mask2former_swin_large

lucid.models.mask2former_swin_large(num_labels: int, *, pretrained_backbone: bool = False, **config_kwargs: Any) Mask2Former

The mask2former_swin_large function builds a Mask2Former model with a Swin-Large backbone preset.

Total Parameters (num_labels=150): 215,488,779

Function Signature

@register_model
def mask2former_swin_large(
    num_labels: int,
    *,
    pretrained_backbone: bool = False,
    **config_kwargs
) -> Mask2Former

Parameters

  • num_labels (int): Number of segmentation classes.

  • pretrained_backbone (bool, optional): If True, initializes the backbone from available pretrained weights.

  • config_kwargs (dict, optional): Additional overrides applied to Mask2FormerConfig.

Returns

  • Mask2Former: Mask2Former model configured with a Swin-Large backbone.

Example Usage

from lucid.models.vision.mask2former import mask2former_swin_large
import lucid

model = mask2former_swin_large(num_labels=150)
x = lucid.random.randn(1, 3, 384, 384)
pred = model.predict(x)
print(pred.shape)

Pretrained Weights

import lucid.models as models
import lucid.weights as W

model = models.mask2former_swin_large(
    num_labels=150,
    weights=W.Mask2Former_Swin_Large_Weights.ADE20K_SEMANTIC,
)
import lucid.models as models
import lucid.weights as W

model = models.mask2former_swin_large(
    num_labels=150,
    weights=W.Mask2Former_Swin_Large_Weights.ADE20K_PANOPTIC,
)