How to Install XGBoost for Python on macOS

XGBoost is a library for developing very fast and accurate gradient boosting models.

It is a library at the center of many winning solutions in Kaggle data science competitions.

In this tutorial, you will discover how to install the XGBoost library for Python on macOS.

Kick-start your project with my new book XGBoost With Python, including step-by-step tutorials and the Python source code files for all examples.

Let’s get started.

How to Install XGBoost for Python on macOS

How to Install XGBoost for Python on macOS
Photo by auntjojo, some rights reserved.

Tutorial Overview

This tutorial is divided into 3 parts; they are:

  1. Install MacPorts
  2. Build XGBoost
  3. Install XGBoost

Note: I have used this procedure for years on a range of different macOS versions and it has not changed. This tutorial was written and tested on macOS High Sierra (10.13.1).

1. Install MacPorts

You need GCC and a Python environment installed in order to build and install XGBoost for Python.

I recommend GCC 7 and Python 3.6 and I recommend installing these prerequisites using MacPorts.

  • 1. For help installing MacPorts and a Python environment step-by-step, see this tutorial:

>> How to Install a Python 3 Environment on Mac OS X for Machine Learning and Deep Learning

  • 2. After MacPorts and a working Python environment are installed, you can install and select GCC 7 as follows:

  • 3. Confirm your GCC installation was successful as follows:

You should see the version of GCC printed; for example:

What version did you see?
Let me know in the comments below.

2. Build XGBoost

The next step is to download and compile XGBoost for your system.

  • 1. First, check out the code repository from GitHub:

  • 2. Change into the xgboost directory.

  • 3. Copy the configuration we intend to use to compile XGBoost into position.

  • 4. Compile XGBoost; this requires that you specify the number of cores on your system (e.g. 8, change as needed).

The build process may take a minute and should not produce any error messages, although you may see some warnings that you can safely ignore.

For example, the last snippet of the compilation might look as follows:

Did this step work for you?
Let me know in the comments below.

3. Install XGBoost

You are now ready to install XGBoost on your system.

  • 1. Change directory into the Python package of the xgboost project.

  • 2. Install the Python XGBoost package.

The installation is very fast.

For example, at the end of the installation, you may see messages like the following:

  • 3. Confirm that the installation was successful by printing the xgboost version, which requires the library to be loaded.

Save the following code to a file called

Run the script from the command line:

You should see the XGBoost version printed to screen:

How did you do?
Post your results in the comments below.

Further Reading

This section provides more resources on the topic if you are looking to go deeper.


In this tutorial, you discovered how to install XGBoost for Python on macOS step-by-step.

Do you have any questions?
Ask your questions in the comments below and I will do my best to answer.

Discover The Algorithm Winning Competitions!

XGBoost With Python

Develop Your Own XGBoost Models in Minutes

...with just a few lines of Python

Discover how in my new Ebook:
XGBoost With Python

It covers self-study tutorials like:
Algorithm Fundamentals, Scaling, Hyperparameters, and much more...

Bring The Power of XGBoost To Your Own Projects

Skip the Academics. Just Results.

See What's Inside

58 Responses to How to Install XGBoost for Python on macOS

  1. Avatar
    Abhishek Dalakoti January 17, 2018 at 8:36 pm #

    can you put for windows?

  2. Avatar
    Juliano Petronetto January 19, 2018 at 5:15 am #

    Or can simply use it in Docker:

  3. Avatar
    Gabriel Kreplak January 19, 2018 at 6:45 am #


    $ gcc -v

    I did not get gcc version but:

    Configured with: –prefix=/Library/Developer/CommandLineTools/usr –with-gxx-include-dir=/usr/include/c++/4.2.1
    Apple LLVM version 9.0.0 (clang-900.0.39.2)
    Target: x86_64-apple-darwin17.3.0
    Thread model: posix
    InstalledDir: /Library/Developer/CommandLineTools/usr/bin

    I didn’t try further (still)

    • Avatar
      Jason Brownlee January 20, 2018 at 8:11 am #

      Looks like you are still using the apple compiler.

