MedMLX/nnunet-autopet-mlx

nnU-Net · AutoPET

Segment tumor labels from paired CT and FDG PET using the autoPET II nnU-Net checkpoint.

GitHub · Files · Docs · Paper

Input	Output
CT + SUV PET · NIfTI	Tumor segmentation

Apple Silicon · Metal · Python 3.13

Quick start

uv venv --python 3.13
source .venv/bin/activate

uv pip install "nnunet-mlx>=0.6.4"
uv tool install huggingface-hub

hf download MedMLX/nnunet-autopet-mlx --revision 459d23802ae289be0f6539d40ca70c7facb1a9de --local-dir models

Replace the example input paths with your own files.

from nnunet_mlx.medmlx import run

result = run(
    weights_path='models',
    image_paths=['ct.nii.gz', 'suv.nii.gz'],
    configuration='3d_fullres_resenc_bs80',
    output_dir='outputs',
)
print(result["outputs"])

Use aligned CT and SUV PET volumes, in the order shown.

About nnunet-mlx

See the project README for package installation and current documentation.

Use & license

For research and education. Not validated for clinical use.

Weights: CC-BY-4.0. Notices.

Paper references

• Source checkpoint: Look Ma, no code: fine tuning nnU-Net for the AutoPET II challenge.

Citation

If you use MedMLX, please cite the MedMLX paper:

@article{sandoval2026medmlx,
  title   = {MedMLX: On-Device Medical Imaging AI for Apple Silicon},
  author  = {Sandoval, A. and Sandoval, J. and Saleem, A.},
  journal = {medRxiv},
  year    = {2026},
  doi     = {10.64898/2026.10.28.26366762}
}