I am trying to create a question-answering model with the word embedding model BERT from google. I am new to this and would really want to use my own corpus for the training
