I believe you can solve it well
Now i have submitted a new one
Are they both the same code/model-wise?
In that case, re-running your 1-3 would be quicker than the new submission and we can stop the new one.
Edit: I noticed they are same. I am cleaning it up right now and you should have results soon in your older one.
yes，they are the same，subset 8 is not completed, it also needs to be run again