# CoinCheung/BiSeNet

Add bisenetv2.  My implementation of BiSeNet

Repository: https://github.com/CoinCheung/BiSeNet
Canonical: https://ross.abutalabs.com/products/bisenet
Language: Python
License: MIT
License Family: permissive
Topics: bisenet, cityscapes, pytorch, cocostuff, tensorrt, ncnn, openvino, triton-inference-server, ade20k
Last push: 2026-04-04T10:28:15+00:00

## Health v2 (maintenance only)
Score: 57/100 (v2, computed 2026-09-02T17:46:02.011165+00:00)
- activity 75, release rhythm 8, longevity 100
- inputs: {"age_days": 2834, "days_push": 151, "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 1638, forks 336 (observed 2026-08-28T04:05:15.114650+00:00)

## What it is
A PyTorch implementation of the BiSeNet V1 and V2 real-time semantic segmentation models, with pretrained weights for Cityscapes, COCO-Stuff, and ADE20K. It also provides export and deployment guides for TensorRT, ncnn, OpenVINO, and Triton Inference Server.

## Use cases
- train a real-time semantic segmentation model on cityscapes
- run fast image segmentation on GPU with pretrained bisenet weights
- deploy a segmentation model to tensorrt or ncnn for inference
- evaluate mIoU of bisenet on ade20k or cocostuff
- segment street scenes for autonomous driving research
- convert a pytorch segmentation model to openvino

## When to choose
- you need fast real-time semantic segmentation with proven mIoU benchmarks
- you want pretrained weights on multiple segmentation datasets
- you plan to deploy segmentation models to edge inference runtimes like tensorrt or ncnn

## When to avoid
- you need state-of-the-art accuracy rather than speed-oriented segmentation
- you work outside PyTorch or need non-CUDA training environments
- you need official paper-exact training settings for ade20k

## Facets
- artifact type: library
- maturity: active
- function: machine-learning, deep-learning, image-processing, computer-vision
- domain: computer-vision, image-processing, deep-learning, machine-learning
- platform: python
- tags: semantic-segmentation, bisenet, pytorch, cityscapes, tensorrt, ncnn, openvino, model-deployment, linux, gpu

## Member repositories
- CoinCheung/BiSeNet (main) score 57

## Provenance
- Observed fields: from GitHub, fetched 2026-08-28T04:05:15.114650+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 2026-08-30T03:46:32.398407+00:00, confidence not recorded.
  - readme: https://github.com/CoinCheung/BiSeNet (fetched 2026-08-28T04:05:15.114650+00:00, sha c9401b0f9433)
- Data as of 2026-08-30T08:39:29.467469+00:00.