Tensorflow之tf.metrics

import tensorflow as tf
import numpy as np

N_TRUE_P = 0
N_PRED_P = 0

def reset_running_variables():
    """ Resets the previous values of running variables to zero """
    global N_TRUE_P, N_PRED_P
    N_TRUE_P = 0
    N_PRED_P = 0

def update_running_variables(labs, preds):
    global N_TRUE_P, N_PRED_P
    N_TRUE_P += ((labs * preds) > 0).sum()
    N_PRED_P += (preds > 0).sum()

def calculate_precision():
    global N_TRUE_P, N_PRED_P
    return float (N_TRUE_P) / N_PRED_P

if __name__ == '__main__':

    labels = np.array([[1,1,1,0],
                       [1,1,1,0],
                       [1,1,1,0],
                       [1,1,1,0]], dtype=np.uint8)

    predictions = np.array([[1,0,0,0],
                            [1,1,0,0],
                            [1,1,1,0],
                            [0,1,1,1]], dtype=np.uint8)

    n_batches = len(labels)

    # #numpy
    # reset_running_variables()
    #
    # for i in range(n_batches):
    #     update_running_variables(labs=labels[i], preds=predictions[i])
    #
    # precision = calculate_precision()
    # print("[NP] SCORE: %1.4f" % precision)

    #tensorflow
    graph = tf.Graph()
    with graph.as_default():
        # Placeholders to take in batches onf data
        tf_label = tf.placeholder(dtype=tf.int32, shape=[None])
        tf_prediction = tf.placeholder(dtype=tf.int32, shape=[None])

        # Define the metric and update operations
        tf_metric, tf_metric_update = tf.metrics.precision(tf_label,
                                                           tf_prediction,
                                                           name="my_metric")

        # Isolate the variables stored behind the scenes by the metric operation
        running_vars = tf.get_collection(tf.GraphKeys.LOCAL_VARIABLES, scope="my_metric")

        # Define initializer to initialize/reset running variables
        running_vars_initializer = tf.variables_initializer(var_list=running_vars)

    with tf.Session(graph=graph) as session:
        session.run(tf.global_variables_initializer())

        # initialize/reset the running variables
        session.run(running_vars_initializer)

        for i in range(n_batches):
            # Update the running variables on new batch of samples
            feed_dict = {tf_label: labels[i], tf_prediction: predictions[i]}
            session.run(tf_metric_update, feed_dict=feed_dict)

        # Calculate the score
        score = session.run(tf_metric)
        print("[TF] SCORE: %1.4f" % score)

 参考：https://zhuanlan.zhihu.com/p/43359894