It is obvious that using the train data from 2 and 3 will give a huge boost in task one.
I doubt that 90% accuracy can be achieved with task 1 data only.
I removed my submission that uses extra data.
But it would be nice if everyone did the same so we have a clean leaderboad.
That is correct. We request all participants to remove entries from the leaderboard that involved the usage of task 2/3 data for task 1.
Posted by: shivaniku @ Jan. 30, 2024, 10:30 a.m.