In order yo create a named entity recognition (NER) classifier model based on your own texts you need to tokenize and format the text document before you annotate it.