classtorch.nn.RNN(*args, **kwargs)
input_size – The number of expected features in the input x
hidden_size – The number of features in the hidden state h
num_layers – Number of recurrent layers. E.g., setting num_layers=2 would mean stacking two RNNs together to form a stacked RNN, with the second RNN taking in outputs of the first RNN and computing the final results. Default: 1
nonlinearity – The non-linearity to use. Can be either ‘tanh’ or ‘relu’. Default: ‘tanh’
bias – If False, then the layer does not use bias weights b_ih and b_hh. Default: True
batch_first – If True, then the input and output tensors are provided as (batch, seq, feature)
dropout – If non-zero, introduces a Dropout layer on the outputs of each RNN layer except the last layer, with dropout probability equal to dropout. Default: 0
bidirectional – If True, becomes a bidirectional RNN. Default: False
首先，RNN这里的序列长度，是动态的，不写在参数里的，具体会由输入的input参数而定
RNN的对象接受的参数，input维度是(seq_len, batch_size, input_dim)，h0维度是(num_layers * directions, batch_size, hidden_dim)
其中，input的seq_len决定了序列的长度，h0是提供给每层RNN的初始输入，所有num_layers要和RNN的num_layers对得上
返回两个值，一个output，一个hn
hn的维度是(num_layers * directions, batch_size, hidden_dim)，是RNN的右侧输出，如果是双向的话，就还有一个左侧输出，其维度不受batch_first的控制
output的维度是(seq_len, batch_size, hidden_dim * directions)，是RNN的上侧输出
关于batch_first：
batch_first 设置之后，输入和输出都是batch_first了，不仅仅是输入要batch_first，输出也自动变为了batch_first
但是！hn的维度不变，还是 (num_layers * num_directions, batch, hidden_size)， batch还是在中间
关于num_layers ：
而num_layers并不是RNN的序列长度，而是堆叠层数，由上一层每个时间节点的输出作为下一层每个时间节点的输入
num_layers in RNN is just stacking RNNs on top of each other. So you get a hidden from each layer and an output only from the topmost layer.
如果numlayers=1, RNN图示如下：
如果numlayers=2, RNN图示如下：
LSTM可以参考下图：
原文地址：