I am really confused about the final results on the leaderboard for Broad1. It seems to be run on a random submission, and not my final submission. How is the selection chosen?
1 Like
The final prediction is taken from the run you selected.
If you did not selected anything, it will be the latest valid run.
I dont remember selecting anything in the old interface. It also didnt take my latest run. The final prediction on one of my model(model2) used a bad prediction but its not possible to select a run if its bad.
From what I have in our database, you made your first selection the 2025-01-24 at 00:30:30 UTC, and updated it for the last time the 2025-01-26 at 15:50:08 UTC.
You likely selected a run and forgot about it.
Thats unfortunate. I might have misclicked it then. Its probably too late to do anything. Appreciate the swift reply.
Ps: Is there a way to see the score of my latest model on the test set? (Not to participate but just to see its performance).
Yes that is too late.
I can compute it manually if you want.
Please give me the run you have in mind.
Can I check submission 14 of my model2 which is the last submission if it’s possible. It’s ok if it’s too much trouble. Thanks a lot.
For prediction #27131, you would have gotten:
- MSE: 0.113
- Spearman: 41.999
And would have ended up 6th (instead of 7th) on the leaderboard.
Thanks! Appreciate the help.
In Broad Challenge 1, there is one winner who appears twice. Will this be revised?
The broad 1 leaderboard is now frozen, but it will not happen anymore for the next competitions.
See this announcement.