What does this diagram mean in terms of number of neurons of a feedforward neural network?
Can someone re-draw it like a feedforward neural network?
Hey there
RNN cells can be shown in different time steps as below:
Or maybe this can illustrate the whole RNN structure better:
The RNN used in the above image is composed of 3 LSTM layers, consisting of 24 cells (green blocks) each. The input layer (pink block) of the network comprises the 6 observed-state variables. The black arrows between different layers indicate all-to-all connections. The purple arrows indicate recurrent connections within each LSTM layer. The last layer consists of 5 output neurons (orange) with linear activation.
Feel free to ask if you need further assistance!
2 Likes