Mask2former Swin Base Ade

--- library_name: lucid license: other tags: - semantic-segmentation - mask2former - lucid datasets: - ade20k pipeline_tag: image-segmentation model-index: - name: mask2former-swin-base results: - task: { type: semantic-segmentation } dataset: { name: ADE20K, type: ade20k } metrics: - { type: mIoU, value: 53.9 } --- # Mask2Former (Swin-Base) > Cheng et al., 2022 — *Masked-attention Mask Transformer for Universal Image Segmentation* (arXiv:2112.01527) [Lucid](https://github.com/ChanLumerico/lucid) port of `facebook/mask2former-swin-base-ade-semantic`, converted to Lucid-native safetensors. ## Available weights | Tag | mIoU | Params | GFLOPs | Size | Source | |---|---|---|---|---|---| | `ADE20K` *(default)* | 53.9 | 106.9M | — | 407.98 MB | facebook | ## Usage ```python import lucid.models as models from lucid.models.weights import Mask2FormerSwinBaseWeights # default tag model = models.mask2former_swin_base(pretrained=True) # explicit tag (enum or string) model = models.mask2former_swin_base(weights=Mask2FormerSwinBaseWeights.ADE20K) model = models.mask2former_swin_base(pretrained="ADE20K") # preprocessing travels with the weights weights = Mask2FormerSwinBaseWeights.ADE20K preprocess = weights.transforms() out = model(preprocess(image)[None]) # SemanticSegmentationOutput: per-pixel class logits (B, C, H, W) seg = out.logits.argmax(axis=1) # (B, H, W) class indices ``` ## Conversion Converted from `facebook/mask2former-swin-base-ade-semantic` via `python -m tools.convert_weights mask2former_swin_base --tag ADE20K`. Key mapping + numerical parity verified against the source. ## License `other` — inherited from the original weights. ## Citation ``` @inproceedings{cheng2022mask2former, title={Masked-attention Mask Transformer for Universal Image Segmentation}, author={Cheng, Bowen and Misra, Ishan and Schwing, Alexander G. and Kirillov, Alexander and Girdhar, Rohit}, booktitle={Proceedings of the IEEE/CVF Conference on Computer Vision and Pattern Recognition (CVPR)}, year={2022} } ```

Source context: 0 downloads · 0 likes · Pipeline image-segmentation · Library lucid · Repo lucid-dl/mask2former-swin-base-ade