Install the released version

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.

Operating system CPU architectures GPU support using CUDA
macOS (versions currently supported by Apple) x86_64 and arm64 no
Linux (compatible with manylinux2014 platform tag ) x86_64 and aarch64 yes
Windows 10 and 11 x86_64 yes

Note

Release binaries for x86_64 CPU architectures are built with SIMD extensions SSE2, SSE3, SSSE3, SSE4 enabled. If you need to run CatBoost on older CPUs that do not support these instruction sets build CatBoost artifacts yourself

Note

This version has CUDA-enabled GPU support out-of-the-box on Linux and Windows.

As of CatBoost 1.2.8, devices with CUDA compute capability >= 3.5 are supported in released packages.

All necessary CUDA libraries are statically linked in the released Linux and Windows binaries, the only installation necessary is the appropriate version of the CUDA driver.

Training or inference on CUDA-enabled GPUs requires NVIDIA Driver of version 450.80.02 or higher.

To install the released R package binary:

  1. Run the following commands:

    install.packages('remotes')
remotes::install_url('BINARY_URL'[, INSTALL_opts = c("--no-multiarch", "--no-test-load")])

    • BINARY_URL is the URL of the released binary.

      It should look like this:

      https://github.com/catboost/catboost/releases/download/v{RELEASE_VERSION}/catboost-R-{OS}-{ARCHITECTURE}-{RELEASE_VERSION}.tgz

      Operating system {OS} value Possible {ARCHITECTURE} values GPU support using CUDA
      Linux (compatible with manylinux2014 platform tag ) linux x86_64 and aarch64 yes
      macOS (versions currently supported by Apple) darwin universal2 - supports both x86_64 and arm64 (Apple silicon) no
      Windows 10 and 11 windows x86_64 yes

    • INSTALL_opts:

      • --no-multiarch is an optional parameter for multiarch support.
      • --no-test-load is an argument which suppresses the testing phase.

    Examples:

    • install the macOS version 1.2.8 of the R package:
    install.packages('remotes')
remotes::install_url('https://github.com/catboost/catboost/releases/download/v1.2.8/catboost-R-darwin-universal2-1.2.8.tgz', INSTALL_opts = c("--no-multiarch", "--no-test-load"))
    • install the Linux version 1.2.8 of the R package for the aarch64 CPU architecture:
    install.packages('remotes')
remotes::install_url('https://github.com/catboost/catboost/releases/download/v1.2.8/catboost-R-linux-aarch64-1.2.8.tgz', INSTALL_opts = c("--no-multiarch", "--no-test-load"))
    • install the Windows version 1.2.8 of the R package for the x86_64 CPU architecture:
    install.packages('remotes')
remotes::install_url('https://github.com/catboost/catboost/releases/download/v1.2.8/catboost-R-windows-x86_64-1.2.8.tgz', INSTALL_opts = c("--no-multiarch", "--no-test-load"))

    Warning

    Do not forget to change the package version to the required one. The example above illustrates the installation of the version 1.2.8.

    To install another version, let's say 0.16.5, change the code to the following:

    install.packages('remotes')
remotes::install_url('https://github.com/catboost/catboost/releases/download/v0.16.5/catboost-R-Windows-0.16.5.tgz', INSTALL_opts = c("--no-multiarch", "--no-test-load"))

Troubleshooting

Installation error

If you use script for installation, try to add the repos parameter to the install.packages('devtools') command.
Select any repository from the list.

Example: install.packages("devtools", repos="http://cran.us.r-project.org").

ERROR: some hard-coded temporary paths could not be fixed

This can be mitigated by adding "--no-staged-install" option to INSTALL_opts parameter of remotes::install_url call

Previous
Overview
Next
conda install