For Python users, Analytics Zoo can be installed either from pip or without pip.

NOTE: We have tested on Python 3.6 and Python 3.7. Support for Python 2.7 has been removed due to its end of life.

Install the latest nightly build wheels for pip

You can find the list of the latest nightly build wheels for pip here. Choose a wheel with timestamp for your platform and click to download it to your local repository.

Note that if you are using Linux, you need to choose the file analytics_zoo-VERSION-TIMESTAMP-py2.py3-none-manylinux1_x86_64.whl. If you are using Mac, choose analytics_zoo-VERSION-TIMESTAMP-py2.py3-none-macosx_10_11_x86_64.whl instead.

After downloading the target wheel file, go to the download directory and run the following command:

pip install analytics_zoo-*-py2.py3-none-*_x86_64.whl

See here for some remarks with regard to pip install.

Install from pip for local usage

You can use the following command to install the latest release version of analytics-zoo via pip easily:

pip install analytics-zoo


  1. Installing analytics-zoo from pip will automatically install pyspark. To avoid possible conflicts, you are highly recommended to unset SPARK_HOME if it exists in your environment.

  2. Please always first call init_nncontext() at the very beginning of your code after pip install. This will create a SparkContext with optimized performance configuration and initialize the BigDL engine.

from zoo.common.nncontext import *
sc = init_nncontext()


  1. We've tested this package with pip 9.0.1. pip install --upgrade pip if necessary.
  2. Pip install supports Mac and Linux platforms.
  3. You need to install Java >= JDK8 before running Analytics Zoo, which is required by pyspark.
  4. bigdl==0.12.2, pyspark==2.4.3, conda-pack==0.3.1 and their dependencies will be automatically installed if they haven't been detected in the current Python environment.

Install from pip for Yarn cluster

You only need to following these steps on your driver node and we only support yarn-client mode for now.

1) Install Conda in your environment.

2) Create a new conda environment (with name "zoo" for example):

conda create -n zoo python=3.6
source activate zoo

3) Install the latest release version of analytics-zoo into the created conda environment.

pip install analytics-zoo

See here if you want to install the latest nightly build version.

4) Download JDK8 and set the environment variable: JAVA_HOME (recommended).

You can also install JDK via conda without setting the JAVA_HOME manually:

conda install -c anaconda openjdk=8.0.152

5) Start python and then execute the following code to create a SparkContext on Yarn for verification:

from zoo import init_spark_on_yarn

sc = init_spark_on_yarn(
    hadoop_conf="path to the yarn configuration folder",
    conda_name="zoo", # The name of the created conda-env

Install without pip

If you choose to install Analytics Zoo without pip, you need to prepare Spark and install necessary Python dependencies.


  1. Download Spark

    • Note that Python 3.6 is only compatible with Spark 1.6.4, 2.0.3, 2.1.1 and >=2.2.0. See this issue for more discussion.
  2. You are recommended to download Analytics Zoo prebuilt release package from the Release Page and extract it. Alternatively, you can also build the Analytics Zoo from source.

  3. Install Python dependencies. Analytics Zoo only depends on numpy and six for now.

For Spark standalone cluster

For Yarn cluster

You can run Analytics Zoo Python programs on Yarn clusters without changes to the cluster (i.e., no need to pre-install any Python dependency).

You can first package all the required dependencies into a conda environment on the local node (where you will run the spark-submit command), and then directly use spark-submit to run the Analytics Zoo Python program on the Yarn cluster using that conda environment.

Follow the steps below to create the conda environment:

1) Install Conda in your environment.

2) Create a new conda environment (with name "environment" for example):

conda create -n environment python=3.6 pip=20.0.2 setuptools=45.2.0 conda-pack==0.3.1
conda activate environment

3) Install python dependencies into the created conda environment. Install the dependencies according to the dependencies listed in requirements.txt. You can add your own dependencies into this file if you wish. The current requirements only contain those needed for running Analytics Zoo Python examples and models.

pip install -r ${ANALYTICS_ZOO_HOME}/bin/requirements.txt

4) Create conda package for the created conda environment

conda pack -o environment.tar.gz

Then you can find the conda package file under the current directory.