blockPolynomNetwork
Neural Network approximating y = u*u + 0.5*u - 2.0 on interval [-1,1]
Extends from NeuralNetwork.Networks.Interfaces.Network.
Information
Training
Neural network parameters trained in HelloWorld.ipynb.
Trained with TensorFlow on 8000 data points from interval [-1,1].
Parameters
| Type | Name | Default | Description |
|---|---|---|---|
| Real | layer_1_weights | {{-0.95248}, {-0.943175}} | |
| Real | layer_1_bias | {0.872633, -0.949252} | |
| Real | layer_2_weights | {{-2.25385, 1.40389}} | |
| Real | layer_2_bias | {0.60548} |
Connectors
| Type | Name | Default | Description |
|---|---|---|---|
| Modelica.Blocks.Interfaces.RealInput[:] | u (from Network) | Connector of Real input signals | |
| Modelica.Blocks.Interfaces.RealOutput[:] | y (from Network) | Connector of Real output signals |
Components
| Type | Name | Default | Description |
|---|---|---|---|
| Layer.Dense | inputLayer | ||
| Layer.Dense | outputLayer |