Models

Logisitic Regression

class distil.utils.models.logreg_net.LogisticRegNet(input_dim, num_classes)[source]

Bases: torch.nn.modules.module.Module

forward(x, last=False)[source]

Defines the computation performed at every call.

Should be overridden by all subclasses.

Note

Although the recipe for forward pass needs to be defined within this function, one should call the Module instance afterwards instead of this since the former takes care of running the registered hooks while the latter silently ignores them.

get_embedding_dim()[source]

Simple Neural Network

class distil.utils.models.simpleNN_net.ThreeLayerNet(input_dim, num_classes, h1, h2)[source]

Bases: torch.nn.modules.module.Module

forward(x, last=False)[source]

Defines the computation performed at every call.

Should be overridden by all subclasses.

Note

Although the recipe for forward pass needs to be defined within this function, one should call the Module instance afterwards instead of this since the former takes care of running the registered hooks while the latter silently ignores them.

get_embedding_dim()[source]
class distil.utils.models.simpleNN_net.TwoLayerNet(input_dim, num_classes, hidden_units)[source]

Bases: torch.nn.modules.module.Module

forward(x, last=False)[source]

Defines the computation performed at every call.

Should be overridden by all subclasses.

Note

Although the recipe for forward pass needs to be defined within this function, one should call the Module instance afterwards instead of this since the former takes care of running the registered hooks while the latter silently ignores them.

get_embedding_dim()[source]

MNIST_NET

class distil.utils.models.mnist_net.MnistNet[source]

Bases: torch.nn.modules.module.Module

forward(x, last=False)[source]

Defines the computation performed at every call.

Should be overridden by all subclasses.

Note

Although the recipe for forward pass needs to be defined within this function, one should call the Module instance afterwards instead of this since the former takes care of running the registered hooks while the latter silently ignores them.

get_embedding_dim()[source]

CIFAR10Net module

class distil.utils.models.cifar10net.CifarNet[source]

Bases: torch.nn.modules.module.Module

forward(x, last=False)[source]

Defines the computation performed at every call.

Should be overridden by all subclasses.

Note

Although the recipe for forward pass needs to be defined within this function, one should call the Module instance afterwards instead of this since the former takes care of running the registered hooks while the latter silently ignores them.

get_embedding_dim()[source]

ResNet

ResNet in PyTorch.

For Pre-activation ResNet, see ‘preact_resnet.py’.

Reference: [1] Kaiming He, Xiangyu Zhang, Shaoqing Ren, Jian Sun

Deep Residual Learning for Image Recognition. arXiv:1512.03385

class distil.utils.models.resnet.BasicBlock(in_planes, planes, stride=1)[source]

Bases: torch.nn.modules.module.Module

expansion = 1
forward(x)[source]

Defines the computation performed at every call.

Should be overridden by all subclasses.

Note

Although the recipe for forward pass needs to be defined within this function, one should call the Module instance afterwards instead of this since the former takes care of running the registered hooks while the latter silently ignores them.

class distil.utils.models.resnet.Bottleneck(in_planes, planes, stride=1)[source]

Bases: torch.nn.modules.module.Module

expansion = 4
forward(x)[source]

Defines the computation performed at every call.

Should be overridden by all subclasses.

Note

Although the recipe for forward pass needs to be defined within this function, one should call the Module instance afterwards instead of this since the former takes care of running the registered hooks while the latter silently ignores them.

class distil.utils.models.resnet.ResNet(block, num_blocks, num_classes=10)[source]

Bases: torch.nn.modules.module.Module

forward(x, last=False)[source]

Defines the computation performed at every call.

Should be overridden by all subclasses.

Note

Although the recipe for forward pass needs to be defined within this function, one should call the Module instance afterwards instead of this since the former takes care of running the registered hooks while the latter silently ignores them.

get_embedding_dim()[source]
distil.utils.models.resnet.ResNet101(num_classes=10)[source]
distil.utils.models.resnet.ResNet152(num_classes=10)[source]
distil.utils.models.resnet.ResNet18(num_classes=10)[source]
distil.utils.models.resnet.ResNet34(num_classes=10)[source]
distil.utils.models.resnet.ResNet50(num_classes=10)[source]