Keras Models --> TensorFlow SavedModel format

export.py

# Mostly copied from https://keras.io/applications/#usage-examples-for-image-classification-models                                           
# Changing it to use InceptionV3 instead of ResNet50                                                                                         
from keras.applications.inception_v3 import InceptionV3, preprocess_input, decode_predictions                                                
from keras.preprocessing import image                                                                                                        
import numpy as np                                                                                                                           
                                                                                                                                             
model = InceptionV3()                                                                                                                        
img_path = 'elephant.jpg'                                                                                                                    
img = image.load_img(img_path, target_size=(299, 299))                                                                                       
x = image.img_to_array(img)                                                                                                                  
x = np.expand_dims(x, axis=0)                                                                                                                
x = preprocess_input(x)                                                                                                                      
preds = model.predict(x)                                                                                                                     
print('Predicted:', decode_predictions(preds, top=3)[0])                                                                                     
                                                                                                                                             
# And now exporting to the TensorFlow SavedModel format.                                                                                     
# Documentation for the SavedModel format can be found here:                                                                                 
# https://github.com/tensorflow/tensorflow/blob/master/tensorflow/python/saved_model/README.md                                               
from keras import backend as K                                                                                                               
import tensorflow as tf                                                                                                                      
                                                                                                                                             
signature = tf.saved_model.signature_def_utils.predict_signature_def(                                                                        
    inputs={'image': model.input}, outputs={'scores': model.output})                                                                         
                                                                                                                                             
builder = tf.saved_model.builder.SavedModelBuilder('/tmp/my_saved_model')                                                                    
builder.add_meta_graph_and_variables(                                                                                                        
    sess=K.get_session(),                                                                                                                    
    tags=[tf.saved_model.tag_constants.SERVING],                                                                                             
    signature_def_map={                                                                                                                      
        tf.saved_model.signature_constants.DEFAULT_SERVING_SIGNATURE_DEF_KEY:                                                                
            signature                                                                                                                        
    })                                                                                                                                       
builder.save()
       
       
# This model can be loaded in other langauges using the C API:                                                                               
# TF_SessionOptions* opts = TF_NewSessionOptions();                                                                                          
# const char* tags[] = {"serve"};  // tf.saved_model.tag_constants.SERVING                                                                   
# TF_LoadSessionFromSavedModel(opts, NULL, "/tmp/my_saved_model", tags, 1, graph, NULL, status);
#
# This is what is used by the:
# - Java API: https://www.tensorflow.org/api_docs/java/reference/org/tensorflow/SavedModelBundle
# - Go API: https://godoc.org/github.com/tensorflow/tensorflow/tensorflow/go#LoadSavedModel
# etc.

Thanks for sharing. This helped me save my tensorflow.keras models.