-
|
Moving here from stack-overflow. Under the hood Scikit-Learn components do a lot of input validation checks for (X,y) shapes, expected scalars/arrays types, ensuring estimator are already fitted, etc. When designing models these checks are extremely useful to catch early bugs, but these checks become runtime overheads in large mature pipelines once in production. Is there currently any way, maybe via a global setting, passing in special keyword arguments, or similar to disable these internal checks within pipelines, estimators & transformers? |
Beta Was this translation helpful? Give feedback.
Replies: 1 comment 3 replies
-
|
You can disable some of the validation checks globally using Alternatively, you can disable those checks locally using |
Beta Was this translation helpful? Give feedback.
-
|
FYI, the following issue has been opened: #21804 |
Beta Was this translation helpful? Give feedback.
-
|
@jjerphan good to know about the |
Beta Was this translation helpful? Give feedback.
-
|
Enjoy! Feel free to participate to the conversations in #21804, if you are interested. |
Beta Was this translation helpful? Give feedback.
You can disable some of the validation checks globally using
sklearn.set_config.Alternatively, you can disable those checks locally using
sklearn.config_context.