What are the different methods and tools to deploy a TensorFlow model on GCP? What is TensorFlow Serving, AI Platform Predictions, Cloud Functions, and how do they compare? This video should also make clear that TF Serving is a good, but not the only way, to put models into production on GCP.