I have trained a model on vertex AI and created a pipeline run. I have uploaded the model on model registry. Suppose my model is getting data from outside for online prediction but that that data has only 20 columns of data and my model require 28 columns after some preprocessing . is there a way we can specify the preprocessing in the api endpoint to get 28 columns the use that data for prediction. This should be part of deployed model only.
LinkedIn
Twitter
