tf.nn.embedding_lookup用法解释

版权声明：本文为博主原创文章，未经博主允许不得转载。 https://blog.csdn.net/littlehaes/article/details/82825951

Welcome to my blog
tf.nn.embedding_lookup( params, ids, …),主要使用params, ids两个参数，函数的功能是从params中挑出索引为ids的元素，并返回一个张量，
假设params的shape是batch * hidden, ids的shape是batch * n
那么函数返回张量的shape是batch *n * hidden

import tensorflow as tf
w = tf.constant([[1,2],[3,4],[5,6]])
res = tf.nn.embedding_lookup(w, [0,1,0])
res2 = tf.nn.embedding_lookup(w, [[0],[1],[0]])
res4 = tf.nn.embedding_lookup(w, [[0,0,0,0],[1,1,1,1],[0,0,0,0]])
with tf.Session() as sess:
	res,res2,res4 = sess.run([res,res2,res4])
print res,res.shape
print res2,res2.shape
print res4,res4.shape

'''
打印结果
res: (3, 2)
[[1 2]
 [3 4]
 [1 2]] 

res2:(3, 1, 2)
[[[1 2]]

 [[3 4]]

 [[1 2]]]

res4: (3, 4, 2)
[[[1 2]
  [1 2]
  [1 2]
  [1 2]]

 [[3 4]
  [3 4]
  [3 4]
  [3 4]]

 [[1 2]
  [1 2]
  [1 2]
  [1 2]]] (3, 4, 2)
'''