# milesial/Pytorch-UNet

PyTorch implementation of the U-Net for image semantic segmentation with high quality images

Repository: https://github.com/milesial/Pytorch-UNet
Canonical: https://ross.abutalabs.com/products/pytorch-unet
Language: Python
License: GPL-3.0
License Family: copyleft
Topics: pytorch, unet, pytorch-unet, semantic-segmentation, convolutional-networks, kaggle, tensorboard, wandb, weights-and-biases, deep-learning, convolutional-neural-networks
Last push: 2024-08-11T03:00:15+00:00

## Health v2 (maintenance only)
Score: 23/100 (v2, computed 2026-09-02T17:46:02.011165+00:00)
- activity 0, release rhythm 8, longevity 100
- inputs: {"age_days": 3304, "days_push": 752, "days_rel": null, "gap_med": null, "n_releases_24m": 0}
- flags: none
- formula: round(0.45*activity + 0.35*rhythm + 0.20*longevity); archived -> min(score, 10)

## Adoption (not part of the score)
Stars 11613, forks 2765 (observed 2026-08-28T04:10:48.969816+00:00)

## What it is
A PyTorch implementation of the U-Net architecture for semantic segmentation of high-resolution images, originally built for Kaggle's Carvana Image Masking Challenge. It includes training and prediction scripts, a pretrained model, and Docker support.

## Use cases
- segment images into semantic classes with a U-Net in PyTorch
- train an image masking model on high definition images
- run medical image segmentation
- do portrait or multiclass segmentation
- reproduce the Carvana image masking challenge solution
- get a pretrained U-Net segmentation model

## When to choose
- you need a well-tested, popular U-Net reference implementation in PyTorch
- you want a pretrained segmentation model with a high Dice score
- you want a simple training script with AMP, TensorBoard, and Weights & Biases support
- you want a Docker image to run training on GPU machines

## When to avoid
- you need segmentation architectures beyond U-Net (e.g., Transformer-based models)
- you need a production inference server rather than training/prediction scripts
- your project cannot accept GPL-3.0 licensing
- you work outside PyTorch (e.g., TensorFlow or JAX)

## Facets
- artifact type: library
- maturity: active
- function: machine-learning, deep-learning, image-processing, computer-vision
- domain: deep-learning, computer-vision, image-processing, machine-learning
- platform: python, cross-platform
- tags: unet, semantic-segmentation, pytorch, image-masking, pretrained-model, kaggle, gpu, docker, linux

## Member repositories
- milesial/Pytorch-UNet (main) score 23

## Provenance
- Observed fields: from GitHub, fetched 2026-08-28T04:10:48.969816+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 2026-08-29T17:15:28.166557+00:00, confidence not recorded.
  - readme: https://github.com/milesial/Pytorch-UNet (fetched 2026-08-28T04:10:48.969816+00:00, sha ee42043b14b7)
- Data as of 2026-08-30T08:39:29.467469+00:00.