ResNet-101 feature-extracting backbone (no classification head).
Builds a ResNet with the paper-cited ResNet-101 topology:
_Bottleneck blocks stacked [3, 4, 23, 3] over four
stages, yielding approximately 44.5M parameters. Reaches 77.37%
ImageNet-1k top-1 in He et al., 2015 (Table 4). Higher-capacity
drop-in replacement for ResNet-50 when accuracy matters more than
inference latency.
Model Size
Parameters
pretrainedbool= FalseReserved for future pretrained-weight loading. Currently
ignored.
**overridesobject= {}Keyword overrides forwarded into
ResNetConfig to
customise individual fields without writing a config by hand.Returns
ResNetBackbone with the ResNet-101 configuration applied (or with
overrides merged on top of it).
Notes
See He et al., "Deep Residual Learning for Image Recognition", CVPR 2016 (arXiv:1512.03385), Table 1. The bulk of the depth lives in stage-3 (23 bottleneck blocks); stages 1, 2, and 4 are identical to ResNet-50. Final-stage output is 2048 channels.
Examples
>>> import lucid
>>> from lucid.models.vision.resnet import resnet_101
>>> model = resnet_101()
>>> x = lucid.randn(1, 3, 224, 224)
>>> out = model(x)
>>> out.last_hidden_state.shape # (B, 2048, H/32, W/32)
(1, 2048, 7, 7)