KeyError：tf.Tensor&#39;Placeholder_6：0&#39;shape = <unknown> dtype =字符串

Question

Can you please explain the problem below? Here is my code fragment in my python notebook:

word2int = {}
int2word = {}

for i,word in enumerate(words):
    word2int[word] = i
    int2word[i] = word

def euclidean_dist(vec1, vec2):
    return np.sqrt(np.sum((vec1-vec2)**2))

def find_closest(word_index, vectors):
    min_dist = 10000 # to act like positive infinity
    min_index = -1
    query_vector = vectors[word_index]
    for index, vector in enumerate(vectors):
        if euclidean_dist(vector, query_vector) < min_dist and not np.array_equal(vector, query_vector):
            min_dist = euclidean_dist(vector, query_vector)
            min_index = index
    return min_index

Z = tf.placeholder(tf.string)
find_closest_word = int2word[find_closest(word2int[Z], vectors)]

# Create SignatureDef metadata for the model
classification_inputs = tf.saved_model.utils.build_tensor_info(Z)
classification_outputs_classes = tf.saved_model.utils.build_tensor_info(find_closest_word)

classification_signature = (
      tf.saved_model.signature_def_utils.build_signature_def(
          inputs={
              tf.saved_model.signature_constants.CLASSIFY_INPUTS:
                  classification_inputs
          },
          outputs={
              tf.saved_model.signature_constants.CLASSIFY_OUTPUT_CLASSES:
                  classification_outputs_classes
          },
          method_name=tf.saved_model.signature_constants.CLASSIFY_METHOD_NAME))

Here is the error message when I run the code fragment above:

---------------------------------------------------------------------------
KeyError                                  Traceback (most recent call last)
<ipython-input-65-016dad8c7403> in <module>()
     12     return min_index
     13 Z = tf.placeholder(tf.string)
---> 14 find_closest_word = int2word[find_closest(word2int[Z], vectors)]

KeyError: <tf.Tensor 'Placeholder_7:0' shape=<unknown> dtype=string>

Updated question:

How can I convert a string tensor Z to a python string so that it can be used as index in word2int ?

Answer 1

From your code, I guess you think Z is the word you pass as input to the network. It isn't, since you define it as Z = tf.placeholder(tf.string) . Z is thus a placeholder object that will eventually be filled with a string in your feed_dict when you run the graph in a tf.Session instance by calling its run() method.

Since your word2int dictionary is just a string-to-index dict, you get the KeyError when you try to use the placeholder as a key.