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

Compile Model

Compile Model Node Documentation

The Compile Model node in AugeLab Studio is used to compile a model by specifying the optimizer, loss function, metrics, and training parameters. It takes the output of a layer node as input and compiles the model with the specified configurations.

Node Overview

The Compile Model node allows you to compile a model by specifying the optimizer, loss function, metrics, and training parameters. It provides the following outputs:
  • PNNModel: The compiled model configuration object that contains the information about the compiled model and training parameters.

Node Properties

  • Node Title: Compile Model
  • Node ID: OP_NODE_AI_2D_COMPILE

Inputs

The Compile Model node has the following input sockets:
  1. 1.
    Layer: The output of a layer node.
  2. 2.
    Optimizer: The optimizer to use for training the model.
  3. 3.
    Loss Function: The loss function to use for training the model.
  4. 4.
    Metrics: The metrics to evaluate the model's performance during training and evaluation.
  5. 5.
    Training Params: The parameters for training the model.

Outputs

The Compile Model node has one output socket:
  1. 1.
    PNNModel: The compiled model configuration object that contains the information about the compiled model and training parameters.

Node Interaction

  1. 1.
    Drag and drop the Compile Model node from the node library onto the canvas in AugeLab Studio.
  2. 2.
    Connect the output of a layer node to the "Layer" input socket of the Compile Model node.
  3. 3.
    Connect the desired optimizer, loss function, metrics, and training parameters to the corresponding input sockets of the Compile Model node.
  4. 4.
    Click on the "Compile Model" button to compile the model with the specified configurations.
  5. 5.
    The compiled model will be displayed in the node.
  6. 6.
    Run the pipeline.
  7. 7.
    The output socket of the Compile Model node will provide the compiled model configuration object (PNNModel) containing the information about the compiled model and training parameters.

Implementation Details

The Compile Model node takes the output of a layer node as input and compiles the model with the specified optimizer, loss function, metrics, and training parameters. It uses the TensorFlow 2 backend for model compilation.
When the "Compile Model" button is clicked, the node retrieves the input configurations from the connected nodes and compiles the model using the specified optimizer, loss function, metrics, and training parameters. The compiled model is stored in a configuration object (PNNModel) that contains the information about the compiled model and training parameters.
The compiled model can be used for training or evaluation by connecting it to other nodes in the pipeline.

Usage

  1. 1.
    Drag and drop the Compile Model node from the node library onto the canvas in AugeLab Studio.
  2. 2.
    Connect the output of a layer node to the "Layer" input socket of the Compile Model node.
  3. 3.
    Connect the desired optimizer, loss function, metrics, and training parameters to the corresponding input sockets of the Compile Model node.
  4. 4.
    Click on the "Compile Model" button to compile the model with the specified configurations.
  5. 5.
    The compiled model will be displayed in the node.
  6. 6.
    Run the pipeline.
  7. 7.
    The output socket of the Compile Model node will provide the compiled model configuration object (PNNModel) containing the information about the compiled model and training parameters.

Notes

  • The Compile Model node requires the output of a layer node as input.
  • Ensure that the connected nodes provide the appropriate configurations for the optimizer, loss function, metrics, and training parameters.
  • The Compile Model node uses the TensorFlow 2 backend for model compilation.
  • The compiled model configuration object (PNNModel) can be used as input to other nodes in the pipeline for training or evaluation of the compiled model.