COCO panoptic checkpoint
According to the model card, this checkpoint is trained for COCO panoptic segmentation in the large Swin-backbone size.
Open Source Model Profile · facebook
mask2former-swin-large-coco-panoptic
mask2former-swin-large-coco-panoptic is a Mask2Former image-segmentation model from facebook. According to the model card, it is the large Swin-backbone version trained for COCO panoptic segmentation.
- Publisher
- Task
- image-segmentation
- Model type
- mask2former
- License
- other
- Library
- transformers
- Publication status
- Accepted · not indexed
Model overview
mask2former-swin-large-coco-panoptic is published by facebook as an image-segmentation model. The captured configuration identifies Mask2FormerForUniversalSegmentation with model type mask2former. According to the model card, it handles COCO panoptic segmentation with a large-sized Swin backbone.
Recorded capabilities
Mask-plus-label paradigm
According to the model card, Mask2Former treats instance, semantic, and panoptic segmentation as mask prediction with labels.
Documented efficiency changes
According to the model card, it uses multi-scale deformable attention, masked-attention decoding, and subsampled-point loss computation.
Use cases in the source record
- Panoptic segmentation of COCO-style images into masks with corresponding labels.
- Universal-segmentation experiments extending the same mask-plus-label paradigm to instance or semantic tasks.
Limitations and unknowns
- No parameter count was extracted for this record.
- No evaluation results were extracted from this record.
- The model card is Hugging Face-authored rather than publisher-authored, so method detail is secondhand.
- Provider state is historical snapshot data, not independently refreshed current availability.
Source and provenance
Source: facebook/mask2former-swin-large-coco-panoptic
Captured: Unknown. Processed: 2026-09-07T19:34:44.238731+00:00.
Mask2Former Mask2Former model trained on COCO panoptic segmentation (large-sized version, Swin backbone). It was introduced in the paper Masked-attention Mask Transformer for Universal Image Segmentation and first released in this repository . Disclaimer: The team releasing Mask2Former did not write a model card for this model so this model card has been written by the Hugging Face team. Model description Mask2Former addresses instance, semantic and panoptic segmentation with the same paradigm: by predicting a set of masks and corresponding labels. Hence, all 3 tasks are treated as if they were instance segmentation. Mask2Former out…
F001F002F003F004F005F006F007F008F009F010F011F012F013