2016-07-07 53 views
1

假设输入到网络的规模是一个placeholder具有可变批量大小,即:获取变量批次尺寸

x = tf.placeholder(..., shape=[None, ...]) 

是有可能得到的x形状已经馈送后? tf.shape(x)[0]仍然返回None

回答

8

如果x具有变量批量大小,获得实际形状的唯一方法是使用tf.shape()运算符。该运算符在tf.Tensor中返回一个符号值,因此它可以用作其他TensorFlow操作的输入,但要获取该形状的具体Python值,则需要将其传递到Session.run()

x = tf.placeholder(..., shape=[None, ...]) 
batch_size = tf.shape(x)[0] # Returns a scalar `tf.Tensor` 

print x.get_shape()[0] # ==> "?" 

# You can use `batch_size` as an argument to other operators. 
some_other_tensor = ... 
some_other_tensor_reshaped = tf.reshape(some_other_tensor, [batch_size, 32, 32]) 

# To get the value, however, you need to call `Session.run()`. 
sess = tf.Session() 
x_val = np.random.rand(37, 100, 100) 
batch_size_val = sess.run(batch_size, {x: x_val}) 
print x_val # ==> "37" 
1

使用x.get_shape().as_list()可以得到张量x的形状。要获得第一维(批量大小),您可以使用x.get_shape().as_list()[0]