Pretrained weights for lucid.models.vgg_16_cls.
VGG-D 16-layer config (138.4 M params, 15.5 GFLOPs, top-1 71.59%).
Attributes
IMAGENET1K_V1WeightEntryImageNet-1k V1 checkpoint (top-1 71.592% / top-5 90.382%),
sourced from
torchvision/VGG16_Weights.IMAGENET1K_V1.DEFAULTWeightEntryAlias for
IMAGENET1K_V1.Notes
Reference: Simonyan, Zisserman, "Very Deep Convolutional Networks for Large-Scale Image Recognition", ICLR 2015 (arXiv:1409.1556).
Examples
>>> from lucid.models import vgg_16_cls
>>> model = vgg_16_cls(pretrained=True).eval()