Chapter 2. New features and enhancements

This release introduces the Caikit Standalone ServingRuntime for KServe, a preinstalled model-serving runtime for the single-model serving platform that supports embeddings models.

With an upgraded version of Caikit-NLP, the runtime includes support for the Embeddings Service, providing inference endpoints for embedding, sentence similarity, and re-ranking tasks.

You can access the endpoints with REST protocol by default. You can also manually configure the endpoints to use gRPC protocol.

For more information, see Model-serving runtimes.

You can now create and use experiments for your data science pipelines. Experiments are workspaces where you can try different configurations of your pipelines. You can also use experiments to organize your pipeline runs into logical groups.

From the Experiments tab of the OpenShift AI dashboard, you can track your pipeline experiments, compare experiment runs, view and track run artifacts, and visualize run metrics.

You can customize the metrics columns of the experiment runs table to display the metrics that are relevant to your use case.

You can also compare metrics for up to 10 runs within an experiment, and view available parameter, scalar metric, confusion matrix, and receiver operating characteristic (ROC) curve data for all selected runs.

For more information, see Managing pipeline experiments.

Elasticsearch is now available as an integrated partner solution in OpenShift AI.

Elasticsearch encompasses all the tools that developers need to build next generation search experiences with generative AI, including a vector database, the ability to use multiple models, and powerful search capabilities for Retrieval Augmented Generation (RAG).

To use Elasticsearch, you need to install the Elastic Operator. For more information, see Elasticsearch (ECK) Operator.

After installing the Elastic Operator, you can enable Elasticsearch under Applications Explore on the OpenShift AI dashboard.