ValueError: No variables to save

gistfile1.txt

#!/usr/bin/python
import time
import os
import tensorflow as tf
import numpy as np
import skimage
import skimage.io
import skimage.transform

import freeze_graph


def outputGraph(sess, graph):
    with graph.as_default():
        checkpoint_dir = '/home/shane/code/tensorflow/graph_saver/'
        checkpoint_state_name = "save_checkpoint.ckpt"
        input_graph = "input_graph.pb"
        output_graph_name = "output_graph.pb"

        # printing variable lists
        print(tf.get_collection(tf.GraphKeys.VARIABLES)) #deprecated
        print(tf.all_variables()) #deprecated
        print(tf.global_variables())
        print(tf.local_variables())
        print(tf.model_variables())
        print(tf.trainable_variables())
        print(tf.moving_average_variables())

        print("Saving to checkpoint")
        saver = tf.train.Saver()
        # Fails on this line
        # ValueError: No variables to save
        saver.save(sess, checkpoint_dir + checkpoint_state_name, global_step=0,
                   latest_filename=checkpoint_state_name)

        print("Saving the graph")
        tf.train.write_graph(graph, checkpoint_dir, input_graph)

        input_graph_path = os.path.join(checkpoint_dir, input_graph)
        input_saver = ""
        input_binary = False
        input_checkpoint_path = os.path.join(checkpoint_dir, checkpoint_state_name) + "-0.index"

        output_node_names = "import/softmax:0"
        restore_op_name = "save/restore_all"
        filename_tensor_name = "save/Const:0"
        output_graph_path = os.path.join(checkpoint_dir, output_graph_name)
        clear_devices = False
        initializer_nodes = ""

        freeze_graph.freeze_graph(input_graph_path,
                                  input_saver,
                                  input_binary,
                                  input_checkpoint_path,
                                  output_node_names,
                                  restore_op_name,
                                  filename_tensor_name,
                                  output_graph_path,
                                  clear_devices,
                                  initializer_nodes)

def main():
    with tf.device('/cpu:0'):
        with open("tensorflow_inception_graph.pb", mode='rb') as f:
            fileContent = f.read()

        graph_def = tf.GraphDef()
        graph_def.ParseFromString(fileContent)

        images = tf.placeholder("float", [ None, 299, 299, 3])

        tf.import_graph_def(graph_def, input_map={ "Mul": images })
        print "graph loaded from disk"

        graph = tf.get_default_graph()
        im = load_image("grace_hopper.jpg")

        with tf.Session() as sess:
            init = tf.global_variables_initializer()

            sess.run(init)
            print "variables initialized"

            batch = im.reshape((1, 299, 299, 3))
            assert batch.shape == (1, 299, 299, 3)

            feed_dict = { images: batch }

            prob_tensor = graph.get_tensor_by_name("import/softmax:0")

            # printing variable lists
            print(tf.get_collection(tf.GraphKeys.VARIABLES)) #deprecated
            print(tf.all_variables()) #deprecated
            print(tf.global_variables())
            print(tf.local_variables())
            print(tf.model_variables())
            print(tf.trainable_variables())
            print(tf.moving_average_variables())

            ITERATIONS = 10
            total = 0
            for i in range(ITERATIONS):
                print(i)
                t1 = time.time()
                prob = sess.run(prob_tensor, feed_dict=feed_dict)
                t2 = time.time()
                total += (t2-t1)
                print_prob(prob[0])
            
            print("%fms per run\n" % ((total/ITERATIONS)*1000))
            
            outputGraph(sess, graph)
            sess.close()

################## Helper Functions #####################

# returns image of shape [299, 299, 3]
# [height, width, depth]
def load_image(path):
  # load image
  img = skimage.io.imread(path)
  img = img / 255.0
  assert (0 <= img).all() and (img <= 1.0).all()
  #print "Original Image Shape: ", img.shape
  # we crop image from center
  short_edge = min(img.shape[:2])
  yy = int((img.shape[0] - short_edge) / 2)
  xx = int((img.shape[1] - short_edge) / 2)
  crop_img = img[yy : yy + short_edge, xx : xx + short_edge]
  # resize to 299, 299
  resized_img = skimage.transform.resize(crop_img, (299, 299))
  return resized_img

# returns the top1 string
def print_prob(prob):
  synset = [l.strip() for l in open('imagenet_comp_graph_label_strings.txt').readlines()]
  #print prob
  print "prob shape", prob.shape
  pred = np.argsort(prob)[::-1]

  # Get top1 label
  top1 = synset[pred[0]]
  print "Top1: ", top1
  # Get top5 label
  top5 = [synset[pred[i]] for i in range(5)]
  print "Top5: ", top5
  return top1

main()