CODEWITHSUNDEEP
machine-learningkerasneural-networkdata-science
new_one
new_one

Reputation: 105

Is it possible to have multiple inputs in a neural network?

I have 2 different inputs where the second one is the output of the first (visible in the image- note that I have not included all of the lines between the network).
I'm trying to build a network where between the first input and the second there is one or more dense layers (fully connected layers), and then between the second input and the output there is again, one or more dense layers. Is this possible?

The network I'm trying to build: The network I'm trying to build

I have started by defining my network like this:

model = Sequential()
model.add(Dense(num_neurons, input_dim=input, activation='relu'))
model.add(Dense(output, activation='softmax'))

model = Sequential()
model.add(Dense(num_neurons, input_dim=input, activation='relu'))
model.add(Dense(output, activation='softmax'))

Since the inputs are a vector I didn't know if maybe using RNN/LSTM is even possible.

Should I go for a different network design?

Upvotes: 0

Views: 2089

Answers (1)

new_one
new_one

Reputation: 105

I created the network using Keras Functional API! It appeared like it allow for much more diversification in the network, allowing many inputs and outputs.

Upvotes: 1

Related Questions