Fetch argument <tf.Tensor 'batch:0' shape=(128, 56, 56, 3) dtype=float32> cannot be interpreted as a Tensor.

Question

I write a test code,and when I run it ,it said Fetch argument cannot be interpreted as a Tensor.I really don't know what's going on .Can somebody tell me how to fix it? thank you very much .Here is the code

# coding=utf-8
from  color_1 import read_and_decode, get_batch, get_test_batch
import color_inference
import cv2
import os
import time
import numpy as np
import tensorflow as tf
import color_train
import math

EVAL_INTERVAL_SECS=10
batch_size=128
num_examples = 10000
crop_size=56
def test(test_x, test_y):
    with tf.Graph().as_default() as g:
        image_holder = tf.placeholder(tf.float32, [batch_size, 56, 56, 3], name='x-input')
        label_holder = tf.placeholder(tf.int32, [batch_size], name='y-input')

        y=color_inference.inference(image_holder)

        num_iter = int(math.ceil(num_examples / batch_size))
        true_count = 0
        total_sample_count = num_iter * batch_size
        saver=tf.train.Saver()
        top_k_op = tf.nn.in_top_k(y, label_holder, 1)
        while True:
            with tf.Session() as sess:
                ckpt=tf.train.get_checkpoint_state(color_train.MODEL_SAVE_PATH)
                if ckpt and ckpt.model_checkpoint_path:
                    saver.restore(sess,ckpt.model_checkpoint_path)
                    global_step=ckpt.model_checkpoint_path.split('/')[-1].split('-')[-1]
                    image_batch, label_batch = sess.run([test_x, test_y])
                    predictions = sess.run([top_k_op], feed_dict={image_holder: image_batch,
                                                                  label_holder: label_batch})
                    true_count += np.sum(predictions)
                    precision = true_count * 1.0 / total_sample_count
                    print("After %s training step,the prediction is :%g",global_step,precision)
                else:
                    print('No checkpoint file found')
                    return
            time.sleep(EVAL_INTERVAL_SECS)

def main(argv=None):
    test_image, test_label = read_and_decode('val.tfrecords')
    test_images, test_labels = get_test_batch(test_image, test_label, batch_size, crop_size)
    test(test_images, test_labels)

if __name__=='__main__':
    tf.app.run()

And the error is here:

File "/home/vrview/tensorflow/example/char/tfrecords/color_test.py", line 57, in <module>
    tf.app.run()
  File "/home/vrview/tensorflow/local/lib/python2.7/site-packages/tensorflow/python/platform/app.py", line 44, in run
    _sys.exit(main(_sys.argv[:1] + flags_passthrough))
  File "/home/vrview/tensorflow/example/char/tfrecords/color_test.py", line 54, in main
    test(test_images, test_labels)
  File "/home/vrview/tensorflow/example/char/tfrecords/color_test.py", line 39, in test
    image_batch, label_batch = sess.run([test_x, test_y])
  File "/home/vrview/tensorflow/local/lib/python2.7/site-packages/tensorflow/python/client/session.py", line 767, in run
    run_metadata_ptr)
  File "/home/vrview/tensorflow/local/lib/python2.7/site-packages/tensorflow/python/client/session.py", line 952, in _run
    fetch_handler = _FetchHandler(self._graph, fetches, feed_dict_string)
  File "/home/vrview/tensorflow/local/lib/python2.7/site-packages/tensorflow/python/client/session.py", line 408, in __init__
    self._fetch_mapper = _FetchMapper.for_fetch(fetches)
  File "/home/vrview/tensorflow/local/lib/python2.7/site-packages/tensorflow/python/client/session.py", line 230, in for_fetch
    return _ListFetchMapper(fetch)
  File "/home/vrview/tensorflow/local/lib/python2.7/site-packages/tensorflow/python/client/session.py", line 337, in __init__
    self._mappers = [_FetchMapper.for_fetch(fetch) for fetch in fetches]
  File "/home/vrview/tensorflow/local/lib/python2.7/site-packages/tensorflow/python/client/session.py", line 238, in for_fetch
    return _ElementFetchMapper(fetches, contraction_fn)
  File "/home/vrview/tensorflow/local/lib/python2.7/site-packages/tensorflow/python/client/session.py", line 274, in __init__
    'Tensor. (%s)' % (fetch, str(e)))
ValueError: Fetch argument <tf.Tensor 'batch:0' shape=(128, 56, 56, 3) dtype=float32> cannot be interpreted as a Tensor. (Tensor Tensor("batch:0", shape=(128, 56, 56, 3), dtype=float32) is not an element of this graph.)

Answer 1

You focused on the wrong part of the error message. The relevant part is

Tensor is not an element of this graph.

The problem is that you create a graph g in your function test, that is not the same one in which placeholders test_x and test_y provided as arguments have been created.

The easiest solution would be to create your graph g in main,

def main(argv=None):
    test_image, test_label = read_and_decode('val.tfrecords')
    with tf.Graph().as_default():
        test_images, test_labels = get_test_batch(test_image, test_label,
                                                  batch_size, crop_size)
        test(test_images, test_labels)