What are the steps to refresh the previously selected model?

[Snowcatcat]

Can someone let me know the steps to refresh a previously selected model? I ran screenshot 1 first, but I am getting errors.

I am getting this error: object 'parallel_processing' not found

I also see this piece of the code, and should I run this after I run the code in screenshot 1?

[kyletgoldberg]

If you have the Robyn (RDS) object saved it will be easiest to use the code in screenshot 3. You don't need to run the code in screenshot 1 prior to running screenshot 3 either, screenshot 1's code is just for the case that you don't have the RDS object.

[kyletgoldberg]

That is expected yes, this is not the RDS object though. To generate that you need to run the robyn_save function which will force you to select one model ID. All of these folders are generated by the robyn_outputs. In the demo.R file, check the portion of the code with robyn_save which is what saves down the .RDS objects for future use.

[Snowcatcat]

Thanks @kyletgoldberg . This is the code I run:

But if I close out R, and open R again and rerun the code, it will just produce a new init folder with different model results. The Robyn_save code above will return an error.

[laresbernardo]

In the last section of the demo.R file you'll see an example to import your previous hyper-parameters, train the model and generate that OutputCollect object. The InputCollect object should be the same you ran when you first ran the code.

dt_hyper_fixed <- data.table::fread("~/Desktop/2022-02-21 11.29 rf11/pareto_hyperparameters.csv")
select_model <- "1_51_11" # your selected model ID
dt_hyper_fixed <- dt_hyper_fixed[solID == select_model]

OutputCollect <- robyn_run(
  # InputCollect must be provided by robyn_inputs with same dataset and parameters as before
  InputCollect = InputCollect,
  dt_hyper_fixed = dt_hyper_fixed)

[Snowcatcat]

Thanks! I followed the steps that you showed above and getting this error. Could you let me know what I possibly did wrong?

[kyletgoldberg]

Does your dt_hyper_fixed object have a lambda hyperparameter?