In [1]:

    

import tensorflow as tf
a = tf.truncated_normal([16, 128, 128, 3])
sess = tf.Session()
sess.run(tf.global_variables_initializer())
sess.run(tf.shape(a))


    

    
Out[1]:





array([ 16, 128, 128,   3], dtype=int32)

In [4]:

    

b = tf.reshape(a, [16, 49152])
sess.run(tf.shape(b))


    

    
Out[4]:





array([   16, 49152], dtype=int32)

In [ ]:

    




    

In [ ]: