LSTM unit outputs

Course Q&A Deep Learning Specialization Sequence Models
1

in the course lab its given that

next_hidden_state, _ , next_cell_state =  post_activation_LSTM_cell(inputs=..., initial_state=[prev_hidden_state, prev_cell_state])

but documentation says

output, next_hidden_state, next_cell state = post_activation_LSTM_cell(inputs=..., initial_state=[prev_hidden_state, prev_cell_state])

even the shapes of the tensors seem to match what the tensorflow documentation says, I’m really confused.

2

Thanks for bringing this up.
Your understanding of lstm __call__ api is correct. The staff have been notified about this.

3

Hi @lightfury9,

Can you share the documentation link with me where you say that ?

Thanks,
Mubsi

4

Here’s an example from the documentation: