I would like to perform one round of inference on the training data before the actual training starts. What is the easiest and optimal way to do that?
I would like to perform one round of inference on the training data before the actual training starts. What is the easiest and optimal way to do that?