OSS Tanbou
View on GitHubProject website

train classification, regression, and ranking models with gradient-boosted decision trees that handle numerical and categorical features

OSS scale score 272.9OSS health 79Update pace 112 commits/30d
About these scores

OSS scale score is an unbounded metric that log-compresses and weights Stars, Watchers, Forks, and Contributors. Discovery score is the current OSS scale score minus the score at discovery. Update pace is commits in the last 30 days, growth momentum is the OSS scale score difference within the recent observation window, and OSS health is a 0–100 rating based on available recency, Community Health, and release data.

Stars
9,134
Primary language
C++
License
Apache-2.0
Repository last updated
Oct 3, 2026
On this page

Overview

CatBoost is a machine-learning library based on gradient boosting over decision trees. It supports numerical and categorical features for classification, regression, ranking, and related tasks. Python, R, Java, and C++ workflows are available, with CPU, GPU, and multi-GPU training options.

Features and best fit

Based on official documentation; not hands-on tested · Content checked:

Key features

Train directly on tabular data containing categorical features

CatBoost highlights support for both numerical and categorical features, reducing the need to make a separate one-hot encoding pipeline mandatory before GBDT training.

Sources: [1]

Use one library for classification, regression, and ranking

The project and Python package target classification, regression, ranking, and other machine-learning tasks through task-specific estimators, losses, and metrics.

Sources: [1][2]

Choose CPU, GPU, or distributed training for the workload

The README documents GPU and multi-GPU training and also points to distributed workflows through Apache Spark and the command-line interface.

Sources: [1]

Best fit

Fits GBDT evaluation on tabular datasets with many categorical values

It is useful for customer, product, event, and similar structured datasets where numerical and categorical columns are mixed and classification, regression, or ranking models are being compared.

Sources: [1]

Before adoption

Validate GPU and distributed environments separately from local CPU use

GPU and Spark execution add hardware, driver, runtime, and cluster requirements beyond a local CPU workflow. Validate the model and feature definitions locally before scaling the execution environment.

Sources: [1]

Source builds require CMake, Conan, Cython, and related tooling

The Python package pyproject declares CMake, Conan, Cython, NumPy, and other build dependencies. Prefer prebuilt packages for normal use and assemble the native toolchain only when source builds are required.

Sources: [3][2]

Pin 1.2.10 and record the model together with the library version

As of October 3, 2026, the latest stable release is 1.2.10. Record the training library version with model artifacts and regression-test predictions and metrics when upgrading.

Sources: [5][4][6]

Official sources

  1. [1]CatBoost v1.2.10 README(2026-10-03)
  2. [2]CatBoost 1.2.10 Python setup.py(2026-10-03)
  3. [3]CatBoost 1.2.10 Python pyproject.toml(2026-10-03)
  4. [4]CatBoost 1.2.10 version metadata(2026-10-03)
  5. [5]CatBoost 1.2.10 release(2026-10-03)
  6. [6]CatBoost Apache-2.0 license(2026-10-03)
Supplemental curator note

CatBoost fits tabular workloads where categorical values should be handled without a separate encoding pipeline. Compare accuracy together with training time, inference latency, and model size on the real dataset before choosing CPU/GPU settings or selecting among GBDT libraries.

Try it in 3 steps

  1. 1

    Install CatBoost 1.2.10 in a virtual environment

    Install the prebuilt Python package with the stable release pinned.

    python3 -m venv .venv && . .venv/bin/activate && python -m pip install catboost==1.2.10
  2. 2

    Train a tiny classification model

    Train and predict on a small CPU-only dataset without downloading external data.

    python -c "from catboost import CatBoostClassifier; m=CatBoostClassifier(iterations=5,verbose=False); m.fit([[0],[1],[2],[3]],[0,0,1,1]); print(m.predict([[2.5]]))"
  3. 3

    Verify the installed version

    Confirm the library version so it can be recorded together with model artifacts.

    python -c "import catboost; print(catboost.__version__)"
Check the official README

Growth

Growth trends · Last 30 days

9,134 Stars

Trend data is still being collected.

Development activity

Last 90 days · weekly

Commits (last 30 days)
112
Open PRs
63

Development activity is still being collected.

Built with

Categories and tags

Categories

Tags

GitHub data

GitHub dataView detailed GitHub data
View on GitHub

GitHub Topics

  • machine-learning
  • decision-trees
  • gradient-boosting
  • gbm
  • gbdt
  • python
  • r
  • kaggle
  • gpu-computing
  • catboost
  • tutorial
  • categorical-features
Stars
9,134
Forks
1,343
Watchers
187
Open issues
673
Contributors
189
Owner type
Organization
Primary language
C++
License
Apache-2.0
Repository last updated
Oct 3, 2026
Write a related article

Share a guide or use case for this OSS in Markdown. Articles are published after administrator approval.

Report incorrect information

Tell us if any listing information is incorrect or outdated.

After reading this page, do you know what to do next?