Why my rank is little although my MAE=0.128 is less than the top rank
share your model 😂. your validation has a leak
You need my module why?
they say on my submission is Missing entries for IDs
but i review my predictions.csv has not any missing entries for IDs
Can't see your submission on the leaderboard mate
My rank is the below of rank list
There are no submissions on your Zindi account my friend
But,I submitted my submission
You are likely overfitting to your validation data or there could be something wrong about the way you are computing the MAE score .
Firstly,I make tow MAE score for two target variables solar panel and water heater, and I take the MAE average.
mae_total=((mae_panel + mae_boil)/2)
You are probabily overfitting on validation data. Try to visually count some of the panels and boilers in the test set and see if your model predicts same. Especially for those with higher counts from sat images.
git gud
share your model 😂. your validation has a leak
You need my module why?
they say on my submission is Missing entries for IDs
but i review my predictions.csv has not any missing entries for IDs
Can't see your submission on the leaderboard mate
My rank is the below of rank list
There are no submissions on your Zindi account my friend
But,I submitted my submission
You are likely overfitting to your validation data or there could be something wrong about the way you are computing the MAE score .
Firstly,I make tow MAE score for two target variables solar panel and water heater, and I take the MAE average.
mae_total=((mae_panel + mae_boil)/2)
You are probabily overfitting on validation data. Try to visually count some of the panels and boilers in the test set and see if your model predicts same. Especially for those with higher counts from sat images.
git gud