TF 2.0 is out! Get hands-on practice at TF World, Oct 28-31. Use code TF20 for 20% off select passes. Register now


TensorFlow 1 version View source on GitHub

Exports a tf.keras.Model as a Tensorflow SavedModel. (deprecated)



Note that at this time, subclassed models can only be saved using serving_only=True.

The exported SavedModel is a standalone serialization of Tensorflow objects, and is supported by TF language APIs and the Tensorflow Serving system. To load the model, use the function tf.keras.experimental.load_from_saved_model.

The SavedModel contains:

  1. a checkpoint containing the model weights.
  2. a SavedModel proto containing the Tensorflow backend graph. Separate graphs are saved for prediction (serving), train, and evaluation. If the model has not been compiled, then only the graph computing predictions will be exported.
  3. the model's json config. If the model is subclassed, this will only be included if the model's get_config() method is overwritten.


import tensorflow as tf

# Create a tf.keras model.
model = tf.keras.Sequential()
model.add(tf.keras.layers.Dense(1, input_shape=[10]))

# Save the tf.keras model in the SavedModel format.
path = '/tmp/simple_keras_model'
tf.keras.experimental.export_saved_model(model, path)

# Load the saved keras model back.
new_model = tf.keras.experimental.load_from_saved_model(path)


  • model: A tf.keras.Model to be saved. If the model is subclassed, the flag serving_only must be set to True.
  • saved_model_path: a string specifying the path to the SavedModel directory.
  • custom_objects: Optional dictionary mapping string names to custom classes or functions (e.g. custom loss functions).
  • as_text: bool, False by default. Whether to write the SavedModel proto in text format. Currently unavailable in serving-only mode.
  • input_signature: A possibly nested sequence of tf.TensorSpec objects, used to specify the expected model inputs. See tf.function for more details.
  • serving_only: bool, False by default. When this is true, only the prediction graph is saved.


  • NotImplementedError: If the model is a subclassed model, and serving_only is False.
  • ValueError: If the input signature cannot be inferred from the model.
  • AssertionError: If the SavedModel directory already exists and isn't empty.