Hiii, I am using the pre-trained word embeddings for text classification. In general top most frequency vocabulary words which are present in the training dataset are used. But here I want to give some selective words from the training data as a features for text classification. Is there any function for using customized vocabulary?. please suggest me.