This repo shows the implementation of loading the word embeddings via Keras for text classification (binary but you can build a model for multiclass class-n as well).
Examples will be provided both for R and Python notebooks.
AUC in our example with a few epochs on a small data is ~ 0.98.