when it report ' Number of crashed target algorithm runs:1',what is the specific reason that causes an algorithm to crash

[belzheng]

when it report ' Number of crashed target algorithm runs:1',what is the specific reason that causes an algorithm to crash?

[eddiebergman]

Hi @belzheng,

There's no a great interface for this at the moment but you can get access to most of this information by using

askl = AutoSklearnClassifier(...)
askl.fit(...)

# A dictionary where the configuration are as keys and any information we have on them as values
askl.automl_.runhistory_.data

Best, Eddie