Hi,
I uploaded a docker for performing validation but I get the following error:"There must at least be one confidence value". This can mean many things, as any error encountered would lead to an empty predictions.tsv file. I have tested this script locally with the pilot data and it works perfectly.
Any idea how can I determine the cause of the error?
Another question would be related to training express lane. Could you please check if everything works on your side. I submitted two models (5-6 hours ago) to test the content (e.g. how training images are stored since the preprocessing script works locally but not on your side) but I can't see my submissions in the Express Lane Dashboard.
Many thanks,
Best,
Anamaria
Created by Anamaria Vizitiu Anamaria I would really suggest you to increase the capacity of your server. Iteration is extremely important for machine learning. I'm ready to give up on Round 1 now. It's mainly because my progress was significantly delayed by docker image configuration and the unavailability of the server. @thomas.yu
Dear Anamaria, and Li,
Due to the the end of the round nearing, there is always an influx of people submitting during this time. Due to this reason, there is a backlog of submissions to both the express lanes and the non express lanes. Apologies for the delay and thank you for your patience.
Best,
Thomas I had the same issue with express lane. I submitted a job 8 hours ago and it hasn't generated any log files yet. Another job two hours ago and it is not even in the dashboard.