# ternaus/TernausNet

UNet model with VGG11 encoder pre-trained on Kaggle Carvana dataset

Repository: https://github.com/ternaus/TernausNet
Canonical: https://ross.abutalabs.com/products/ternausnet
Homepage: https://arxiv.org/abs/1801.05746
Language: Python
License: MIT
License Family: permissive
Topics: pytorch, image-segmentation
Last push: 2022-10-26T02:20:58+00:00

## Health v2 (maintenance only)
Score: 32/100 (v2, computed 2026-09-03T02:39:23.370411+00:00)
- activity 0, release rhythm 35, longevity 100
- inputs: {"age_days": 3168, "days_push": 1408, "days_rel": null, "gap_med": null, "n_releases_24m": 0}
- flags: no_releases
- formula: round(0.45*activity + 0.35*rhythm + 0.20*longevity); archived -> min(score, 10)

## Adoption (not part of the score)
Stars 1063, forks 244 (observed 2026-08-28T04:03:26.291888+00:00)

## What it is
TernausNet is a PyTorch implementation of the U-Net architecture with a VGG11 encoder pre-trained on ImageNet for image segmentation. It was part of the winning solution of the Kaggle Carvana Image Masking Challenge and ships with pre-trained weights via pip.

## Use cases
- segment images with a pretrained UNet in PyTorch
- train a binary image segmentation model on a small dataset
- segment medical or satellite imagery
- fine-tune a UNet with an ImageNet-pretrained encoder
- reproduce a Kaggle Carvana image masking solution

## When to choose
- you need a proven UNet variant with pretrained encoder weights for segmentation
- you have limited training data and want faster convergence
- you want a simple pip-installable PyTorch segmentation model

## When to avoid
- you need multi-class or modern architectures like DeepLab or SegFormer
- you need a actively developed library with frequent updates
- you work outside PyTorch

## Facets
- artifact type: library
- maturity: maintenance
- function: machine-learning, deep-learning, computer-vision, image-processing
- domain: computer-vision, deep-learning, machine-learning, image-processing
- platform: python
- tags: unet, image-segmentation, pytorch, pretrained-encoder, vgg11, semantic-segmentation

## Member repositories
- ternaus/TernausNet (main) score 32

## Provenance
- Observed fields: from GitHub, fetched 2026-08-28T04:03:26.291888+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 2026-08-30T06:56:41.420212+00:00, confidence not recorded.
  - readme: https://github.com/ternaus/TernausNet (fetched 2026-08-28T04:03:26.291888+00:00, sha 79c4b21cf63e)
  - homepage: https://arxiv.org/abs/1801.05746 (fetched 2026-08-29T12:58:29.753259+00:00, sha 484d4af0f8f6)
  - site_page: https://info.arxiv.org/about/donate.html (fetched 2026-08-29T12:58:29.757046+00:00, sha cca9c3a11c56)
  - site_page: https://info.arxiv.org/about/ourmembers.html (fetched 2026-08-29T12:58:29.760516+00:00, sha 47cbc55ff1de)
  - site_page: https://info.arxiv.org/about (fetched 2026-08-29T12:58:29.762397+00:00, sha a1f16f915a9a)
  - registry_pypi: https://pypi.org/pypi/ternausnet/json (fetched 2026-08-29T12:58:29.764582+00:00, sha 9ac1d18b1926)
  - site_page: https://info.arxiv.org/labs/index.html (fetched 2026-08-29T12:58:29.758894+00:00, sha b14a8d05a0ec)
- Data as of 2026-08-30T08:39:29.467469+00:00.