What does the presenter mean by 'serving dataset'?

I do not think I fully understand what the presenter means by the serving dataset. Could someone explain it to me because I know that it is a very important concept.

Serving dataset refers to the data your model is responding to in the real environment. One can create a dataset of the production environment from logs or streaming the model inputs / outputs to a store.


Hi @Nathan_Angell ,

Can you please elaborate a bit more?

You have TensorFlow Datasets and TensorFlow Serving.

Just to remind that TensorFlow Datasets is a collection of datasets ready to use. They are exposed as tf.data.Datasets

TensorFlow Serving is a flexible, high-performance serving system for machine learning models, designed for production environments. TensorFlow Serving makes it easy to deploy new algorithms and experiments.


Please confirm what serving dataset refers to with respect to the lecture.

Two types of datasets exist - training dataset is used to build/train model and the serving dataset is what the users are requesting predictions for.