Expand description
Neural network layers and models using the array protocol.
This module provides neural network layers and models that work with any array type implementing the ArrayProtocol trait.
Structs§
- Batch
Norm - Batch normalization layer.
- Conv2D
- Convolutional layer.
- Conv2D
Builder - Builder for creating Conv2D layers
- Dropout
- Dropout layer.
- Linear
- Linear (dense/fully-connected) layer.
- MaxPool2D
- Max pooling layer.
- Multi
Head Attention - Multi-head attention layer.
- Sequential
- Sequential model that chains layers together.
Traits§
- Layer
- Trait for neural network layers.
Functions§
- create_
simple_ cnn - Example function to create a simple CNN model.