Data Science now Supports Text Embeddings Inference (TEI) Framework

Data Science AI Quick Actions now supports TEI (Text Embeddings Inference) frame work. You can use TEI through a Bring Your Own Container approach:
  • Push the container image to OCIR.
  • Select TEI as inference container.
  • Specifying the image URI during model registration.

e5-mistral-7b-instruct has been added as a service verified model

The Environment Explorer now only has two tabs:
  • Base models
  • Fine-tuned models.
Under Base models, you can find the service cached models, service verified models that you can register to use in AI Quick Actions and any model you have registered to bring into the service. Each of these models has its own model card with tags such as "Ready to Deploy" and "Ready to Register" to help you differentiate them.

For more information, see the Data Science documentation on Service Managed Inference Containers.