      You may need to tell macports that you want to use the newly installed compiler.

    • Avatar
      Steven Limpert June 16, 2018 at 8:00 pm #

      I had the same issue. However, I resolved it by using the command “hash -r” to get Terminal to recognize the change. According to this,, opening a new Terminal window would have worked as well. After using “hash -r”, “gcc -v” gave the same result shown in this tutorial.

  4. Avatar
    Sergey January 20, 2018 at 11:09 pm #

    Hi Jason

    All worked great till

    make -j8

    I received fatal error and then – some errors

    /opt/local/include/gcc7/c++/cwchar:44:10: fatal error: wchar.h: No such file or directory

    Do you have any ideas?

    • Avatar
      Jason Brownlee January 21, 2018 at 9:11 am #

      Oh dear. No good ideas sorry.

      Perhaps try posting to stack overflow or the xgboost issue tracker:

      Or try checking out the v0.6 version in case something changed in the recent 0.7 release?

    • Avatar
      MC February 17, 2018 at 6:30 am #

      Same error

    • Avatar
      Shuqing Ke March 26, 2020 at 3:09 am #

      Hi guys,

      I have the same errors. Have you figured out how to resolve this please?Thank you in advance!

      • Avatar
        Jason Brownlee March 26, 2020 at 7:59 am #

        Perhaps try installing via pip instead:

  5. Avatar
    cgironda January 23, 2018 at 12:13 am #

    Hi Jason,

    I just used,

    conda install -c conda-forge xgboost

    So I did not follow steps 1, 2 and 3. However, it depends on tastes.


    • Avatar
      Jason Brownlee January 23, 2018 at 8:01 am #

      Great tip, thanks!

    • Avatar
      MC February 17, 2018 at 6:30 am #

      Yes, this worked.

    • Avatar
      zyx December 18, 2019 at 4:18 pm #

      I upvote this one! It worked for me as well. Saved many steps!

    • Avatar
      Mandy Wu February 5, 2020 at 3:42 am #

      Upvote! it worked for me, thank you cgironda!

    • Avatar
      Solihin March 25, 2020 at 1:38 pm #

      Great tip! Thanks

    • Avatar
      Santiago April 1, 2020 at 7:53 am #

      Thanks a lot!

  6. Avatar
    Rajat Sharma January 30, 2018 at 11:23 pm #

    Why do you think that python is a better option for machine learning than R or Matlab?

  7. Avatar
    Arvind Ravichandran February 9, 2018 at 3:07 am #

    Thank you very much! Worked brilliantly.

  8. Avatar
    Matt Smith February 22, 2018 at 4:23 am #

    Brilliant! I was previously trying to use homebrew to install gcc and then editing the file, as basically all other tutorials tell you to, but was constantly getting errors. This is the first method that worked for me (on mac OS Sierra 10.12.6); thank you!

  9. Avatar
    Gary March 19, 2018 at 7:29 am #

    cgironda you are a star! You could spend a day trying to answer all the questions out there related to unsupported option '-fopenmp' and you wouldn’t even come close to getting to all of them. You rescued me after about 4 hours of fussing with this. Thanks!

  10. Avatar
    Alvic May 15, 2018 at 12:00 pm #

    Hi Jason,

    Have followed your instructions up to step 2 build XG boost and this is the message I saw on my terminal.

    a – build/metric/metric.o
    a – build/metric/multiclass_metric.o
    a – build/metric/rank_metric.o
    a – build/objective/multiclass_obj.o
    a – build/objective/objective.o
    a – build/objective/rank_obj.o
    a – build/objective/regression_obj.o
    a – build/predictor/cpu_predictor.o
    a – build/predictor/predictor.o
    a – build/tree/tree_model.o
    a – build/tree/tree_updater.o
    a – build/tree/updater_colmaker.o
    a – build/tree/updater_fast_hist.o
    a – build/tree/updater_histmaker.o
    a – build/tree/updater_prune.o
    a – build/tree/updater_refresh.o
    a – build/tree/updater_skmaker.o
    a – build/tree/updater_sync.o
    ld: can’t open output file for writing: xgboost, errno=21 for architecture x86_64
    collect2: error: ld returned 1 exit status
    make: *** [xgboost] Error 1
    make: *** Waiting for unfinished jobs….
    Alvics-MacBook-Pro-2:xgboost alviceugenejosol$ a – b

