Thank you for sharing ! Your step 5 is particularly interesting as I spent a lot of time trying to make something meaningful to propose 5 different predictions.
I am too afraid of a leaderboard shake up to share my approach at the moment: the test set seems small as I observed large differences between my local CV and LB results so far…
I was wondering about the process for the final model selection. Will the latest model be used ? Or will the best model on the (available) test set will be used ? Or the best model on the whole test set (in which case there is a strong incentive to submit as many models as possible) ?