I am building a text classification using Keras and I am using GloVe to initialize the weights. I was wondering if it is possible to construct a bigram word embedding?
