composer.models.resnet.model#

composer.models.resnet.model

Classes

ComposerClassifier

Implements the base logic that all classifiers can build on top of.

ComposerResNet

ResNet model family extending ComposerClassifier.

Initializer

An enumeration.

Attributes

  • List

  • Optional

class composer.models.resnet.model.ComposerResNet(model_name, num_classes, pretrained=False, groups=1, width_per_group=64, initializers=None)[source]#

Bases: composer.models.base.ComposerClassifier

ResNet model family extending ComposerClassifier.

See this paper for details on the residual network architecture.

Parameters

  • model_name (str) – Name of the ResNet model instance. Either [“resnet18”, “resnet34”, “resnet50”, “resnet101”, “resnet152”].

  • num_classes (int) – Number of classes for the model.

  • pretrained (bool) – If true, use ImageNet pretrained weights. (default: False).

  • groups (int) – Number of filter groups for the 3x3 convolution layer in bottleneck blocks. (default: 1).

  • width_per_group (int) – Initial width for each convolution group. Width doubles after each stage. (default: 64).

  • initializers (List[Initializer], optional) – Initializers for the model. None for no initialization. (default: None).