Get the binary

The following installation methods are supported:

Note that there are additional system requirements and peculiarities if training on GPU support is required.

Download

An up-to-date list of available CatBoost releases and the corresponding binaries for different operating systems is available in the Download section of the releases page on GitHub.

Note.

Only CUDA 8.0 is supported. Windows users must  build the binary from a local copy if GPU support is required and the installed version of CUDA differs from 8.0.

Build the binary from a local copy

To install the command-line package from a local copy of the CatBoost repository:
  1. Install the libc header files on macOS and Linux:
    • macOS: xcode-select --install
    • Linux: install the appropriate package (for example, libc6-dev on Ubuntu)
  2. Clone the repository:

    git clone https://github.com/catboost/catboost.git
  3. Open the catboost/catboost/app catalog from the local copy of the CatBoost repository.

  4. Run the following command:

    ../../ya make -r [-o <output directory>] [-DCUDA_ROOT=<path to the CUDA directory>]

    Use the DCUDA_ROOT option if training on GPU support is required.

GPU system requirements

Devices of compute capability 3.5 and higher are supported.

GPU peculiarities

  • Some training parameters are missing but will be added in future releases
  • Multiple train runs with the same random seed may result in different formulas because of the float summation order