    • Avatar
      Jason Brownlee May 15, 2018 at 2:45 pm #

      Ouch. I have not seen this.

      Perhaps try a make clean, then try building again?
      Perhaps try posting the error to the xgboost user group?

  11. Avatar
    Akeel September 20, 2018 at 2:47 am #

    Hi Jason,
    Worked like a charm. Thanks for your detailed instructions.

  12. Avatar
    Paul M Washburn November 4, 2018 at 7:36 am #

    Step 2 did not work …

    Pauls-MacBook-Pro:xgboost pmw$ make -j8
    Makefile:31: MAKE [/Library/Developer/CommandLineTools/usr/bin/make] – checked OK
    c++ -DDMLC_LOG_CUSTOMIZE=1 -std=c++11 -Wall -Wno-unknown-pragmas -Iinclude -Idmlc-core/include -Irabit/include -I/include -O3 -funroll-loops -msse2 -fPIC -fopenmp -MM -MT build/c_api/c_api.o src/c_api/ >build/c_api/c_api.d
    c++ -DDMLC_LOG_CUSTOMIZE=1 -std=c++11 -Wall -Wno-unknown-pragmas -Iinclude -Idmlc-core/include -Irabit/include -I/include -O3 -funroll-loops -msse2 -fPIC -fopenmp -MM -MT build/logging.o src/ >build/logging.d
    c++ -DDMLC_LOG_CUSTOMIZE=1 -std=c++11 -Wall -Wno-unknown-pragmas -Iinclude -Idmlc-core/include -Irabit/include -I/include -O3 -funroll-loops -msse2 -fPIC -fopenmp -MM -MT build/learner.o src/ >build/learner.d
    c++ -DDMLC_LOG_CUSTOMIZE=1 -std=c++11 -Wall -Wno-unknown-pragmas -Iinclude -Idmlc-core/include -Irabit/include -I/include -O3 -funroll-loops -msse2 -fPIC -fopenmp -MM -MT build/c_api/c_api_error.o src/c_api/ >build/c_api/c_api_error.d
    c++ -DDMLC_LOG_CUSTOMIZE=1 -std=c++11 -Wall -Wno-unknown-pragmas -Iinclude -Idmlc-core/include -Irabit/include -I/include -O3 -funroll-loops -msse2 -fPIC -fopenmp -MM -MT build/common/common.o src/common/ >build/common/common.d
    c++ -DDMLC_LOG_CUSTOMIZE=1 -std=c++11 -Wall -Wno-unknown-pragmas -Iinclude -Idmlc-core/include -Irabit/include -I/include -O3 -funroll-loops -msse2 -fPIC -fopenmp -MM -MT build/common/hist_util.o src/common/ >build/common/hist_util.d
    c++ -DDMLC_LOG_CUSTOMIZE=1 -std=c++11 -Wall -Wno-unknown-pragmas -Iinclude -Idmlc-core/include -Irabit/include -I/include -O3 -funroll-loops -msse2 -fPIC -fopenmp -MM -MT build/common/host_device_vector.o src/common/ >build/common/host_device_vector.d
    c++ -DDMLC_LOG_CUSTOMIZE=1 -std=c++11 -Wall -Wno-unknown-pragmas -Iinclude -Idmlc-core/include -Irabit/include -I/include -O3 -funroll-loops -msse2 -fPIC -fopenmp -MM -MT build/data/data.o src/data/ >build/data/data.d
    clang: clangclang: : error: errorerrorunsupported option ‘-fopenmp’: : unsupported option ‘-fopenmp’unsupported option ‘-fopenmp’

