CatBoostRegressor: How to assign weights to response variables?

[fokoa]

I'm working on a Multigression problem. Indeed, I have 13 response variables to predict. The problem statement tells us the weight of each response variable. When I train my Catboost model without considering the weights, I get a bad score. So, I would like to train my CatBoostRegressor() with the weights of the response variables: I would like its equivalent to CatBoostClassifier(auto_class_weights=).

I leaned towards a custom objective function because my model will have to seek the best parameters while taking into account the weights of the response variables.
This link was a bit helpful for me (Section 5. Multiregression): https://github.com/catboost/catboost/blob/master/catboost/tutorials/custom_loss/custom_loss_and_metric_tutorial.ipynb

Unfortunately, I cannot assign my weights to the "weights" parameter.

Please, can someone have a solution to my problem? How to assign weights to response variables at CatBoostRegressor?

Thank you!