This is a companion discussion topic for the original entry at https://dphi.tech/practice/challenge/57
This is a companion discussion topic for the original entry at https://dphi.tech/practice/challenge/57
Hi all/moderators,
The score of 100 is attained by following the paper suggested in the data section (https://www.sciencedirect.com/science/article/pii/S1319157817300149)
If we go down at the bottom of the paper we get the corresponding github repo that already has the dataset:- https://github.com/urwithajit9/clamp
I intended to use this data as extra training data but after 1-2 submissions realized the dataset mentioned over here has the test data as well. Hence my model is training on this data as well and giving a score of 100.
If possible remove my submission as it was unintentional from my end and came as a result of modelling bias.
Hey,
Someone who has scored 98+, can you pls share the notebook once datathon is over.
thanks.
@akashpb thank you for keeping us informed. We will remove the submission for now, could you please make the submission that you got while building the model on the training dataset?