AI Model Packaging Component

1. Create a source folder and add all the files that needs to be packaged (entrypoint.py, requirements.txt).

2. Custom packaging Logic (pack.py), packages the model into an edge configuration package. You can define the input/output and dependencies.

3. Adding all files under src folder for packaging, with defining i/o types. If the model is MLFlow, specify the artifacts download path.

   
   
   

4. Packaging pipeline(ai_sdk_packaging.py), downloads the model & registers the model created.

Parent topic:AI Model Packaging & Delivery