resnext-101-32x4d

ResNeXt-101 (32x4d)

Xie et al., 2017 — Aggregated Residual Transformations for Deep Neural Networks (arXiv:1611.05431)

Lucid port of timm/resnext101_32x4d.gluon_in1k, converted to Lucid-native safetensors.

Available weights

Usage

import lucid.models as models
from lucid.models.weights import ResNeXt101_32x4dWeights

# default tag
model = models.resnext_101_32x4d_cls(pretrained=True)

# explicit tag (enum or string)
model = models.resnext_101_32x4d_cls(weights=ResNeXt101_32x4dWeights.GLUON_IN1K)
model = models.resnext_101_32x4d_cls(pretrained="GLUON_IN1K")

# preprocessing travels with the weights
weights = ResNeXt101_32x4dWeights.GLUON_IN1K
preprocess = weights.transforms()
logits = model(preprocess(image)[None]).logits

Conversion

Converted from timm/resnext101_32x4d.gluon_in1k via python -m tools.convert_weights resnext_101_32x4d --tag GLUON_IN1K. Key mapping + numerical parity verified against the source.

License

apache-2.0 — inherited from the original weights.

Citation

@inproceedings{xie2017aggregated,
  title={Aggregated Residual Transformations for Deep Neural Networks},
  author={Xie, Saining and Girshick, Ross and Doll{\'a}r, Piotr and Tu, Zhuowen and He, Kaiming},
  booktitle={CVPR}, year={2017}
}