I\'m trying to use fasttext word embeddings as input for a SVM for a text classification task. I averaged the word vectors over each sentence, and for each sentence I want t