Model Class Reference

Class for the ResNet generator model. More...

Inheritance diagram for Model:

Collaboration diagram for Model:

Public Member Functions
	__init__ (self)
	Constructor to create a new model instance.
	forward (self, z)
	Function to perform a forward pass.

Public Attributes
	fc = nn.Linear(96, 49152)
	Fully-connected layer.
	blocks
	Sequence of residual block layers.
	norm = nn.BatchNorm2d(16)
	Batch normalization layer.
	conv = nn.Conv2d(16, 1, 3, 1, 1)
	Convolutional layer.

Detailed Description

Class for the ResNet generator model.

ResNet generator model.

Definition at line 86 of file resnet.py.

Constructor & Destructor Documentation

__init__()

__init__

(

self

)

Constructor to create a new model instance.

Definition at line 91 of file resnet.py.

    def __init__(self):
        super().__init__()
        # fully-connected inputs
        self.fc = nn.Linear(96, 49152)
        # stack of residual blocks
        self.blocks = nn.ModuleList(
            [
                # (256, 8, 24)
                ResidualBlock(256, 256),
                # (256, 16, 48)
                ResidualBlock(256, 128),
                # (128, 32, 96)
                ResidualBlock(128, 64),
                # (64, 64, 192)
                ResidualBlock(64, 32),
                # (32, 128, 384)
                ResidualBlock(32, 16),
                # (16, 256, 768)
            ]
        )
        # norm and conv outputs
        self.norm = nn.BatchNorm2d(16)
        self.conv = nn.Conv2d(16, 1, 3, 1, 1)
 

Member Function Documentation

forward()

forward	(		self,
			z )

Function to perform a forward pass.

Compute the model output for a given input.

Definition at line 117 of file resnet.py.

    def forward(self, z):
        """Compute the model output for a given input."""
        z = self.fc(z)
        z = z.view(-1, 256, 8, 24)
        for block in self.blocks:
            z = block(z)
        z = self.norm(z)
        z.relu_()
        z = self.conv(z)
        return z.tanh_()
 

Member Data Documentation

blocks

blocks

Initial value:

=  nn.ModuleList(
            [
                # (256, 8, 24)
                ResidualBlock(256, 256),
                # (256, 16, 48)
                ResidualBlock(256, 128),
                # (128, 32, 96)
                ResidualBlock(128, 64),
                # (64, 64, 192)
                ResidualBlock(64, 32),
                # (32, 128, 384)
                ResidualBlock(32, 16),
                # (16, 256, 768)
            ]
        )

Sequence of residual block layers.

Definition at line 96 of file resnet.py.

conv

conv = nn.Conv2d(16, 1, 3, 1, 1)

Convolutional layer.

Definition at line 113 of file resnet.py.

fc

fc = nn.Linear(96, 49152)

Fully-connected layer.

Definition at line 94 of file resnet.py.

norm

norm = nn.BatchNorm2d(16)

Batch normalization layer.

Definition at line 112 of file resnet.py.

---

The documentation for this class was generated from the following file:

• pxd/scripts/pxd/background_generator/models/resnet.py