Hi everyone,
Excuse me If you find my question very intuitive because I’m still new to Pytorch.
I have some confusion about torch.nn.LSTMCell Class.
According to the documentation, the inputs to the LSTM cell as follows: input and (h_0, c_0), where input is of shape (batch, input_size) which is a tensor containing input features.
The problem is that I can’t understand why in the provided example the given input has different dimensions, instead of being:
input(batch,input_size)
the given input is:
input(seq_len,input_size)
As long as I understand, the number 6 in the example represents the batch size.
Here is the example used in the documentation:
rnn = nn.LSTMCell(10, 20)
>>> input = torch.randn(6, 3, 10)
>>> hx = torch.randn(3, 20)
>>> cx = torch.randn(3, 20)
>>> output = []
>>> for i in range(6):
hx, cx = rnn(input[i], (hx, cx))
output.append(hx)