    clang: clang: error: unsupported option ‘-fopenmp’error:
    unsupported option ‘-fopenmp’
    clang: error: unsupported option ‘-fopenmp’
    clang: error: unsupported option ‘-fopenmp’
    clang: error: unsupported option ‘-fopenmp’
    make: *** [build/common/common.o] Error 1
    make: *** Waiting for unfinished jobs….
    make: *** [build/c_api/c_api_error.o] Error 1
    make: *** [build/learner.o] Error 1
    make: *** [build/logging.o] Error 1
    make: *** [build/c_api/c_api.o] Error 1
    make: *** [build/common/host_device_vector.o] Error 1
    make: *** [build/common/hist_util.o] Error 1
    make: *** [build/data/data.o] Error 1

    • Avatar
      Jason Brownlee November 5, 2018 at 6:06 am #

      Prhaps try installing with pip on your workstation instead?


  13. Avatar
    Chris February 20, 2019 at 1:35 am #

    After installing MacPorts, I tried port install gcc but got an error that no port named gcc7 existed. Tried several variants. Finally had to do a port install self update, and then the rest worked like a charm. So maybe add the self update as step one in the Installing MacPorts section. Everything else worked great. Thanks for this and all your other articles.

  14. Avatar
    Diane February 5, 2020 at 3:52 am #

    When I ran the step:
    make -j4
    I got an error:
    clang: error: unsupported option ‘-fopenmp’
    clang: error: unsupported option ‘-fopenmp’
    clang: clang: error: unsupported option ‘-fopenmp’error: unsupported option ‘-fopenmp’

    make: *** [build/c_api/c_api_error.o] Error 1
    make: *** Waiting for unfinished jobs….
    make: *** [build/c_api/c_api.o] Error 1
    make: *** [build/learner.o] Error 1
    make: *** [build/logging.o] Error 1

    any thoughts for how to fix that?

  15. Avatar
    özgür koska February 7, 2020 at 7:43 am #

    Hi Jason,
    Thank you for your great jobs!
    At the last step I got the belowmentioned error. Any fix Suggestion?

    apple@Apple-MacBook-Pro python-package % sudo python install
    Traceback (most recent call last):
    File “”, line 277, in
    TypeError: ‘encoding’ is an invalid keyword argument for this function

    • Avatar
      Jason Brownlee February 7, 2020 at 8:29 am #

      Perhaps ensure you are using Python 3.6 or higher?

      • Avatar
        özgür February 9, 2020 at 3:20 am #

        Thank you very much Jason.
        After your comment I got the issue.

        sudo python3 install solved the issue.


  16. Avatar
    sandeep March 3, 2020 at 5:44 pm #

    Hi Jason,
    Thank you for your great jobs!
    At the last step I got the belowmentioned error. Any fix Suggestion?
    INFO:XGBoost build_ext:Building from source. /Users/smehta27/Desktop/Dental/xgboost/lib/libxgboost.dylib
    error: [Errno 2] No such file or directory: ‘cmake’: ‘cmake’

    • Avatar
      Jason Brownlee March 4, 2020 at 5:50 am #

      Looks like you need to install make, perhaps via macports or brew.

  17. Avatar
    Jose Quinonez March 17, 2020 at 7:37 am #

    Hi Jason,
    Great tutorial as always!

