История

Brian 89948de905 R GPU Dockerfiles (#341 ) * Added dockerfiles for aztk r gpu * Removed tensorflow/cntk cpu versions * Enabled configure option for shared lib for python * Removed temp paths * Replaced python version with aztk version * Cleaning up args * Added variable definition for tensorflow, shorten lines		2018-01-22 16:36:36 -08:00
..
spark1.6.3	R GPU Dockerfiles (#341 )	2018-01-22 16:36:36 -08:00
spark2.1.0	R GPU Dockerfiles (#341 )	2018-01-22 16:36:36 -08:00
spark2.2.0	R GPU Dockerfiles (#341 )	2018-01-22 16:36:36 -08:00
README.md	Feature/python container (#210 )	2017-11-09 00:43:32 -08:00

README.md

Base AZTK Docker image

This Dockerfile is used to build the aztk-base image used by this toolkit. This Dockerfile produces the Docker image that is selected by AZTK by default.

You can modify this Dockerfile to build your own image.

How to build this image

This Dockerfile takes in a single variable at build time that allows you to specify your desired Spark version: SPARK_VERSION_KEY.

By default, this image will also be installed with python v3.5.4 as a requirement for this toolkit.

# For example, if I want to use Spark 1.6.3 I would build the image as follows:
docker build \
    --build-arg SPARK_VERSION_KEY=spark-1.6.3-bin-hadoop2.6 \
    -t <my_image_tag> .

SPARK_VERSION_KEY is used to locate which version of Spark to download. These are the values that have been tested and known to work:

spark-1.6.3-bin-hadoop2.6
spark-2.1.0-bin-hadoop2.7
spark-2.2.0-bin-hadoop2.7

For a full list of supported keys, please see this page

NOTE: Do not include the '.tgz' suffix as part of the Spark version key.