Objects that Simplify the Creation of Neural Networks
Simplify the creation of neural networks using networks of filter, image, and state nodes.
- Metal Performance Shaders
Graphs in Metal Performance Shaders offer a higher level graph API, intended to simplify the creation of neural networks. The graph is a network of
MPSCNNKernel objects—each of the lower level
MPSCNNKernel subclasses has an associated object that is a subclass of the
Neural Network Graphs
An optimized representation of a graph of neural network image and filter nodes.
A placeholder node denoting the position of a neural network image in a graph.
The protocol that provides resource identification.
Convolution Layer Nodes
A representation of a convolution kernel with binary weights and an input image using binary approximations.
Fully Connected Layer Nodes
Normalization Layer Nodes
Kernel Concatenation Nodes
Loss Layer Nodes
A placeholder node denoting the per-element weight buffer used by loss and gradient loss kernels.
A protocol that defines methods that determine whether and when neural network training parameters are updated.
A texture that may have more than four channels for use in convolutional neural networks.
A texture for use in convolutional neural networks that stores transient data to be used and discarded promptly.