Tensorflow error: "The tensor must be from the same graph as the tensor ..."

Question

I am trying to train a simple binary logistic regression classifier using Tensorflow (version 0.9.0) very similar to the initial tutorial and the following error occurs when installing the model:

ValueError: Tensor("centered_bias_weight:0", shape=(1,), dtype=float32_ref) must be from the same graph as Tensor("linear_14/BiasAdd:0", shape=(?, 1), dtype=float32). 

Here is my code:

 import tempfile import tensorflow as tf import pandas as pd # Customized training data parsing train_data = read_train_data() feature_names = get_feature_names(train_data) labels = get_labels(train_data) # Construct dataframe from training data features x_train = pd.DataFrame(train_data , columns=feature_names) x_train["label"] = labels y_train = tf.constant(labels) # Create SparseColumn for each feature (assume all feature values are integers and either 0 or 1) feature_cols = [ tf.contrib.layers.sparse_column_with_integerized_feature(f,2) for f in feature_names ] # Create SparseTensor for each feature based on data categorical_cols = { f: tf.SparseTensor(indices=[[i,0] for i in range(x_train[f].size)], values=x_train[f].values, shape=[x_train[f].size,1]) for f in feature_names } # Initialize logistic regression model model_dir = tempfile.mkdtemp() model = tf.contrib.learn.LinearClassifier(feature_columns=feature_cols, model_dir=model_dir) def eval_input_fun(): return categorical_cols, y_train # Fit the model - similarly to the tutorial model.fit(input_fn=eval_input_fun, steps=200) 

I feel that I am missing something critical ... maybe something that was supposed to be in the textbook but not explicitly mentioned?

In addition, every time I call fit (), I get the following warning:

 WARNING:tensorflow:create_partitioned_variables is deprecated. Use tf.get_variable with a partitioner set, or tf.get_partitioned_variable_list, instead. 

Answer 1

When you execute model.fit , LinearClassifier creates a separate tf.Graph based on the Ops contained in your eval_input_fun function. But during the creation of this graph, LinearClassifier does not have access to the categorical_cols and y_train that you saved globally.

Solution: move all Ops definitions (and their dependencies) inside eval_input_fun