Using custom model and ec2 instance in sagemaker ground truth

[harit7]

I am trying to run a labelling job for which the ml model can be run on cheaper ec2-instances. However I do not want to change the default active learning algorithm of sagemaker ground truth. It seems in the notebook on using custom model uses step functions to rewrite the entire active learning workflow. I was wondering if it is possible to use a custom model/ec2-instance with minimal changes to the default sagemaker groundtruth settings?
@ajaykarpur