tf.sparse_to_dense()

tf.sparse_to_dense

import tensorflow as tf   

import numpy as np

SIZE=6

CLASS=10

label=np.random.randint(0,10,size=SIZE) 

label=np.reshape(label,[SIZE,1])

index = np.reshape(np.arange(SIZE), [SIZE, 1])

#use a matrix  

concated = tf.concat([index, label], 1)  

onehot_labels = tf.sparse_to_dense(concated, [SIZE, CLASS], 1.0, 0.0)  

#use a vector  

concated2=tf.constant([1,3,4])  

onehot_labels2 = tf.sparse_to_dense(concated2, [ CLASS], 1.0, 0.0)

#use a scalar  

concated3=tf.constant(5)  

onehot_labels3 = tf.sparse_to_dense(concated3, [ CLASS], 1.0, 0.0)  

with tf.Session() as sess:  

    sess.run(tf.global_variables_initializer())

    result1=sess.run(onehot_labels)  

    result2 = sess.run(onehot_labels2)  

    result3 = sess.run(onehot_labels3)  

    print ("This is result1:")  

    print (result1)  

    print ("This is result2:")  

    print (result2)  

    print ("This is result3:")  

    print (result3)