This model service example is based on the "Models-as-a-Service" reference implementation found here: https://github.com/rh-aiservices-bu/models-aas.
To stand up your own instance, follow the instructions in the reference implementation's repository, and select granite-8b-code-instruct-4k as the deployed model.
The following model is deployed in the example:
- granite-8b-code-instruct
If you wish to replicate this example:
-
Fork this repository.
-
Follow the instructions above to deploy the model server and models.
-
Modify the
catalog-info.yamlas needed: replace any links (e.g. anydomain.comreferences), and update anyResourcesas needed (for example, adding/removing models). -
Select
Register Existing Componentin Backstage, and copy the catalog link into import form.