Hi, we suspect that some of our approaches for subtask 2 (based on multi-label) are weak for certain traits and one column is all zeros.
When we submit those runs we get the following error.
WARNING: Your kernel does not support swap limit capabilities or the cgroup is not mounted. Memory limited without swap.
Traceback (most recent call last):
File "/tmp/codalab/tmpl9PsMk/run/program/score.py", line 105, in
recall1 = tp / float(tp + fn)
ZeroDivisionError: float division by zero
We are not sure if we would have time to improve our models for task 2 in order to detect some of these traits. However, maybe our models are better with the rest of the traits.
It is possible to modify the program to take this into account?
Kind regards
Posted by: JoseAGD @ May 17, 2022, 8:57 p.m.Hi
We have patched the scoring program so that errors of the type "divide by zero" do not cause an erroneous submission. Can you please try again?
Best
Posted by: hugojair @ May 17, 2022, 10:52 p.m.I have resubmitted the file and all seems to work fine. Thanks very much
I know that the leader board is private. Anyway, it is possible to observe my own score?
Kind regards
Posted by: JoseAGD @ May 18, 2022, 7:02 a.m.I am sorry, we cannot share this info. We will reveal the leaderboard to everyone once the competition is over
Best
Posted by: hugojair @ May 18, 2022, 2:04 p.m.No problem. We will wait!
Posted by: JoseAGD @ May 18, 2022, 2:57 p.m.