Serving a TensorFlow Model

This tutorial shows you how to use TensorFlow Serving components to export a trained TensorFlow model and use the standard tensorflow_model_server to serve it. If you are already familiar with TensorFlow Serving, and you want to know more about how the server internals work, see the TensorFlow Serving advanced tutorial.

This tutorial uses the simple Softmax Regression model introduced in the TensorFlow tutorial for handwritten image (MNIST data) classification. If you do not know what TensorFlow or MNIST is, see the MNIST For ML Beginners tutorial.

The code for this tutorial consists of two parts:

  • A Python file ( that trains and exports the model.

  • A C++ file which is the standard TensorFlow model server that discovers new exported models and runs a gRPC service for serving them.

Before getting started, please complete the prerequisites.

Train And Export TensorFlow Model

As you can see in, the training is done the same way it is in the MNIST For ML Beginners tutorial. The TensorFlow graph is launched in TensorFlow session sess, with the input tensor (image) as x and output tensor (Softmax score) as y.

Then we use TensorFlow Serving SavedModelBuilder module to export the model. SavedModelBuilder saves a “snapshot” of the trained model to reliable storage so that it can be loaded later for inference.

from tensorflow.python.saved_model import builder as saved_model_builder
export_path_base = sys.argv[-1]
export_path = os.path.join(
print 'Exporting trained model to', export_path
builder = saved_model_builder.SavedModelBuilder(export_path)
      sess, [tag_constants.SERVING],

SavedModelBuilder.__init__ takes the following argument:

  • export_path is the path of the export directory. SavedModelBuilder will create the directory if it does not exist. In the example, we concatenate the command line argument and FLAGS.model_version to obtain the export directory. FLAGS.model_version specifies the version of the model. You should specify a larger integer value when exporting a newer version of the same model. Each version will be exported to a different sub-directory under the given path.

You can add meta graph and variables to the builder using SavedModelBuilder.add_meta_graph_and_variables() with the following arguments:

  • sess is the TensorFlow session that holds the trained model you are exporting.

  • tags is the set of tags with which to save the meta graph.

  • signature_def_map specifies the map of user-supplied key for a signature to a tensorflow::SignatureDef to add to the meta graph. Signature specifies what type of model is being exported, and the input/output tensors to bind to when running inference. The special signature key serving_default specifies the default serving signature. signature_def_utils.build_signature_def() accepts the following arguments:

    • inputs={'images': tensor_info_x} specifies the input tensor info.

    • outputs={'scores': tensor_info_y} specifies the scores tensor info.

    • images and scores are tensor alias names. They can be whatever unique strings you want, and they will become the logical names of tensor x and y that you refer to for tensor binding when sending prediction requests later. For instance, if x refers to the tensor with name ‘long_tensor_name_foo’ and y refers to the tensor with name ‘generated_tensor_name_bar’, builder will store tensor logical name to real name mapping (‘images’ -> ‘long_tensor_name_foo’ and ‘scores’ -> ‘generated_tensor_name_bar’) and allow user to refer to these tensors with their logical names when running inference.

    • method_name is the method used for the inference. For Prediction requests, it should be set to tensorflow/serving/predict.

Let’s run it!

Clear the export directory if it already exists:

$>rm -rf /tmp/mnist_model
$>bazel build //tensorflow_serving/example:mnist_saved_model
$>bazel-bin/tensorflow_serving/example/mnist_saved_model /tmp/mnist_model
Training model...


Done training!
Exporting trained model to /tmp/mnist_model
Done exporting!

Now let’s take a look at the export directory.

$>ls /tmp/mnist_model

As mentioned above, a sub-directory will be created for exporting each version of the model. FLAGS.model_version has the default value of 1, therefore the corresponding sub-directory 1 is created.

$>ls /tmp/mnist_model/1
saved_model.pb variables

Each version sub-directory contains the following files:

  • saved_model.pb is the serialized tensorflow::SavedModel. It includes the the one or more graph definitions of the model, as well as metadata of the model such as signatures.

  • variables are files that hold the serialized variables of the graphs.

With that, your TensorFlow model is exported and ready to be loaded!

Load Exported Model With Standard TensorFlow Model Server

$>bazel build //tensorflow_serving/model_servers:tensorflow_model_server
$>bazel-bin/tensorflow_serving/model_servers/tensorflow_model_server --port=9000 --model_name=mnist --model_base_path=/tmp/mnist_model/

Test The Server

We can use the provided mnist_client utility to test the server. The client downloads MNIST test data, sends them as requests to the server, and calculates the inference error rate.

To run it:

$>bazel build //tensorflow_serving/example:mnist_client
$>bazel-bin/tensorflow_serving/example/mnist_client --num_tests=1000 --server=localhost:9000
Inference error rate: 10.5%

We expect around 91% accuracy for the trained Softmax model and we get 10.5% inference error rate for the first 1000 test images. This confirms that the server loads and runs the trained model successfully!