Only schedule 1 cell in Seeq Data Lab

[Kato]

I have a model that can predict a variable based on various different parameters. The first part of my Seeq Data Lab project builds the model, and the second part applies the model on live data and pushes the results back to Seeq. How do I schedule the cells with the model application to run every minute, but let the model building cells run every month?

[Patrick]

Hi Kato - Is there a reason you would not want to split this into two separate projects and schedule them independently? 

For example, you could store the model training output in a pickle df.to_pickle('filename.pkl') and read it in your model execution script pd.read_pickle('filename.pkl'). 

Every month, the training script would be scheduled to update the .pkl file with the new parameters.

The model execution notebook would be scheduled on a higher frequency using the parameters in the .pkl file.

 

[Kato]

This worked perfectly! Thank you so much.