Always same evaluation results

#1

We submitted our numpy-based codes several times, changing the function numpy/train_numpy.py.
However, every evaluation results in the leaderboard show the very similar values, fcatorVAE = 0.0, sum_of_all_scores = 0.747-0.749, even if the function train_numpy.py is completely different.

Did I make mistakes in the submission process?

#2

The leaderboard shows your best submission not the latest one.
You can see the score for each evaluation in your gitlab issues.
Hope this helps!

Let me know if you have any queries.

Best of Luck!

#3

Of course, I have checked my gitlab issues.
But the score for each evaluation in my gitlab issues are always the same value.
If I submit a code which output all zero matrix, if I submit a code which output random matrix,
the system always shows the same value.

#4

Hi @komattasan,

I see the latest commit (80edc707c0a6b821cbd346ce0ec25301cc989e2f) on your repository hasn’t changed since 2 weeks. Along with it all the submissions/tags have been made using the same hash as well.

Are you sure you add adding your updated train_numpy.py as a commit before resubmitting solution i.e. new tag push? If not, please do it like:

git add <file-changed>  # train_numpy.py in this example
git commit -m "Your commit message"
git push origin master
#5

Thank you for your reply. I did not check the repo page itself.

However, I’m using that command …
It is surprising that there is no change in my repo.
Maybe, my environment has a problem, in a network proxy or something.
I will try again.