Tensorflow使用变量出现错误： List of Tensors when single Tensor expected

1.背景：

import tensorflow as tf
a = tf.constant(tf.random_normal([2, 2]))    # 运行该代码出现错误
print(a)

将tf.random_normal传入给tf.constant发生错误:
TypeError: List of Tensors when single Tensor expected

2. 问题出现原因：

查看tf.constant()和tf.random_normal()是如何使用的

2.1 tf.random_normal函数定义:

def random_normal(shape,
                  mean=0.0,
                  stddev=1.0,
                  dtype=dtypes.float32,
                  seed=None,
                  name=None):

• Returns: A tensor of the specified shape filled with random normal values.

2.2 tf.constant函数定义:

def constant(value, dtype=None, shape=None, name="Const", verify_shape=False)

• value: A constant value (or list) of output type dtype.
• Returns: A Constant Tensor.

2.3 问题的原因：

结论: 因为tf.random_normal返回值是一个Tensor,但是tf.constat传入的形参是list二者类型是不匹配的,所以出现错误。

3. 如何解决：

3.1 方法1：

Use NumPy to generate the random value and put it in a tf.constant()

some_test = tf.constant(
    np.random.normal(loc=0.0, scale=1.0, size=(2, 2)).astype(np.float32))

3.2 方法2：

Potentially faster, as it can use the GPU to generate the random numbers,Use TensorFlow to generate the random value and put it in a tf.Variable

some_test = tf.Variable(
    tf.random_normal([2, 2], mean=0.0, stddev=1.0, dtype=tf.float32)
sess.run(some_test.initializer)

查看tf.Variable函数定义:

def __init__(self,
               initial_value=None,
               trainable=True,
               collections=None,
               validate_shape=True,
               caching_device=None,
               name=None,
               variable_def=None,
               dtype=None,
               expected_shape=None,
               import_scope=None,
               constraint=None):
               ...
               )

• args: initial_value: A Tensor, or Python object convertible to a Tensor,which is the initial value for the Variable.

下次遇到这种函数参数的问题，一定要查看这个函数是如何使用的，形参和返回值都是些什么。

参考文章：

1. TensorFlow: generating a random constant
2. ipython笔记