Hi
I am trying to test the inference process in the express lane, and i got the following error.
>Your Submission to the Digital Mammography challenge (submission ID 7741066) has failed to complete its scoring phase. >The message is:
>
> inference is not allowed.
Can you help what is the problem? My submission file looks like:
inference=docker.synapse.org/syn7....
Thanks
Dezso
Created by Dezso Ribli riblidezso hi, bruce, thanks for your explanation. but i don't know how that is even meaningful, if one is just 'not be blocked for the entire 5 weeks'......
now that you have empty nodes, can you please just leave me alone on an empty one? i think that is how a 'queue' typically should work. i found i always have to wait for days for something to run, **very often over 5 days. and in the end, i got it run because i ask on the forum.** otherwise, it is not even predictable how long it even would be. in fact, i only had it run once on the cluster without complaining for long waiting. **or you can tell me how long i am expected to wait, 5 days, 10 days, 15 days or forever?** because if it is 15 days, then during the waiting i can do a different project which is more meaningful.
i don't think it is behaving normally. and i don't think it is happening for everyone, because there are people who submitted 3 times and received results in a day. there are people complaining for just waiting for 7 hours. i have to wait for days, every time. > can i ask my jobs to be allocated to another node?
In general the answer is 'no'. However, at this early phase there are unoccupied machines so we disassociated your account from its current node allowing it to be associated with, and run on, an unused node. As the challenge progresses, each node will have a list of associated accounts and jobs will have to wait for others to complete before they can run.
> clearly, if this person makes another submission, then i would miss the whole 1st phase.
That's not true. The policies we put in place are meant to ensure that if you submit a job early in a five week round you will not be blocked for the entire five weeks by other jobs on your assigned node.
You should see that submission 7777770 is running and 7791755 is queued to run afterward. ok. thanks. i waited for now 2 days and my job didn't even start, and so i submitted another one, which never appear on the submission board, i guess it is because it is the second one in queue, so it never appeared. i think i am sharing the node with someone who runs a single job for 3-10 days based on the open period. if so, can i ask my jobs to be allocated to another node? clearly, if this person makes another submission, then i would miss the whole 1st phase.
thanks All the past exams of a subject are healthy in the training and SC2 datasets. The case that you describe may however occur in the Pilot Data, which we have release month ago and for which we decided to provide all the data that we have for a set of subjects. there is a person in pilot who is first time cancer and second time not, are we supposed to label 1 or 0 ?
thanks sc1's cross-walk file is about half the size of sc2's since sc1 only has one exam per subject while sc2 has an exam sequence. is the sc1 cross talk or walk file much shorter than the sc2? since most images are gone?
thanks We have corrected the submission pipeline. Would you please submit again and say how it goes?