efficientformer_l3(pretrained: bool = False, overrides: object = {})EfficientFormer-L3 backbone (Li et al., 2022).
Builds the canonical EfficientFormer-L3 configuration:
depths=(4, 4, 12, 6), embed_dims=(64, 128, 320, 512),
drop_path_rate=0.1. Approximately 30.9M parameters.
Model Size
Parameters
pretrainedbool= FalseIf
True, loads ImageNet-1k pretrained weights when
available. Defaults to False.**overridesobject= {}Keyword overrides on top of the canonical L3 config.
Returns
EfficientFormerAn EfficientFormer backbone returning a flat
feature.
Notes
EfficientFormer-L3 reaches 82.4% top-1 on ImageNet-1k (Li et al., 2022, Table 4).
Examples
>>> import lucid
>>> from lucid.models.vision.efficientformer import efficientformer_l3
>>> model = efficientformer_l3()
>>> x = lucid.randn(1, 3, 224, 224)
>>> model.forward_features(x).shape
(1, 512)