Training Flow Overview
About Estimator
Estimator is a high-level API of TensorFlow and is introduced in TensorFlow 1.10 released in 2018. It greatly streamlines the programming process of machine learning. Estimator has many advantages, for example, good support for distribution, simplified model creation, and code sharing between model developers.
To use the Estimator API to develop a training script, perform the following steps.
Step |
Description |
|---|---|
Preprocess the data. |
Create the input function input_fn. |
Construct a model. |
Construct the model function model_fn. |
Configure run parameters. |
Instantiate Estimator and pass an object of the Runconfig class as the run parameter. |
Perform training. |
Call the training method Estimator.train() in Estimator to train the model with a fixed number of steps using the specified input. |