Open exalate-issue-sync[bot] opened 1 year ago
Erin LeDell commented: This is caused because there are no models built in the specified time-frame (and when you print the object it tries to collect and print model metrics for the leader). If you look only at the leaderboard, that will print fine (1-row with NAs), but not if you try to print the automl object itself. One way to address this is to forcing that one model is always trained in AutoML even it it exceeds max_runtime_secs
. The other way is to check if the leaderboard is NA and print a message saying "there are no models in the leaderboard."
JIRA Issue Migration Info
Jira Issue: PUBDEV-5866 Assignee: Tomas Fryda Reporter: Megan Kurka State: Open Fix Version: N/A Attachments: N/A Development PRs: N/A
I am seeing errors when using the automl model in R if only one model was built:
{code:R} data.hex = h2o.importFile(path = "https://s3.amazonaws.com/h2o-public-test-data/smalldata/demos/bank-additional-full.csv", destination_frame = "data.hex") myX = 1:20 myY="y" automl <- h2o.automl(x = myX, y = myY, training_frame = data.hex, max_runtime_secs = 1, project_name = "automl")
automl An object of class "H2OAutoML" Slot "project_name": [1] "automl"
Slot "leader": Error in is.H2OFrame(x) : trying to get slot "metrics" from an object of a basic class ("NULL") with no slots {code}