MedMLX/flexible-unet-effnetb2-mlx

FlexibleUNet · Endoscopy

Segment surgical tools in RGB endoscopy frames with an EfficientNet-B2 FlexibleUNet.

GitHub · Files · Docs · Paper

Input	Output
RGB endoscopy frame	Full-resolution tool mask

Apple Silicon · Metal · Python 3.13

Quick start

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

uv pip install "flexible-unet-mlx>=0.2.6"
uv tool install huggingface-hub

hf download MedMLX/flexible-unet-effnetb2-mlx --revision ccbbc596f2a9da8477e25cba0d7548cbb2191feb --local-dir models

Replace the example input paths with your own files.

from flexible_unet_mlx.medmlx import run

result = run(
    image_path='frame.jpg',
    weights_path='models/flexibleunet.safetensors',
    output_dir='outputs',
)
print(result["outputs"])

Use & license

For research and education. Not validated for clinical use.

Weights: Apache-2.0. Notices.

Paper references

• Decoder architecture: U-Net: Convolutional Networks for Biomedical Image Segmentation.
• Encoder architecture: EfficientNet: Rethinking Model Scaling for Convolutional Neural Networks.

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}
}