scope.py

import numpy as np
import tensorflow as tf
import tensorflow.contrib.slim as slim

def mapping(feat, is_training=True, reuse=False):
    batch_norm_params = {'is_training': is_training, 'decay': 0.9, 'updates_collections': None}
    with tf.variable_scope("mapping") as scope:
        if reuse:
            scope.reuse_variables()
        net = slim.fully_connected(feat, 16
                        , activation_fn       = tf.nn.sigmoid
                        , weights_initializer = tf.truncated_normal_initializer(stddev=0.01)
                        , normalizer_fn       = slim.batch_norm
                        , normalizer_params   = batch_norm_params
                        , scope='m_fc1')
        net = slim.dropout(net, keep_prob=0.7, is_training=is_training, scope='m_dr')  
        out = slim.fully_connected(net, 2
                        , activation_fn       = None
                        , normalizer_fn       = None
                        , scope='m_fc2')
        return out 
feat1    = tf.placeholder(tf.float32, [None, 6])
feat2    = tf.placeholder(tf.float32, [None, 6])
is_train = tf.placeholder(tf.bool)
map1   = mapping(feat1, is_train, False)
map2   = mapping(feat2, is_train, True) 

# INIT
init = tf.global_variables_initializer()
sess = tf.Session()
sess.run(init)

# CHECK
feat  =  np.random.rand(2, 6)
feeds = {feat1: feat, feat2: feat, is_train: True}
m1val = sess.run(map1, feed_dict=feeds)
m2val = sess.run(map2, feed_dict=feeds)

print (m1val)
print (m2val)

[[ 0.37606525 -0.41489923]
[-0.07807794 -0.81576711]]
[[-0.69362152 -0.15968615]
[-0.56324488 -0.42829311]]