Learnt how to use pre-trained models from TensorFlow Hub with tf.keras for text classification. Transfer learning makes it possible to save training resources and to achieve good model generalization even when training on a small dataset. Demonstration was done by training with several different TF-Hub modules.