You signed in with another tab or window. Reload to refresh your session.You signed out in another tab or window. Reload to refresh your session.You switched accounts on another tab or window. Reload to refresh your session.Dismiss alert
For part 2 we are asked to have only a train set and a test set and to train from scratch only the best model.
In the first part , we have a train, a cross validation and a test set. Our F1 score are calculated from the cross validation set.
Part 2 then asks to print all the F1 scores calculated from test sets. As the sets are different from part 1, we need to train from scratch all the models to do that...
I think you should remove
For the second script solar_system_census.py, only a train and test
set are necessary as one is simply looking to the performance.
or put back the old version of the subject : asking to print the f1 score of the cross validation set and only the best model on test set
Fixed on:
Github
Gitlab
The text was updated successfully, but these errors were encountered:
For part 2 we are asked to have only a train set and a test set and to train from scratch only the best model.
In the first part , we have a train, a cross validation and a test set. Our F1 score are calculated from the cross validation set.
Part 2 then asks to print all the F1 scores calculated from test sets. As the sets are different from part 1, we need to train from scratch all the models to do that...
I think you should remove
For the second script solar_system_census.py, only a train and test
set are necessary as one is simply looking to the performance.
or put back the old version of the subject : asking to print the f1 score of the cross validation set and only the best model on test set
Fixed on:
The text was updated successfully, but these errors were encountered: