tensorflow报错:Shape must be rank 2 but is rank 3 for 'MatMul' (op: 'MatMul')
来源:互联网 发布:淘宝客平台源码 编辑:程序博客网 时间:2024/05/19 12:13
tensorflow矩阵相乘,秩不同报错
在tensorflow中写了这样一句:
y_out = tf.matmul(outputs, W)
其中,outputs的shape为[16,336,400],W的shape为[400,1]
出现以下报错:
Shape must be rank 2 but is rank 3 for 'MatMul' (op: 'MatMul') with input shapes: [16,336,400], [400,1].
Numpy下同样的写法没有问题
import numpy as npA = np.array([[[1, 2, 3, 4], [5, 6, 7, 8], [9, 0, 1, 2]], [[4, 3, 2, 1], [8, 7, 6, 5], [2, 1, 0, 9]]])print(A)print(A.shape)print('---------------------------')B = np.array([[1], [2], [3], [4]])print(B)print(B.shape)print('---------------------------')C = np.matmul(A, B)print(C)print(C.shape)
输出结果:
[[[1 2 3 4] [5 6 7 8] [9 0 1 2]] [[4 3 2 1] [8 7 6 5] [2 1 0 9]]](2, 3, 4)---------------------------[[1] [2] [3] [4]](4, 1)---------------------------[[[30] [70] [20]] [[20] [60] [40]]](2, 3, 1)
解决办法
方案一
import numpy as npimport tensorflow as tfsess = tf.Session()A = np.array([[[1, 2, 3, 4], [5, 6, 7, 8], [9, 0, 1, 2]], [[4, 3, 2, 1], [8, 7, 6, 5], [2, 1, 0, 9]]])B = np.array([[1], [2], [3], [4]])A = tf.cast(tf.convert_to_tensor(A), tf.int32) # shape=[2, 3, 4]B = tf.cast(tf.convert_to_tensor(B), tf.int32) # shape=[4, 1]#-----------------------------------------修改部分(开始)-----------------------------------------#要想让A和B进行tf.matmul操作,第一个维数必须一致。因此要把B先tile后转成[2, 4, 1]维B_ = tf.tile(B, [2, 1])# B的第一维复制2倍,第二维复制1倍B = tf.reshape(B_, [2, 4, 1])# 或 更通用的改法:#B_ = tf.tile(B, [tf.shape(A)[0], 1])#B = tf.reshape(B_, [tf.shape(A)[0], tf.shape(B)[0], tf.shape(B)[1]])#-----------------------------------------修改部分(结束)-----------------------------------------#此时就可以matmul了C = tf.matmul(A, B)print('C:',C.get_shape().as_list())sess.run(C)
输出结果:
('C:', [2, 3, 1])array([[[30], [70], [20]], [[20], [60], [40]]], dtype=int32)
方案二
import numpy as npimport tensorflow as tfsess = tf.Session()A = np.array([[[1, 2, 3, 4], [5, 6, 7, 8], [9, 0, 1, 2]], [[4, 3, 2, 1], [8, 7, 6, 5], [2, 1, 0, 9]]])B = np.array([[1], [2], [3], [4]])A = tf.cast(tf.convert_to_tensor(A), tf.int32) # shape=[2, 3, 4]B = tf.cast(tf.convert_to_tensor(B), tf.int32) # shape=[4, 1]#-----------------------------------------修改部分(开始)-----------------------------------------#把A的前两个维度展为一个维度A = tf.reshape(A, [-1, 4])#此时就可以matmul了C = tf.matmul(A, B)# print('C:',C.get_shape().as_list()) #结果: [6, 1]#再把C的前两个维度还原C = tf.reshape(C, [2, 3, 1])#-----------------------------------------修改部分(结束)-----------------------------------------print('C:',C.get_shape().as_list())sess.run(C)
输出结果:
('C:', [2, 3, 1])
阅读全文
0 0
- tensorflow报错:Shape must be rank 2 but is rank 3 for 'MatMul' (op: 'MatMul')
- tensorflow报错:Shape must be rank 2 but is rank 3 for 'MatMul' (op: 'MatMul')
- ValueError: Shape must be rank 0 but is rank 1 for 'train_data/ReadFile' (op: 'ReadFile') with input
- matmul
- TensorFlow中的一个重要op---MatMul的实现(一)
- tf.matmul (API r1.3)
- TensorFlow入门(二)--Rank,Shape,Type
- tensorflow报错:for unknown op
- tf.matmul
- tf.matmul()
- maven的pom文件报错: must be "pom" but is "jar"
- TensorFlow中的一个重要ops---MatMul的实现(二)
- maven引用jar报错:spring-core:jar must be a valid version but is '${spring.version}
- Rank
- rank()
- rank
- RANK
- Rank
- 初使用Hbuider如何新建web项目
- editText软键盘,挤屏,关闭
- jdbc基础操作
- IOS dispatch_once_t
- Centos7安装、配置SVN服务,并设置开机自启动
- tensorflow报错:Shape must be rank 2 but is rank 3 for 'MatMul' (op: 'MatMul')
- Java中的数组
- ubuntu ssh 安装
- 深入理解操作系统原理之文件系统
- iOS中的Socket编程,Socket基础知识(一)
- Vue中引入ElementUI库
- nginx教程2:日志
- 企业服务总线ESB简介
- java生成Https证书,及证书导入的步骤和过程