tf.nn.static_bidirectional_rnn ; tf.contrib.rnn.static_bidirectional_rnn 讲解

最新推荐文章于 2020-02-08 23:10:12 发布

UESTC_20172222

最新推荐文章于 2020-02-08 23:10:12 发布

阅读量1.2k

点赞数 2

CC 4.0 BY-SA版权

分类专栏： tensorflow 学习文章标签： tf.contrib.rnn.static_bidirectional tf.nn.static_bidirectional_rnn tensorflow

本文链接：https://blog.youkuaiyun.com/qq_32806793/article/details/85240304

tensorflow 学习专栏收录该内容

30 篇文章

订阅专栏

本文深入解析了tf.nn.static_bidirectional_rnn函数，详细介绍了其参数设置和返回值含义，包括前向和反向RNNCell实例的使用，以及如何处理序列长度和初始状态。

摘要生成于 C知道，由 DeepSeek-R1 满血版支持，前往体验 >

tf.nn.static_bidirectional_rnn h 和tf.contrib.rnn.static_bidirectional_rnn是一样的

tf.nn.static_bidirectional_rnn

别名:
tf.contrib.rnn.static_bidirectional_rnn
tf.nn.static_bidirectional_rnn

tf.nn.static_bidirectional_rnn(
    cell_fw,
    cell_bw,
    inputs,
    initial_state_fw=None,
    initial_state_bw=None,
    dtype=None,
    sequence_length=None,
    scope=None
)
'''
Args:
	cell_fw: An instance of RNNCell, to be used for forward direction.
	cell_bw: An instance of RNNCell, to be used for backward direction.
	inputs: A length T list of inputs, each a tensor of shape [batch_size, input_size], or a nested tuple of such elements.
	initial_state_fw: (optional) An initial state for the forward RNN. This must be a tensor of appropriate type and shape [batch_size, cell_fw.state_size]. If cell_fw.state_size is a tuple, this should be a tuple of tensors having shapes [batch_size, s] for s in cell_fw.state_size.
	initial_state_bw: (optional) Same as for initial_state_fw, but using the corresponding properties of cell_bw.
	dtype: (optional) The data type for the initial state. Required if either of the initial states are not provided.
	sequence_length: (optional) An int32/int64 vector, size [batch_size], containing the actual lengths for each of the sequences.
	scope: VariableScope for the created subgraph; defaults to "bidirectional_rnn"
Returns:
	A tuple (outputs, output_state_fw, output_state_bw) where: outputs is a length T list of outputs (one for each input), which are depth-concatenated forward and backward outputs. output_state_fw is the final state of the forward rnn. output_state_bw is the final state of the backward rnn.
	'''

参数说明

cell_fw: An instance of RNNCell, 前向RNNCell 实例
cell_bw: An instance of RNNCell, 反向RNNCell实例
inputs: 输入，是一个长度为 T列表, 每个元素是一个 a tensor（shape [batch_size, input_size]）
initial_state_fw: （可选的)前向RNN的初始状态。这必须是类型和形状都满足的tensor([batch_size,cell_fw.state_size])。如果cell_fw.state_size是一个元组，那么它应该也是一个张量元组，元组的每个元素的shape为 [batch_size,s] for s in cell_fw.state_size
**initial_state_bw:**和initial_state_fw要求一样
sequence_length: 可选）int32/int64类型的向量大小为 [batch_size]，包含每个序列的实际长度。

返回：
一个元组：（outputs，output_state_fw，output_state_bw）
**outputs：**是长度T的list（每个时间步骤对应一个），这些大小为[batch_size,2*hidden_units]，这里不理解的请看下面这个图
**output_state_fw:**是正向rnn的最后state(c,h)
**output_state_bw:**是向后rnn的最后state(c,h)

双向RNN