New issue
Have a question about this project? Sign up for a free GitHub account to open an issue and contact its maintainers and the community.
By clicking “Sign up for GitHub”, you agree to our terms of service and privacy statement. We’ll occasionally send you account related emails.
Already on GitHub? Sign in to your account
[Changed behavior] n_iter_no_change should be attached with early_stopping, not model #19743
Comments
There is no bug here. Basically if
While if activating
@IchiruTake did I miss something in which case do not hesitate to correct me. |
If I did not miss anything (and that there is no regression) @IchiruTake do you want to submit a PR to improve the documentation? |
This will thus make the definition to be easier to understandI am just a beginner developer, thus not good as fixing some other people's code. I am trying to improve myself.
|
Fixed in #19818 |
I have used Neural Network to validate the result. Surprisingly, n_iter_no_change is attached directly into the model instead, although it followed the docmuentation, but get confused for this hyper-parameter. The data is performed on AND gate. This happens on version 0.22 -> 0.24.
Solution: Changed n_iter_no_change so that this hyper-parameter works only if early_stopping=True
CASE #1:
CASE #2:
Result for Case #1:
Result for Case #2:
Note: The speed compared on Neural Net was better compared to 0.22
Version 0.22.post1: Executing Time: 2.344778s
Version 0.24.1: Executing Time: 2.290114s (2.38 % better)
The text was updated successfully, but these errors were encountered: