Acumos marketplace download of onboarded model docker image

Question

I managed to generate and successfully onboard the Iris classification example (Python) model (https://github.com/acumos/acumos-python-client/blob/master/examples/tensorflow_example.py) to Acumos (our own installation). I then downloaded the Docker image of the Acumos created model to my local machine using the "Download" button and selecting the model_xxxx.tar file. I imported the Docker image to Docker and started it using docker run (docker run -d --name iris -p 127.0.0.1:3330:3330 id). I expected that there would be a Swagger API description at http://127.0.0.1:3330 but there was not. I am misunderstanding something here? I looks like that there is a listener at http://127.0.0.1:3330/classify but I have not been able to verify that since I do not know how to invoke it.

Grateful for all help :-). I have tried to locate a coherent tutorial describing how to actually do this but with no luck so far.

Answer 1

Please see the soup-to-nuts guide which describes in some detail how to invoke the service in a deployed Acumos model:

https://wiki.acumos.org/display/AC/Soup-to-Nuts+Example%3A+Onboarding%2C+Downloading%2C+Deploying%2C+and+Using+a+Python-Based+Model+in+Acumos

You might want to comment on this open Jira issue about providing a Swagger endpoint:

https://jira.acumos.org/browse/ACUMOS-732

Answer 2

You can open Docker Desktop, run the container and from the Container list, hover mouse over the container name and click on Open in Browser. This will show you the Swagger API description with all methods and artifacts, and of course will allow you to try it out.