tensor is not element of this graph - tensorflow

Question

i get Tensor Tensor("Placeholder:0", shape=(128, 784), dtype=float32) is not an element of this graph. when i want to train & test my graph. The stack-trace is not really helpful, I can't get more information out of it. I really don't understand how this error happens, every data set should be automatically added to graphFully through with graphFully.as_default():and I later call it with with tf.Session(graph=graph) as session:.
advice on how to simplify the graph would be welcome. I want to define several graphs and compare them, therefore the "complicated" structure.
my graph:
##fully connected with hidden layer
def createFullyConnected():
graphFully = tf.Graph()
with graphFully.as_default():
def constructGraph(dataset, weights1, biases1, weights2, biases2):
logits1 = tf.matmul(dataset, weights1) + biases1
hiddenl = tf.nn.relu(logits1)
logits2 = tf.matmul(hiddenl, weights2) + biases2
return logits2
def weight_variable(shape):
initial = tf.truncated_normal(shape, stddev=0.01)
return tf.Variable(initial, name='weights')
def bias_variable(shape):
initial = tf.constant(0.0, shape=shape)
return tf.Variable(initial, name='biases')
# Input data. For the training data, we use a placeholder that will be fed
# at run time with a training minibatch.
tf_train_dataset = tf.placeholder(tf.float32,
shape=(batch_size, image_size * image_size), name='train_data')
tf_train_labels = tf.placeholder(tf.float32, shape=(batch_size, num_labels), name='train_labels')
tf_valid_dataset = tf.constant(validation[0], name='valid_labels')
tf_test_dataset = tf.constant(test[0], name='test_labels')
# Variables.
with tf.name_scope('hidden') as scope:
weights1 = weight_variable([image_size * image_size, 1024])
biases1 = bias_variable([1024])
weights2 = weight_variable([1024, num_labels])
biases2 = bias_variable([num_labels])
# Training computation.
logits = constructGraph(tf_train_dataset, weights1, biases1, weights2, biases2)
loss = tf.reduce_mean(
tf.nn.softmax_cross_entropy_with_logits(labels=tf_train_labels, logits=logits))
# Optimizer.
optimizer = tf.train.GradientDescentOptimizer(0.5).minimize(loss)
# Predictions for the training, validation, and test data.
train_prediction = tf.nn.softmax(logits)
valid_prediction = tf.nn.softmax(constructGraph(tf_valid_dataset, weights1, biases1, weights2, biases2))
test_prediction = tf.nn.softmax(constructGraph(tf_test_dataset, weights1, biases1, weights2, biases2))
# We write the graph out to the `logs` directory
tf.summary.FileWriter("logs", graphFully).close()
return (graphFully, optimizer, train_prediction, valid_prediction, test_prediction)
and the evaluation:
def evaluate(graph, optimizer, train_prediction, valid_prediction, test_prediction):
num_steps = 3001
train_dataset = train[0]
train_labels = train[1]
valid_labels = validation[1]
test_labels = test[1]
outlier_labels = outlier[1]
with tf.Session(graph=graph) as session:
tf.global_variables_initializer().run()
print("Initialized")
for step in range(num_steps):
# Pick an offset within the training data, which has been randomized.
# Note: we could use better randomization across epochs.
offset = (step * batch_size) % (train_labels.shape[0] - batch_size)
# Generate a minibatch.
batch_data = train_dataset[offset:(offset + batch_size), :]
batch_labels = train_labels[offset:(offset + batch_size), :]
# Prepare a dictionary telling the session where to feed the minibatch.
# The key of the dictionary is the placeholder node of the graph to be fed,
# and the value is the numpy array to feed to it.
feed_dict = {tf_train_dataset : batch_data, tf_train_labels : batch_labels}
_, l, predictions = session.run(
[optimizer, loss, train_prediction], feed_dict=feed_dict)
if (step % 500 == 0):
print("Minibatch loss at step %d: %f" % (step, l))
print("Minibatch accuracy: %.1f%%" % accuracy(predictions, batch_labels))
print("Validation accuracy: %.1f%%" % accuracy(
valid_prediction.eval(), valid_labels))
print("Test accuracy: %.1f%%" % accuracy(test_prediction.eval(), test_labels))

Answer 1

Here the errors occur due to out of scope of the two placeholder tf_train_dataset and tf_train_labels. You need to access these two tensors from th graph inside evaluate function.
def evaluate(...):
...
tf_train_dataset = graph.get_tensor_by_name('train_data:0')
tf_train_labels = graph.get_tensor_by_name('train_labels:0')
with tf.Session(graph=graph) as session:
...