Hello all.
I have completed the machine learning courses and the help but am struggling to put it together. I have a spreadsheet with data that I am using to predict the value in two columns out of a given set of values.
So I figure I need to use a classification model. I logged into vertexAI and created a bucket for myself called my_apps next I went into vertexAI and created a dataset in us-central1. I then clicked on the “Train New Model” button figuring that I would use autoML as I was just getting started. This led me to the AutoML on pipelines page. Next I chose “AutoML” and walked through the “Create a pipeline” wizard. I set the template gallery pipeline to “Auto ML for Tabular cladssification”, region as us-central1, one off run schedule, output cloud storage location as my bucket, objective as classification, target column as “Customer”, checked “create an evaluation on the trained model”, Next in the Training Options I removed the columns that I did not want to include in the training after generating training statistics and correlations. Next I set the budget to 1 as I have <100k data rows and clicked on submit.
Unfortunately I got this message.
To run this pipeline, the service account “122065963851-compute@developer.gserviceaccount.com” needs the following roles and/or permissions:
- storage.objects.get
- storage.objects.create
However when I went to the IAM to create the permissions for the service account, I could not fine the roles specified in the message above. What am I missing? Is it expecting me to create the roles first myself?
Thanks in advance.