Applying word embedding function on a large pandas dataset

Question

We have a pandas dataframe with one column (\'message\') and 3.9 million rows and need to convert these messages to their word embeddings using Google\'s Universal Sentence