Pretrained weights for lucid.models.cspresnet_50_cls.
Wang et al. CVPRW 2020 CSP-ResNet-50 (21.6 M params, top-1 79.58%).
Attributes
RA_IN1KWeightEntryImageNet-1k RandAugment recipe checkpoint (top-1 79.58% at
256x256, per timm's
results-imagenet.csv for this exact tag).
The authors' README figure of 76.6 belongs to their own
darknet-trained model, which is a different checkpoint.
sourced from timm/cspresnet50.ra_in1k.DEFAULTWeightEntryAlias for
RA_IN1K.Notes
Reference: Wang, Liao, Wu, Chen, Hsieh, Yeh, "CSPNet: A New Backbone that can Enhance Learning Capability of CNN", CVPRW 2020 (arXiv:1911.11929).
Examples
>>> from lucid.models import cspresnet_50_cls
>>> model = cspresnet_50_cls(pretrained=True).eval()