# wolny/pytorch-3dunet

3D U-Net model for volumetric semantic segmentation written in pytorch

Repository: https://github.com/wolny/pytorch-3dunet
Canonical: https://ross.abutalabs.com/products/pytorch-3dunet
Language: Jupyter Notebook
License: MIT
License Family: permissive
Topics: semantic-segmentation, unet, unet-pytorch, 3d-unet, groupnorm, pytorch-3dunet, pytorch, dice-coefficient, volumetric-data, 3d-segmentation, residual-unet
Last push: 2025-12-16T14:01:15+00:00

## Health v2 (maintenance only)
Score: 63/100 (v2, computed 2026-09-03T02:20:16.233290+00:00)
- activity 57, release rhythm 50, longevity 100
- inputs: {"age_days": 2903, "days_push": 260, "days_rel": 334, "gap_med": 5.0, "n_releases_24m": 7}
- flags: none
- formula: round(0.45*activity + 0.35*rhythm + 0.20*longevity); archived -> min(score, 10)

## Adoption (not part of the score)
Stars 2416, forks 563 (observed 2026-08-28T04:06:50.323038+00:00)

## What it is
A PyTorch implementation of 3D U-Net and its variants (residual, squeeze-and-excitation) for volumetric semantic segmentation, with 2D U-Net support as well. It supports training for binary/multi-class segmentation and regression tasks like denoising, using HDF5 input data.

## Use cases
- segment 3D volumetric microscopy data
- train a 3D U-Net in PyTorch
- semantic segmentation of medical volumes
- denoise volumetric data with a U-Net
- run connectomics boundary segmentation
- train a 2D U-Net for cell segmentation

## When to choose
- you need volumetric (3D) semantic segmentation with a proven U-Net implementation
- your data is stored in HDF5 and you want configurable training with dice loss and GroupNorm variants
- you work in microscopy, connectomics, or medical imaging segmentation

## When to avoid
- you need instance segmentation or out-of-the-box pretrained models for general images
- you want a lightweight inference-only tool without a training pipeline
- your project is not Python/PyTorch based

## Facets
- artifact type: library
- maturity: active
- function: machine-learning, deep-learning, image-processing, computer-vision
- domain: deep-learning, computer-vision, image-processing, bioinformatics, healthcare
- platform: python, windows
- tags: unet, semantic-segmentation, volumetric-data, 3d-segmentation, pytorch, hdf5, connectomics, microscopy, medical-imaging, dice-coefficient, linux, macos, gpu

## Member repositories
- wolny/pytorch-3dunet (main) score 63

## Provenance
- Observed fields: from GitHub, fetched 2026-08-28T04:06:50.323038+00:00.
- Health v2: computed from the inputs above; adoption is never an input.
- Inferred fields (summary, facets, guidance): AI-extracted, prompt v1, taxonomy v1, on 2026-08-30T02:33:10.250645+00:00, confidence not recorded.
  - readme: https://github.com/wolny/pytorch-3dunet (fetched 2026-08-28T04:06:50.323038+00:00, sha 0a124aedcd39)
- Data as of 2026-08-30T08:39:29.467469+00:00.