Updates to Lite Plan: AI OpenScale service
  • IBM Cloud Platform
  • Dallas
  • Description
    Effective December 14, 2018, we are making changes to inclusions in the Lite plan of IBM Cloud AI OpenScale. IBM AI OpenScale on IBM Cloud provides an enterprise-grade environment to operate and optimize AI applications with trust and transparency. AI OpenScale is currently available as a Lite plan with a limit of up to 20 deployed models that can be monitored.

    New limits on instances
    As we introduce a paid, Standard plan for the service, we are making some changes to the Lite plan to bring newer capabilities to users. In addition to the bias detection and explainability functionality, we are also introducing the ability to de-bias transactions in runtime. The following limits will apply to new and existing instances of the AI OpenScale service beginning December 14, 2018:
    • Monitor: The ability to monitor a maximum of two deployed models at any time
    • Explainability: Twenty transactions explained per month across a maximum of five deployed models
    • Bias detection: Two protected features per model and 50,000 payload rows across a maximum of five deployed models per month
    • De-baising: Two protected features per model and 50,000 rows across a maximum of five deployed models per month
    • Quality monitoring: 50,000 feedback records across a maximum of five models per month
    This information originated in the IBM AI OpenScale: Updates to Lite Plan article within the IBM Cloud platform blog.