Distributed Machine Learning with Google Cloud ML
Couple of issues: In several places ml engine is still used instead of ai platform, and since the introduction of global students will need to set up a region in order for the lab to finish. The point grader will only accept us-central1, so I had to manually create a model resource and deploy it there for us-central1, in order for the grader to work. By default the commands in the tutorial created a resource in global for me.
would be nice if both python version and tensorflow updated
ML Model command on the last "Deploying" section should be modified, not to provide region details. Otherwise, ML model is created with the wrong endpoint and the next command for the V1 creation fails. Please, update.
There is a problem in last stage, Deployment of model script is not working due to region is not correctly specified.
The last step did not register for points, even though it all worked and the python script returned a response from the cloud endpoint.