    I have Mac OS Catalina and I had problems installing MacPorts, but this link was helpful:

    Thank you

  18. Avatar
    angel May 12, 2020 at 3:14 pm #

    Hi Jason, Thanks for sharing.
    I worked from step 2 and step3
    For stop 2, I did not find similar display as yours. Instead, it becomes:
    Makefile:23: MAKE [/Library/Developer/CommandLineTools/usr/bin/make] – checked OK
    c++ -c -DDMLC_LOG_CUSTOMIZE=1 -std=c++11 -Wall -Wno-unknown-pragmas -Iinclude -Idmlc-core/include -Irabit/include -I/include -O3 -funroll-loops amalgamation/ -o amalgamation/xgboost-all0.o
    In file included from amalgamation/
    amalgamation/../src/common/ warning: unused variable ‘ReadErr’
    auto ReadErr = [&fname]() {
    In file included from amalgamation/
    In file included from amalgamation/../src/metric/
    In file included from dmlc-core/include/dmlc/./registry.h:9:
    In file included from /Library/Developer/CommandLineTools/usr/bin/../include/c++/v1/map:479:
    In file included from /Library/Developer/CommandLineTools/usr/bin/../include/c++/v1/__tree:15:
    /Library/Developer/CommandLineTools/usr/bin/../include/c++/v1/memory:2338:5: warning:
    delete called on non-final ‘xgboost::tree::QuantileHistMaker::Builder’
    that has virtual functions but non-virtual destructor
    delete __ptr;
    /Library/Developer/CommandLineTools/usr/bin/../include/c++/v1/memory:2651:7: note:
    in instantiation of member function
    requested here
    amalgamation/../src/tree/ note: in instantiation
    of member function
    >::reset’ requested here
    builder_.reset(new Builder(
    In file included from amalgamation/
    In file included from amalgamation/../src/data/
    In file included from amalgamation/../src/data/./sparse_page_dmatrix.h:17:
    In file included from amalgamation/../src/data/ellpack_page_source.h:13:
    amalgamation/../src/metric/../common/hist_util.h:695:10: warning: private field
    ‘nthread_’ is not used [-Wunused-private-field]
    size_t nthread_ { 0 };
    3 warnings generated.

    Then, when doing step 3: error message occurs:
    INFO:XGBoost build_ext:Building from source. /Users/
    INFO:XGBoost build_ext:Run CMake command: [‘cmake’, ‘xgboost’, ‘-GUnix Makefiles’, ‘-DUSE_OPENMP=1’, ‘-DUSE_CUDA=0’, ‘-DUSE_NCCL=0’, ‘-DBUILD_WITH_SHARED_NCCL=0’, ‘-DHIDE_CXX_SYMBOLS=1’, ‘-DUSE_HDFS=0’, ‘-DUSE_AZURE=0’, ‘-DUSE_S3=0’, ‘-DPLUGIN_LZ4=0’, ‘-DPLUGIN_DENSE_PARSER=0’]
    error: [Errno 2] No such file or directory: ‘cmake’

    • Avatar
      Jason Brownlee May 13, 2020 at 6:23 am #

      I’m sorry to hear that. Perhaps try posting/searching on stackoverflow or the github issues for the xgboost project?

  19. Avatar
    Danielle T September 4, 2020 at 6:37 am #

    Hello, thank you for sharing. I encounter a problem within the second section.
    I can’t find a solution for this. I am using Catalina and wonder if this is part of the problem? I did see your previous comment to someone else’s post on installing make via port or brew. Brew repeatedly fails on my system (internet time out problems). Do you happen to know how I can install make via macports?

    Many thanks!

    Danielles-MacBook-Pro:Desktop danielleturvill$ cd xgboost/
    Danielles-MacBook-Pro:xgboost danielleturvill$ cp make/ ./
    cp: make/ No such file or directory
    Danielles-MacBook-Pro:xgboost danielleturvill$ ls
    CITATION gputreeshap
    CMakeLists.txt amalgamation include appveyor.yml jvm-packages
    Jenkinsfile cmake plugin
    Jenkinsfile-win64 cub python-package
    LICENSE demo rabit
    Makefile dev src dmlc-core tests
    R-package doc

    • Avatar
      Jason Brownlee September 4, 2020 at 7:58 am #

      Newer versions do seem to cause problems. I recommend using an older version of XGBoost, such as 1.0.2

      You can checkout this version directly or install it directly via pip, e.g.

