Hey, thanks for trying to reproduce the error, the unreproducibilty of it means it's probably related to now models being found for the dataset in that time limit.

I would honestly advocate for the simplest solutions:

fit_ensemble gives a ValueError with ensemble_size = 0.
As for show_models(), I'm currently on my phone and can't easily check, but I recall it only shows Ensemble members. It would be good to extend this to all models but I think for simplicity sake, a warning and an empty dict will suffice.

I have been working on a generic backend for auto-sklearn and auto-pytorch at automl_common which should help with model and ensemble interaction. This is in the pipeline for the end of this month so I wouldn't spend to much time on perfecting this PR.

AutoML::fit_ensemble with ensemble_size =0 causes crash #1365

Description

Metadata

Assignees

Labels

Type

Projects

Milestone

Relationships

Development

Issue actions

`AutoML::fit_ensemble` with `ensemble_size =0` causes crash #1365