Difference between types of predict

[LorenzoF6]

If i want to perform predictions on a model (such as patchcore) it is the best choice to use the .predict() function from the class Engine or is a better solution to use torch or opevino inferencer not only for inference time but also for type of output?
And if i want to see the result of inference is correct to use the Visualizer API from the library?
Thansk