VGG-11 with BatchNorm — feature-extracting backbone.
Builds a VGG with the VGG-11 topology
arch=(1, 1, 2, 2, 2) and a lucid.nn.BatchNorm2d layer
after each Conv + ReLU pair. BatchNorm was added in the
timm / reference-framework reimplementations of VGG (not in the
original paper) and substantially improves convergence speed and
final accuracy; the topology is otherwise identical to
vgg_11. Approximately 132.9 M parameters.
Model Size
Parameters
pretrainedbool= FalseReserved for future pretrained-weight loading. Currently
ignored.
**overridesobject= {}Keyword overrides forwarded into
VGGConfig.Returns
VGGBackbone with the VGG-11-BN configuration applied (or with
overrides merged on top of it).
Examples
>>> import lucid
>>> from lucid.models.vision.vgg import vgg_11_bn
>>> model = vgg_11_bn()
>>> x = lucid.randn(1, 3, 224, 224)
>>> out = model(x)
>>> out.last_hidden_state.shape
(1, 512, 7, 7)