R notebook: “Error in load_model(model_output_path) : unused argument (model_output_path)”

Insurance pricing game
#1

Hi all

For those of you using the R notebook, if you encounter this error, please first:

  1. Either restart your notebook to factory runtime (Runtime -> Factory rest runtime) or
  2. Simply visit the original template and copy that into your own drive

That should fix the issue. Note that this change is to allow the load_model functions more flexibility beyond just using the .RData format.

Happy pricing :money_with_wings:
Ali

1 Like
Inference failed errors
Inference failed issue (model not found issue)
#2

@alfarzan, thank you!
Your suggestions actually solved the problem with “Error in load_model() […]”, but unfortunately my submission still ends up with the error message “Inference failed errors”.
In the log there is a detailed error description and there is still a reference to LoadModel function. Thus I presume that the issue is quite similar to that you described above, isn’t it?
Do you have any other ideas? My latest submission is 114872. Thanks in advance!

#3

Hi @imfilip

This one seems to be a xgboost specifcic error. For those models often they should be saved in a way according to the package itself.

See this discussion on how to do that. Hope that clarifies things a bit :slight_smile:

#4

Uuu @alfarzan, that was a devilish change! Now, thanks to this post, I am back in play after ten failures :wink:

1 Like
#5

Hah! Yes sorry about it, I understand it must be a bit annoying.

Had to do it to allow people on notebooks to submit things other than .RData files :slight_smile:

1 Like