> Hunger for data

Dear organizers, I have a request. Can you somehow really increase the amount of data for train and test sets. May be 10-100X times.
It's not so interesting when some participants scored exactly f1_score = 1.00 on updated data in one day.

Posted by: kuzhamuratov @ Sept. 22, 2020, 8:23 a.m.

Dear participant!

Thank you for your request!

Yesterday we have made an update on the data: new sources are added both to the test and train data

Wish you luck!

Posted by: Rybolos @ Sept. 22, 2020, 9:48 a.m.

Yea, I see it. However, it didn't change the picture. Previous dataset ~ 20_000 and the current ~23_900.
In my opinion, it doesn't solve the problem organizers tried to solve.

Posted by: kuzhamuratov @ Sept. 22, 2020, 10 a.m.
Post in this thread