mask2former_swin_base(pretrained: bool | str = False, weights: Mask2FormerSwinBaseWeights | None = None, overrides: object = {})Mask2Former with Swin-Base backbone (Cheng et al., CVPR 2022).
Builds a Mask2FormerForSemanticSegmentation with the Swin-Base
backbone (embed_dim = 128, depths = (2, 2, 18, 2),
num_heads = (4, 8, 16, 32)). ADE20K validation mIoU of 53.9%.
Model Size
Parameters
pretrainedbool or str= FalsePretrained-weight selector.
False → random init; True
→ the DEFAULT tag; a tag string → that specific checkpoint.Explicit weights enum member. Takes precedence over
pretrained.**overridesobject= {}Keyword overrides forwarded into
Mask2FormerConfig.Returns
Mask2FormerForSemanticSegmentationSegmentation model with the Mask2Former-Swin-B configuration applied
(or with overrides merged on top of it).
Examples
>>> import lucid
>>> from lucid.models.vision.mask2former import mask2former_swin_base
>>> model = mask2former_swin_base()
>>> x = lucid.randn(1, 3, 384, 384)
>>> out = model(x)
>>> out.logits.shape[1]
150