CMDS-AD checkpoints

This repository contains the released CMDS-AD few-shot anomaly-detection checkpoints and the final LoRA weights used by the CMDS-AD pipeline.

CMDS-AD uses cross-modal feature mapping between RGB and 3D representations. The released checkpoint package contains models trained for both mapping directions:

• 2d_to_3d: RGB/image features to 3D features.
• 3d_to_2d: 3D features to RGB/image features.

For each dataset, category, and shot setting, both directions are available. The package includes the 1-shot, 2-shot, and 4-shot checkpoints. The 4shot_mlp training variant is not included in this public package.

Repository layout

checkpoints/
├── mvtec3dad/<category>/<shot>/
│   ├── 2d_to_3d/
│   │   ├── model_est_final.pth
│   │   └── model_real_final.pth
│   └── 3d_to_2d/
│       ├── model_est_final.pth
│       └── model_real_final.pth
└── eyecandies/<category>/<shot>/
    ├── 2d_to_3d/
    └── 3d_to_2d/

weights/lora/
├── mvtec3dad/<category>/final_lora.safetensors
└── eyecandies/<category>/final_lora.safetensors

The model filenames intentionally use only the information needed for inference. Internal experiment labels are not part of the public naming scheme.

Downloading a checkpoint

The files can be downloaded with huggingface_hub:

from huggingface_hub import hf_hub_download

repo_id = "JesseCai/CMDS-AD"

path_2d_to_3d = hf_hub_download(
    repo_id=repo_id,
    filename="checkpoints/mvtec3dad/bagel/1shot/2d_to_3d/model_est_final.pth",
)

path_3d_to_2d = hf_hub_download(
    repo_id=repo_id,
    filename="checkpoints/mvtec3dad/bagel/1shot/3d_to_2d/model_est_final.pth",
)

The exact category and shot components can be changed without changing the download method. The CMDS-AD source repository contains the corresponding inference scripts and configuration details.

Downloading a LoRA weight

from huggingface_hub import hf_hub_download

lora_path = hf_hub_download(
    repo_id="JesseCai/CMDS-AD",
    filename="weights/lora/mvtec3dad/bagel/final_lora.safetensors",
)

The LoRA files are provided for the RGB generation stage. Stable Diffusion, Marigold, DINO, PointMAE, and the datasets are not duplicated in this repository; please obtain those dependencies from their official sources as described in the CMDS-AD documentation.

Intended use

These weights are intended for research and reproducibility of the CMDS-AD method on the MVTec 3D-AD and Eyecandies benchmarks. They are not intended for safety-critical inspection decisions without additional validation on the target production domain.

The checkpoints are released for inference and evaluation. Users who need to reproduce the complete training pipeline should follow the data preprocessing, normal estimation, LoRA training, feature extraction, mapping training, and inference instructions in the source repository.

Citation

If you use CMDS-AD, please cite the accompanying paper:

@article{cai2026cmdsad,
  title   = {CMDS-AD},
  author  = {Cai, Junhao and others},
  journal = {arXiv preprint arXiv:2606.20300},
  year    = {2026}
}

Paper: https://arxiv.org/abs/2606.20300
Project page: https://cmds-ad.github.io/
Source code: https://github.com/Junhaocai27/CMDS-AD

License and third-party components

The CMDS-AD source code and released weights are provided under the license stated in the source repository. Third-party components and pretrained foundation models remain subject to their original licenses and terms of use.