I've encountered an issue that I think is tied to the collect_metrics() function. I am manually assigning the rows in my dataframe into different folds for cross-validation of the lambda hyperparameter of lasso. After tuning with my custom 'grid_folds', collect_metrics shows all NAs for the metric I am using (roc_auc) to select the best lambda hyperparameter. I think select_best() fails because of this.
This issue came up after I updated my tidymodels software on my local machine today. However, I can run my analyses on my university HPC without any issues (which has the older software).
Below I've provided a reproducible example of my issue. I ran it on both my local machine and on my HPC. The last step to select the best lambda does not run on my local machine, but it does run on my HPC. I've included session info for both my local machine and my HPC.
This issue has been automatically locked. If you believe you have found a related problem, please file a new issue (with a reprex: https://reprex.tidyverse.org) and link to this issue.
The problem
I've encountered an issue that I think is tied to the collect_metrics() function. I am manually assigning the rows in my dataframe into different folds for cross-validation of the lambda hyperparameter of lasso. After tuning with my custom 'grid_folds', collect_metrics shows all NAs for the metric I am using (roc_auc) to select the best lambda hyperparameter. I think select_best() fails because of this.
This issue came up after I updated my tidymodels software on my local machine today. However, I can run my analyses on my university HPC without any issues (which has the older software).
Below I've provided a reproducible example of my issue. I ran it on both my local machine and on my HPC. The last step to select the best lambda does not run on my local machine, but it does run on my HPC. I've included session info for both my local machine and my HPC.
thank you!!
josh
Reproducible example