Comment on page

Conv. Trans. Layer 2D

Conv. Trans. Layer 2D Node Documentation
The Conv. Trans. Layer 2D node in AugeLab Studio is used to add a 2D transposed convolutional layer to a convolutional neural network (CNN) model. It allows you to specify the filter size, kernel size, dilation size, and activation function for the layer.
Node Overview
The Conv. Trans. Layer 2D node adds a 2D transposed convolutional layer to a CNN model. It provides the following outputs:
Keras Layer: The 2D transposed convolutional layer configured with the specified parameters.
Node Properties
Node Title: Conv. Trans. Layer 2D
Node ID: OP_NODE_AI_2D_CONV_TRANS_LAYER
Inputs
The Conv. Trans. Layer 2D node has no input sockets.
Outputs
The Conv. Trans. Layer 2D node has one output socket:
1.Keras Layer: The 2D transposed convolutional layer configured with the specified parameters.
Node Interaction
1.Drag and drop the Conv. Trans. Layer 2D node from the node library onto the canvas in AugeLab Studio.
2.Set the filter size, kernel size, dilation size, and activation function in the node's properties.
3.The configured 2D transposed convolutional layer will be displayed in the node.
4.Connect the output socket of the Conv. Trans. Layer 2D node to other nodes in the pipeline to build a CNN model.
Implementation Details
The Conv. Trans. Layer 2D node creates a 2D transposed convolutional layer using the Keras API. The layer is configured with the specified filter size, kernel size, dilation size, and activation function.
The Conv. Trans. Layer 2D node extends the NodeCNN base class, which provides the common functionality for CNN-related nodes.
When the node is evaluated, the getKerasLayer method is called to create and configure the Keras layer based on the specified parameters. The getKerasLayer method is implemented in the node's subclasses to return the appropriate Keras layer type (Conv2D, SeparableConv2D, or Conv2DTranspose).
The node supports various activation functions, including relu, sigmoid, softmax, softplus, softsign, tanh, selu, elu, and exponential. The activation function can be selected from a dropdown menu in the node's properties.
The configured 2D transposed convolutional layer is outputted through the Keras Layer output socket, which can be connected to other nodes in the pipeline for further processing.
Usage
1.Drag and drop the Conv. Trans. Layer 2D node from the node library onto the canvas in AugeLab Studio.
2.Set the filter size, kernel size, dilation size, and activation function in the node's properties.
3.The configured 2D transposed convolutional layer will be displayed in the node.
4.Connect the output socket of the Conv. Trans. Layer 2D node to other nodes in the pipeline to build a CNN model.
Notes
The Conv. Trans. Layer 2D node is used to add a 2D transposed convolutional layer to a CNN model.
The Conv. Trans. Layer 2D node requires the Keras library to be installed.
Ensure that the specified kernel size is an odd number to maintain symmetry.
The Conv. Trans. Layer 2D node supports various activation functions for non-linear activation of the layer.
The configured 2D transposed convolutional layer can be connected to other nodes in the pipeline to build a CNN model.

Function Blocks - AI Model Elements - Previous

Conv. Sep. Layer 2D

Next - Function Blocks - AI Model Elements

Convolutional Layer 2D

Last modified 4mo ago