Installing IBM Watson Machine Learning Accelerator with WML or Watson Studio on IBM Cloud Pak for Data 3.5.3
After you install WML Accelerator 1.2.3 you can connect it to IBM Watson® Machine Learning or IBM Watson Studio Local on IBM® Cloud Pak for Data 3.5.3.
About this task
If you plan on using WML Accelerator with IBM Watson Studio Local or WML, you must complete the following after installing WML Accelerator.
Before you begin- Ensure that you have installed WML Accelerator: Installing WML Accelerator. Note:
- If you are installing WML Accelerator for the first time, or upgrading from a previous version of WML Accelerator, you must configure or rebuild your Anaconda environment to use the WML Accelerator deep learning libraries package: Configure a system for deep learning
- As part of the WML Accelerator installation, you created two instance groups, one distributed training (using the wmla-ig-template-2.3.3 template) and one for elastic distributed training (using the wmla-ig-edt-template-2.3.3 template). These instance groups will be used by WML when pushing training jobs to WML Accelerator.
- Ensure that you have installed WML: Installing IBM Watson Machine Learning.
- Review the list of known issues and limitations: Known issues in the WML integration
- Consider the following regarding user authentication and execution:
- If LDAP is used with IBM Watson Studio Local, a common LDAP server can be used by both WML Accelerator and IBM Watson Studio Local for storing user credentials. To do so, you must configure LDAP for WML Accelerator, see: Configuring user authentication for PAM and default clients
- If LDAP is not used, each IBM Watson Studio Local user that wants to run training jobs, must be added to WML Accelerator. You must create an OS user with the same username. You must ensure that each user has the same UID, group ID (GID), and password on all hosts in the cluster.
- In either case, users must be assigned to the instance groups and they must have a Data Scientist or Consumer user role. Roles can be assigned by the Consumer administrator or the cluster administrator. See: Adding a consumer user or Assigning roles to users or user groups
Procedure
Results
What to do next
Get started
After you have successfully connected WML Accelerator with Watson Machine Learning and Watson Studio, here are a few links to get you started:Notes:
- Before adding any hosts to the cluster, you must manually install the WML Accelerator license conda package and accept the license on the management or compute host that is being added to the cluster. See this topic for instructions: Installing the WML Accelerator license on an additional node.
- When installing IBM Spectrum Conductor™ on a compute node, do not add it to cluster until after installing Deep Learning Impact on the compute node.