Here are the installation modes for TensorFlow:
Can any of these installations be automated through Cloudbreak? If not, do you already have a recommended way of installing TensorFlow on HDP?
@Neeraj Sabharwal I guess there is an article which shows that both Spark and TensorFlow can work together - https://databricks.com/blog/2016/01/25/deep-learning-with-spark-and-tensorflow.html
So wanted to know if there are any recommended ways of installing TensorFlow with HDP
@Karthik Gopal I did see that link but no luck on finding an exact information on Hadoop integration with Tensorflow. Spark can run on HDFS and we can HDP on spark to use Tensorflow. Theoretically, This makes sense.
Do you have access to support?
On sandbox, install pip then follow the instructions for pip install, write an article about you success here, great question.
I have installed it on an HDF 2.0 Node on Centos7 and on an HDP 2.5 Node on Centos 7 on OpenStack. This will even install easier on Ubuntu.
Make sure you have Python 2.7+ and PIP installed.
sudo pip install --upgrade https://storage.googleapis.com/tensorflow/linux/cpu/tensorflow-0.7.1-cp27-none-linux_x86_64.whl
Tensorflow on Spark by Yahoo http://yahoohadoop.tumblr.com/post/157196317141/open-sourcing-tensorflowonspark-distributed-deep https://github.com/yahoo/TensorFlowOnSpark https://github.com/yahoo/TensorFlowOnSpark/wiki/GetStarted_YARN https://github.com/yahoo/TensorFlowOnSpark/wiki/Conversion https://github.com/yahoo/TensorFlowOnSpark/tree/master/examples/slim https://github.com/tensorflow/models/tree/master/slim