Ensemble of different models?

How do we combine different models? Does the training will save only one model.joblib or can we create model1.joblib, model2.joblib, model3.joblib etc…I tried doing this and even I loaded all of these inference for ensembling —but it is throwing error in submission.

“Essential container in task exited with exit code 1” is the error.

1 Like

This is a generic error telling you that your code crashed.
You need to look at the trace for more info. Maybe if you could share your Run ID, I could take a look.

(Runs are private, its fine to share IDs)

1 Like

The latest run id is #165

Sorry for the delay.

The error is:

_catboost.CatBoostError: /src/catboost/catboost/libs/train_lib/dir_helper.cpp:20: Can't create train working dir: catboost_info

Here is a potential fix:

EDIT: Add link to the documentation

I dont understand how to use above fix.

I have added the above code just after model.fit (). Here is the error am getting locally.

CatBoostError: You can’t change params of fitted model.

let me know if you want me to submit the code in the server. you can correct that code for me please.

What happen if you put it before?