🖥
🖥
🖥
🖥
AugeLab Studio Manual
English
Ask or search…
K
Comment on page

Fully Connected

Fully Connected Node Documentation

The Fully Connected node in AugeLab Studio is used to add a fully connected layer to a neural network. Fully connected layers, also known as dense layers, are responsible for learning and combining the features extracted by previous layers to make predictions.

Node Overview

The Fully Connected node adds a fully connected layer to the network. It has the following properties:
  • Node Title: Fully Connected
  • Node ID: OP_NODE_AI_FULLY_CONNECTED

Inputs

The Fully Connected node does not have any input sockets. It represents a layer in the network architecture.

Outputs

The Fully Connected node does not have any output sockets. It is a middle layer that modifies the network architecture.

Node Properties

The Fully Connected node has the following configurable properties:
  • Output Size: The number of output units/neurons in the fully connected layer.
  • Activation Function: The activation function applied to the output of the fully connected layer.

Node Interaction

  1. 1.
    Drag and drop the Fully Connected node from the node library onto the canvas in AugeLab Studio.
  2. 2.
    Configure the properties of the Fully Connected node by selecting the desired output size and activation function.
  3. 3.
    Connect the Fully Connected node to other nodes in the network to modify the network architecture.

Implementation Details

The Fully Connected node is implemented using the Dense layer from the Keras library. The node's getKerasLayer method returns a Keras dense layer.
The dense layer is a fully connected layer in which each neuron is connected to every neuron in the previous layer. It learns and combines the features extracted by the previous layers to make predictions. The output size of the dense layer determines the number of neurons in the layer.
The activation function is applied to the output of the dense layer, introducing non-linearity to the network. Common activation functions include relu, sigmoid, softmax, tanh, and more.

Usage

  1. 1.
    Drag and drop the Fully Connected node from the node library onto the canvas in AugeLab Studio.
  2. 2.
    Configure the properties of the Fully Connected node by selecting the desired output size and activation function.
  3. 3.
    Connect the Fully Connected node to other nodes in the network to modify the network architecture.
  4. 4.
    Train the network using the modified architecture to observe the effects of the fully connected layer on model performance.

Notes

  • The Fully Connected node adds a fully connected layer to a neural network.
  • The fully connected layer learns and combines the features extracted by previous layers to make predictions.
  • The Fully Connected node does not have any input or output sockets. It is a middle layer that modifies the network architecture.
  • Ensure that the Fully Connected node is connected to other nodes in the network to have an impact on the model's behavior.
  • The Fully Connected node requires the Keras library to be installed.