Installation for Windows users


Download latest binary distribution of BigARTM from Explicit download links can be found at Downloads section (64 bit only).

The distribution will contain pre-build binaries, command-line interface and BigARTM API for Python. The distribution also contains a simple dataset. More datasets in BigARTM-compatible format are available in the Downloads section.

You may also try BigARTM from pre-built docker container, as described in Installation as Docker container.

Use BigARTM from command line

Download the latest package from Unpack it to C:\BigARTM. Open command line or power shell and change working directory to C:\BigARTM\bin. Now you may use BigARTM command line utility bigartm.exe. Go to BigARTM Command Line Utility for examples and further directions.

Configure BigARTM Python API

  1. Install Python, for example from Anaconda distribution ( You are free to choose Python 2.7 or Python 3, both are supported with BigARTM. You have to choose 64 bit Python installation.

  2. Install Python packages that BigARTM depend on.

    • numpy >= 1.9.2 (skip this if you install via Anaconda)
    • pandas >= 0.16.2 (skip this if you install via Anaconda)
    • protobuf >= 3.0.0 (not included in Anaconda; install by running pip install protobuf==3.0.0)
    • tqdm (not included in Anaconda; install by running pip install tqdm)
  3. Add C:\BigARTM\bin folder to your PATH system variable, and add C:\BigARTM\python to your PYTHONPATH system variable:

    set PATH=%PATH%;C:\BigARTM\bin

    Remember to change C:\BigARTM if you unpack to another location.

  4. Now you can used BigARTM from Python shell or from ipython notebook. Refer to Python Tutorial or Python Guide for examples, or to Python Interface for documentation.

If you are getting errors when configuring or using Python API, please refer to Troubleshooting chapter in Installation for Linux and Mac OS-X users. The list of issues is common between Windows and Linux. Normally you should install protobuf for python with pip install. An alternative way is to install it from the package that comes with BigARTM. This is described here.