tf.shape(a)和a.get_shape()

import tensorflow as tf

a = tf.constant([[1,2,3],[4,5,6]])

print(type(a.get_shape()))
print(type(a.shape))
print(type(tf.shape(a)))

得到：

<class 'tensorflow.python.framework.tensor_shape.TensorShape'>
<class 'tensorflow.python.framework.tensor_shape.TensorShape'>
<class 'tensorflow.python.framework.ops.Tensor'>

tf.shape(a)是返回一个Tensor

a.get_shape()和a.shape返回的是一个TensorShape

接着：

print(type(a.get_shape().as_list()))

得到的Python自带的list

<class 'list'>

同理于：tf.rank(a)和a.get_shape().ndims

#