Visualizing the learned embeddings