tensorflow中的tensor值的获取:
import tensorflow as tf #定义变量a a=tf.Variable([[[1,2,3],[4,5,6]],[[7,8,9],[10,11,12]]]) #定义索引 indics=[[0,0,0],[0,1,1],[0,1,2]] #把a中索引为indics的值取出 b=tf.gather_nd(a,indics) #初始化 init=tf.global_variables_initializer() with tf.Session() as sess: #执行初始化 sess.run(init) #打印结果 print(a.eval()) print(b.eval())