  20. Avatar
    Minnie December 2, 2020 at 6:02 pm #

    Hi Jason,

    I am trying to install xgboost integrated with GPU support, on my MacOS Mojave(10.14.6) from last 3 days, however, no success has been reached. I tried 2 approaches:

    1. pip install xgboost

    xgboost is installed here and it runs successfully without GPU option(i.e., without tree_method=’gpu_hist’).

    I want to run with gpu_hist by giving “tree_method=’gpu_hist’ ” in tree parameters. When I gave “tree_method=’gpu_hist’ ” in tree parameters, following error has come:

    XGBoostError: [12:10:34] /Users/travis/build/dmlc/xgboost/src/gbm/../common/common.h:153: XGBoost version not compiled with GPU support.
    Stack trace:
    [bt] (0) 1 libxgboost.dylib 0x000000012256ba60 dmlc::LogMessageFatal::~LogMessageFatal() + 112
    [bt] (1) 2 libxgboost.dylib 0x00000001225f92b3 xgboost::gbm::GBTree::ConfigureUpdaters() + 531
    [bt] (2) 3 libxgboost.dylib 0x00000001225f8b97 xgboost::gbm::GBTree::Configure(std::__1::vector<std::__1::pair<std::__1::basic_string<char, std::__1::char_traits, std::__1::allocator >, std::__1::basic_string<char, std::__1::char_traits, std::__1::allocator > >, std::__1::allocator<std::__1::pair<std::__1::basic_string<char, std::__1::char_traits, std::__1::allocator >, std::__1::basic_string<char, std::__1::char_traits, std::__1::allocator > > > > const&) + 967
    [bt] (3) 4 libxgboost.dylib 0x0000000122611a0c xgboost::LearnerConfiguration::Configure() + 1500
    [bt] (4) 5 libxgboost.dylib 0x0000000122611e68 xgboost::LearnerImpl::UpdateOneIter(int, std::__1::shared_ptr) + 120
    [bt] (5) 6 libxgboost.dylib 0x000000012256331d XGBoosterUpdateOneIter + 157
    [bt] (6) 7 libffi.7.dylib 0x0000000102102ead ffi_call_unix64 + 85
    [bt] (7) 8 ??? 0x00007ffeee291da0 0x0 + 140732894092704

    2. My second approach:

    git clone –recursive
    cd xgboost/
    make -j4
    cd python-package
    python3 install

    Though it installs xgboost, but throws following error whille running this statement:

    AttributeError: dlsym(0x7ffe9aed62f0, XGDMatrixSetDenseInfo): symbol not found

    Can you please help in installing xgboost with GPU support?

    • Avatar
      Jason Brownlee December 3, 2020 at 8:15 am #

      Try checking out an older version, such as 1.0.1.

      I find the older version compile easily and work just as well.

      • Avatar
        Minnie December 5, 2020 at 9:33 pm #

        Jason, xgboost 1.0.1 is installed with ‘pip install xgboost==1.0.1′ . And it works as well

        However, it is without GPU support and I want to install xgboost with GPU Support

        I got following error, when I ran xgboost with tree_method=’gpu_hist’:
        (import xgboost as xgb
        bst=xgb.XGBClassifier(n_estimators=20, max_depth=9, learning_rate=0.05,n_jobs=4,gamma=5,min_child_weight=20, eval_metric=’rmse’, missing=-999, tree_method=’gpu_hist’)
        XGBoostError: [15:58:10] /private/var/folders/4n/s3z85_zs0z7_12y8mmg2yqd00000gn/T/pip-install-xq6d26qe/xgboost_a072b8d552e84085a86dcf21271bcc77/xgboost/include/xgboost/gbm.h:166: XGBoost version not compiled with GPU support.

        Can you please help in installing xgboost with GPU supprt?

        • Avatar
          Jason Brownlee December 6, 2020 at 6:59 am #

          Sorry, I am not an expert on GPU support in XGBoost, perhaps contact the XGBoost developers directly, e.g. post an issue on their github project.

Leave a Reply