Preparing to install IBM watsonx.ai

Before you install the IBM watsonx.ai service, make sure that your cluster meets the system requirements for the IBM watsonx.ai service depending on the installation mode you use and the models you choose to install.

  1. Choose whether to install one or both of the watsonx.ai™ service components: the core component or the model gateway component.

    For details, see Specifying installation components.

  2. Choose whether to install the full service or watsonx.ai lightweight engine.

    For details, see Choosing an IBM watsonx.ai installation mode.

  3. Choose the foundation models that you want to install. You can also bring you own custom foundation models after the service is installed.

    Gather the extra resources that are needed to support hosting foundation models for inferencing, including specialized GPUs. The number and size of GPUs required depends on the foundation models that you want to install. You can optionally configure NVIDIA Multi-Instance GPU (MIG) to split one GPU into multiple GPU instances on a single node.

    For details about the types of foundation models you can deploy, see System requirements for foundation models in IBM watsonx.ai and Partitioning GPU processors in IBM watsonx.ai.