# yu-changqian/TorchSeg

Fast, modular reference implementation and easy training of Semantic Segmentation algorithms in PyTorch.

Repository: https://github.com/yu-changqian/TorchSeg
Canonical: https://ross.abutalabs.com/products/torchseg
Language: Python
License: MIT
License Family: permissive
Last push: 2020-03-11T09:55:12+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": 2792, "days_push": 2366, "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 1410, forks 249 (observed 2026-08-28T04:04:38.890035+00:00)

## What it is
A fast, modular PyTorch reference implementation for training and evaluating semantic segmentation models such as FCN, DFN, BiSeNet, PSPNet, and PSANet. It supports distributed multi-GPU training and ships pre-trained models with benchmark results on PASCAL VOC, Cityscapes, ADE20K, and CamVid.

## Use cases
- train a semantic segmentation model in pytorch
- benchmark FCN PSPNet BiSeNet on cityscapes
- get pretrained segmentation models with mean IoU results
- distributed multi-gpu training for image segmentation
- reproduce DFN segmentation paper results
- segment images into per-pixel classes

## When to choose
- you need a modular PyTorch codebase for semantic segmentation research
- you want fast distributed multi-GPU training of segmentation models
- you need reference benchmarks and pre-trained weights for FCN, DFN, BiSeNet, PSPNet, or PSANet

## When to avoid
- you need actively maintained code with recent PyTorch version support
- you want general-purpose image classification or object detection
- you need a production inference service rather than a training framework

## Facets
- artifact type: library
- maturity: maintenance
- function: machine-learning, deep-learning, image-processing, computer-vision
- domain: computer-vision, image-processing, deep-learning, machine-learning
- platform: python
- tags: semantic-segmentation, pytorch, cityscapes, pascal-voc, distributed-training, model-zoo, linux, gpu

## Member repositories
- yu-changqian/TorchSeg (main) score 23

## Provenance
- Observed fields: from GitHub, fetched 2026-08-28T04:04:38.890035+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 2026-08-30T04:38:30.460838+00:00, confidence not recorded.
  - readme: https://github.com/yu-changqian/TorchSeg (fetched 2026-08-28T04:04:38.890035+00:00, sha 7eed01c288cb)
- Data as of 2026-08-30T08:39:29.467469+00:00.