This blog walks you through the steps to deploy IBM Watson® Machine Learning Community Edition (WML CE) on your IBM Cloud™ Virtual Private Cloud (VPC) on an IBM Power® environment in the next gen IBM Cloud®.

How to get WML CE 1.70?

WML CE 1.7.0 is available as a no charge offering from IBM.

What is WML CE?

A little less than a year ago, we released Watson Machine Learning Community Edition (WML CE) 1.6.0, which was a ground up reorganizing, refocusing, and retooling of our IBM PowerAI product. We started up the continuous delivery engine, wrapped it in a flexible Linux-distribution-independent conda-themed paint job and have been tirelessly working ever since to make our distribution of artificial intelligence and machine learning packages the most consumable, secure, and enterprise-friendly set available anywhere.

WML CE is a no charge offering that is based on open source software. It is a bundle that includes conda packaging, updated versions of the most popular deep learning and machine learning frameworks such as TensorFlow and Caffe, and a simplified and unified package dependency control. For more information and a deeper technical dive into WML CE, see the following links:

Deploying WML CE in VPC on Power

The following section walks you through the steps for deploying WML CE in your VPC environment.

Step 1: Create a virtual server instance (VSI) in your VPC environment

You can find step-by-step guide for creating a VSI instance at:
https://developer.ibm.com/linuxonpower/2020/04/09/getting-started-with-virtual-private-cloud-vpc-on-ibm-power/

For WML CE, it is recommended to provision the 24 vCPU, 224 GB RAM, and two NVIDIA 32 GB Tesla V100. This profile can be accessed by selecting the following options: POWER for Processor architecture, Ubuntu Linux for Image, All profiles, and GPU as shown in the following figure.

fig1

After creating VSI, attach a floating IP address to your VSI instance. This IP address uses Secure Shell (SSH) to connect to our VSI.

fig2

GPU profiles do take a bit more time to start up. After the GPU profile is up and running, we can use SSH to connect to our VSI.

As a troubleshooting tip, if at any point in the process an issue arises, stopping and restarting the instance from the CLI can often be the solution.

Step 2: Install NVIDIA drivers

To install the NVIDIA drivers in your VSI, run the following commands.

#Install NVIDIA Driver
apt update
apt dist-upgrade
wget http://us.download.nvidia.com/tesla/440.33.01/nvidia-driver-local-repo-ubuntu1804-440.33.01_1.0-1_ppc64el.deb
sudo dpkg -i nvidia-driver-local-repo-ubuntu1804-440.*.deb
sudo apt-key add /var/nvidia-driver-local-repo-440.*/*.pub
sudo apt-get update
sudo apt-get install cuda-drivers
systemctl daemon-reload
systemctl enable nvidia-persistenced
export WMLCE_VERSION=1.7.0
export PYTHON_VERSION=3.7

While running the apt dist-upgrade command, you will be prompted to configure the cloud.cfg file. Enter Y for this option.

fig3

To check if the update is completed correctly, use the nvidia-smi command.

fig4

Step 3: Install Miniconda

Run the following commands to install Miniconda:

#Install Miniconda
wget https://repo.anaconda.com/miniconda/Miniconda3-latest-Linux-ppc64le.sh
bash Miniconda3-latest-Linux-ppc64le.sh -b -f
rm Miniconda3-latest-Linux-ppc64le.sh

Step 4: Get Conda command to work

To get the conda command to work, run the following command:

#Get conda command to work
eval "$($HOME/miniconda3/bin/conda shell.bash hook)"

Step 5: Set up conda environment

Run the following commands to set up the conda environment.

#Setup conda environment
conda config --prepend channels https://public.dhe.ibm.com/ibmdl/export/pub/software/server/ibm-ai/conda/
conda create --name wmlce_env_${WMLCE_VERSION}  python=${PYTHON_VERSION} -y
conda activate wmlce_env_${WMLCE_VERSION}

Again, the WMLCE_VERSION right now is 1.7.0 and the PYTHON_VERSION is 3.7

fig5

Step 6: Install WML CE

Run the following commands to install WML CE:

#Install wmlce
export IBM_POWERAI_LICENSE_ACCEPT=yes
conda install powerai=${WMLCE_VERSION} -y

fig6

After the installation is complete, the following output is displayed.

fig7

Install all of WML CE / Install only Tensorflow

#Install all of WMLCE
conda install powerai=${WMLCE_VERSION} -y

#Install only tensorflow
conda install tensorflow-gpu powerai-release=${WMLCE_VERSION} -y

Wrap up

This blog provided step-by-step instructions for deploying WML CE in the IBM Cloud VPC on Power environment. You should now have a VSI that can use the most popular deep learning and machine learning frameworks.

Join The Discussion

Your email address will not be published. Required fields are marked *