MaskFormer Swin base build
Captured config identifies MaskFormerForInstanceSegmentation, and the card describes a base-sized version with a Swin backbone.
Open Source Model Profile · facebook
maskformer-swin-base-ade
maskformer-swin-base-ade is a MaskFormer image-segmentation model from facebook. Its model card documents a Swin-base build trained on ADE20k semantic segmentation.
- Publisher
- Task
- image-segmentation
- Model type
- maskformer
- License
- other
- Library
- transformers
- Publication status
- Accepted · not indexed
Model overview
maskformer-swin-base-ade is published by facebook as an image-segmentation model. The captured configuration identifies MaskFormerForInstanceSegmentation with a maskformer model type. According to the model card, it is a base-sized MaskFormer model with a Swin backbone trained on ADE20k semantic segmentation.
Recorded capabilities
ADE20k semantic checkpoint
According to the model card, this checkpoint is trained for ADE20k semantic segmentation.
Unified mask-label paradigm
The card says the model addresses instance, semantic, and panoptic segmentation by predicting masks with corresponding labels.
Hugging Face written card
The model card carries a disclaimer that the Hugging Face team, not the releasing team, wrote the card.
Use cases in the source record
- Semantic segmentation on ADE20k-style scenes using this documented checkpoint.
- Mask-and-label segmentation experiments that treat semantic output under the card's unified instance-style paradigm.
Limitations and unknowns
- No parameter count was extracted from this record.
- No evaluation results were extracted from this record.
- Provider state is historical snapshot data, not independently refreshed current availability.
- Card content was written by the Hugging Face team rather than the releasing team, so publisher-level detail is limited.
Source and provenance
Source: facebook/maskformer-swin-base-ade
Captured: Unknown. Processed: 2026-09-07T19:34:44.364640+00:00.
MaskFormer MaskFormer model trained on ADE20k semantic segmentation (base-sized version, Swin backbone). It was introduced in the paper Per-Pixel Classification is Not All You Need for Semantic Segmentation and first released in this repository . Disclaimer: The team releasing MaskFormer did not write a model card for this model so this model card has been written by the Hugging Face team. Model description MaskFormer 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. Intended uses…
F001F002F003F004F005F006F007F008F009F010F011F012F013