admin管理员组

文章数量:1355564

I use mlflow and hyperopt for tuning a model, and trying to figure out hyperopt sampling methods.

I have directly used line of codes from the documentation, as such: my code:

space = {"batch_size": hp.quniformint("batch_size", 16, 256),}

but I get the error:

AttributeError: module 'hyperopt.hp' has no attribute 'quniformint'

Otherwise, my code works with sampling implementation like hp.choice.

Did anyone came across this error? I know that the hyperopt documentation has not always been updated, that could explain this error. Otherwise I basically want to sample uniformly descrete values between 16 and 256 with increment of 16. How else could I do that?

本文标签: hyperparametersHyperopt attribute 39quniformint39 not recognizedStack Overflow