mask2former_swin_small(pretrained: bool | str = False, weights: Mask2FormerSwinSmallWeights | None = None, overrides: object = {})Mask2Former with Swin-Small backbone (Cheng et al., CVPR 2022).
Builds a Mask2FormerForSemanticSegmentation with the Swin-Small
backbone (embed_dim = 96, depths = (2, 2, 18, 2),
num_heads = (3, 6, 12, 24)). Same transformer head as
mask2former_swin_tiny; ADE20K validation mIoU of 51.3%.
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-S configuration applied
(or with overrides merged on top of it).
Examples
>>> import lucid
>>> from lucid.models.vision.mask2former import mask2former_swin_small
>>> model = mask2former_swin_small()
>>> x = lucid.randn(1, 3, 384, 384)
>>> out = model(x)
>>> out.logits.shape[1]
150