# 155th solution

Thank you to organizers and congrats to all participants.
I started working after I found
[@tubotubo](https://www.kaggle.com/tubotubo)
's public notebook & his github repo.
[https://www.kaggle.com/code/tubotubo/cmi-submit](https://www.kaggle.com/code/tubotubo/cmi-submit)
[https://github.com/tubo213/kaggle-child-mind-institute-detect-sleep-states](https://github.com/tubo213/kaggle-child-mind-institute-detect-sleep-states)
First time for me to tackle on Deep learnning techninque instead of GBDT, I learned so many things.
My solution
Basically my solution comes from tubotubo 's public notebook.
Differences are following;
Use top-5 models instead of best model on an epoch.
Ensembling on variety of K-Folds
I found validation score fluctuates on learning, also private data is 75% of test dataset, I decided to ensemble as much as I can.
A Tip tried but not worked
Here is a tip I couldn't find on the discussion. This is not an effective tip. I'd appreciate any feedbacks.
Mean/Standardization on each series_id, instead of whole dataset
I wanted my model to learn the difference btwn awake/sleep instead of each seires_id's difference.
So I applied mean/standardization to anglez/enmo on each series_id, but it didn't work as I expected.