Skip to content

add input/output borders files support from python package#656

Merged
arcadia-devtools merged 3 commits into
catboost:masterfrom
necnec:feature/add-input-output-borders-file-to-python-package
Feb 14, 2019
Merged

add input/output borders files support from python package#656
arcadia-devtools merged 3 commits into
catboost:masterfrom
necnec:feature/add-input-output-borders-file-to-python-package

Conversation

@necnec

@necnec necnec commented Feb 2, 2019

Copy link
Copy Markdown
Contributor

I hereby agree to the terms of the CLA available at: https://yandex.ru/legal/cla/?lang=en

During training and prediction, CatBoost splits the range of values of each floating point feature into intervals, and uses these intervals instead of the true values.
Using the same feature borders during training and prediction improves prediction accuracy.
Currently python users can neither save feature borders after training, nor load them during prediction; while this is possible with the command-line CatBoost tool.

Your task is to close this gap.
You need to add input_borders_file and output_borders_file parameters to classes CatBoost* in the CatBoost python package, and pass them via cython to the C++ implementation.

@annaveronika

Copy link
Copy Markdown
Contributor

@arcadia-devtools Ship it!

@annaveronika

Copy link
Copy Markdown
Contributor

please agree to the cla, see exaples in other pr-s

cdef EPredictionType EPredictionType_Probability "EPredictionType::Probability"
cdef EPredictionType EPredictionType_RawFormulaVal "EPredictionType::RawFormulaVal"


Copy link
Copy Markdown
Contributor

Choose a reason for hiding this comment

The reason will be displayed to describe this comment to others. Learn more.

Please undo whitespace change

@arcadia-devtools

Copy link
Copy Markdown
Collaborator

@annaveronika, internal review request created: 701583

@annaveronika

Copy link
Copy Markdown
Contributor

@necnec Could you please resolve the conflicts?

@annaveronika

Copy link
Copy Markdown
Contributor

@necnec Do you plan to proceed with the pr?

@annaveronika annaveronika mentioned this pull request Feb 12, 2019
@Metamorphus

Copy link
Copy Markdown

@necnec Please, fix the conflict, it's very quick. I can do it for you, if you give me write access or I can create my PR with your code, if you don't mind.

@kizill

kizill commented Feb 14, 2019

Copy link
Copy Markdown
Member

@arcadia-devtools Ship it!

1 similar comment
@annaveronika

Copy link
Copy Markdown
Contributor

@arcadia-devtools Ship it!

@arcadia-devtools

Copy link
Copy Markdown
Collaborator

@annaveronika, internal review request created: 724218

@arcadia-devtools
arcadia-devtools merged commit cdf20a9 into catboost:master Feb 14, 2019
@annaveronika

Copy link
Copy Markdown
Contributor

@necnec Thank you very much for contribution!

Sign up for free to join this conversation on GitHub. Already have an account? Sign in to comment

Labels

None yet

Projects

None yet

Development

Successfully merging this pull request may close these issues.

6 participants