-
Story
-
Resolution: Done
-
Major
-
None
-
5
-
False
-
-
False
-
-
-
DEVAI Sprint 3278, DEVAI Sprint 3279, DEVAI Sprint 3280, DEVAI Sprint 3281
Story (Required)
As a user of Software Templates trying to use a model from model catalog to tie into my application I want a reference to the catalog entry to avoid needing to input the required information manually.
The ability to selecting a model from the model catalog under the AI Software Templates would make the process of connecting a model from the catalog more streamlined.
Background (Required)
The rolling demo (https://github.com/redhat-ai-dev/ai-lab-template/pull/72) showcased choosing a model server from the model catalog when creating a ChatBot application using Software Templates. With Full Bridge Functionality being prioritized for RHDH 1.8 release plan, having this feature in Software Templates would allow a user to take better advantage of this functionality.
Out of scope
Model Catalog Bridge / scaffolder features (i.e. display naming and data preprocessing requirements)
Approach (Required)
Create an additional entry for toggling "Model server from the catalog" to enable the path of selecting a model from the catalog. This can be enabled/disabled by an environment variable under the AI Software Template skeleton.
We will use dropdown groupings from the UX design mockups to separate model server paths by their group, "Select an existing model server" / "Create a new model server".
For this path, use the ai-model entity picker to choose the model server endpoint and model name of the existing model/model server.
Dependencies
<Describes what this story depends on. Dependent Stories and EPICs should be linked to the story.>
Acceptance Criteria (Required)
Implement additional and unselectable model server entries "Select an existing model server" and "Create a new model server" to serve as groupings for the existing entries, see UX design mockups for reference
Implement additional model server entry "Model server from the catalog" under "Select an existing model server" grouping for enabling the path of selecting a model from the catalog
Implement ai-model entity picker for LLM templates and model-server entity picker for non-LLM templates that shows when "Model server from the catalog" path is selected
Implement same barer token properties as the "Existing model server" path when "Model server from the catalog" path is selected
Implement action to pull required model server information from the chosen ai-model or model-server catalog entity to pass to GitOps template for the model server connection (should be the same values normally inputted by the user with the "Existing model server" path)
Test template changes on an instance with a model catalog to ensure an application can be deployed connected to an existing model server tied to the selected catalog entity
documentation updates (design docs, release notes etc)
demo needed
SOP required
education module update (Filled by RHDHPAI team only)
R&D label required (Filled by RHDHPAI team only)
Done Checklist
Code is completed, reviewed, documented and checked in
Unit and integration test automation have been delivered and running cleanly in continuous integration/staging/canary environment
Continuous Delivery pipeline(s) is able to proceed with new code included
Customer facing documentation, API docs, design docs etc. are produced/updated, reviewed and published
Acceptance criteria are met
If the Grafana dashboard is updated, ensure the corresponding SOP is updated as well
- is blocked by
-
RHIDP-10594 Add API URL(s) annotations to model server entities within model catalog
-
- Closed
-