Confidence Scores for custom SpaCy NER Model

[davesuketu215]

I would like to get the confidence score of my custom model in predicting a certain entity label in unseen data.
Is there any convenient approach to get the confidence scores on the prediction of entities for SpaCy, apart from using the beam parser?
I know that the Scores module of SpaCy gives precision, recall, f1-scores. But rather what I want is something different as stated above. Any techniques to use?

[davesuketu215]

Is Beam Parser the only way to get "Confidence Score" of a Spacy Model? Or can we derive confidence score from the output of nlp.evaluate ? Or can any other API help here?
I am really new to Spacy NER world and so need a small guidance on this aspect. I know my question is trivial and silly but a clarification would help a noob like me. :)