lucid-dl/wide-resnet-101-2

Wide ResNet-101-2

Zagoruyko & Komodakis, 2016 — Wide Residual Networks (arXiv:1605.07146)

Lucid port of torchvision/Wide_ResNet101_2_Weights.IMAGENET1K_V1, converted to Lucid-native safetensors.

Available weights

Tag	acc@1	acc@5	Params	GFLOPs	Size	Source
IMAGENET1K_V1 (default)	78.848	94.284	126.9M	22.753	484.62 MB	torchvision

Usage

import lucid.models as models
from lucid.models.weights import WideResNet101Weights

# default tag
model = models.wide_resnet_101_cls(pretrained=True)

# explicit tag (enum or string)
model = models.wide_resnet_101_cls(weights=WideResNet101Weights.IMAGENET1K_V1)
model = models.wide_resnet_101_cls(pretrained="IMAGENET1K_V1")

# preprocessing travels with the weights
weights = WideResNet101Weights.IMAGENET1K_V1
preprocess = weights.transforms()
logits = model(preprocess(image)[None]).logits

Conversion

Converted from torchvision/Wide_ResNet101_2_Weights.IMAGENET1K_V1 via python -m tools.convert_weights wide_resnet_101 --tag IMAGENET1K_V1. Key mapping + numerical parity verified against the source.

License

bsd-3-clause — inherited from the original weights.

Citation

@inproceedings{zagoruyko2016wide,
  title={Wide Residual Networks},
  author={Zagoruyko, Sergey and Komodakis, Nikos},
  booktitle={BMVC}, year={2016}
}