Skip to content

CatBoost integration guide#

Open in Colab

CatBoost is a high-performance open source library for gradient boosting on decision trees. This guide will show you how to:

  • Upload experiment datasets
  • Upload CatBoost model parameters and attributes
  • Upload training results to Neptune

CatBoost example visualized in a Neptune dashboard

See example in Neptune  Code examples 

Before you start#

  • Sign up at
  • Create a project for storing your metadata.

  • Have Neptune and CatBoost installed.

    To follow the example, you will also need to install ipython, ipywidgets, and scikit-learn.

    pip install -U catboost neptune ipython ipywidgets scikit-learn
    pip install --user -U scikit-learn
    conda install -c conda-forge catboost neptune ipywidgets scikit-learn
Passing your Neptune credentials

Once you've registered and created a project, set your Neptune API token and full project name to the NEPTUNE_API_TOKEN and NEPTUNE_PROJECT environment variables, respectively.

export NEPTUNE_API_TOKEN="h0dHBzOi8aHR0cHM.4kl0jvYh3Kb8...6Lc"

To find your API token: In the bottom-left corner of the Neptune app, expand the user menu and select Get my API token.

export NEPTUNE_PROJECT="ml-team/classification"

Your full project name has the form workspace-name/project-name. You can copy it from the project settings: Click the menu in the top-right → Edit project details.

On Windows, navigate to SettingsEdit the system environment variables, or enter the following in Command Prompt: setx SOME_NEPTUNE_VARIABLE 'some-value'

While it's not recommended especially for the API token, you can also pass your credentials in the code when initializing Neptune.

run = neptune.init_run(
    project="ml-team/classification",  # your full project name here
    api_token="h0dHBzOi8aHR0cHM6Lkc78ghs74kl0jvYh...3Kb8",  # your API token here

For more help, see Set Neptune credentials.

Start a run#

import neptune

run = neptune.init_run() # (1)!
  1. If you haven't set up your credentials, you can log anonymously:


We'll use the run object we just created to log metadata. You'll see the metadata appear in the app.

Log training metadata#

from catboost import CatBoostClassifier

model = CatBoostClassifier()

plot_file = "training_plot.html"
    eval_set=(X_eval, y_eval),

The training plot can be uploaded as an interactive plot file:


You can log the training metrics to Neptune using the = operator.

from neptune.utils import stringify_unsupported

run["training/best_score"] = stringify_unsupported(model.get_best_score())
run["training/best_iteration"] = stringify_unsupported(model.get_best_iteration())

You can save and upload your predictions to Neptune as a CSV file using the to_csv() method to view them in an interactive table.

titanic_test.to_csv("results.csv", index=False)



For more information on logging and visualizing DataFrames, see How to use Neptune with pandas.

Upload model metadata to Neptune#

Here's how you can upload the model binary to Neptune:



To upload model attributes, you can use the = operator to assign particular values and dictionaries to specific fields:

run["model/attributes/tree_count"] = model.tree_count_
run["model/attributes/feature_importances"] = dict(
    zip(model.feature_names_, model.get_feature_importance())
run["model/attributes/probability_threshold"] = model.get_probability_threshold()

The following method will fetch the model parameters and upload them to the "model/parameters" namespace:

run["model/parameters"] = stringify_unsupported(model.get_all_params())

To stop the connection to Neptune and sync all data, call the stop() method:


You can check your results in the following namespaces, located in the All metadata tab of the run in the Neptune app:

  • model namespace
  • training namespace
  • data namespace

See in Neptune