OpenViSTAI · Agents · Blockchain · Crypto
Submit a project
Menu
AI Models

Adaptsegnet

Adaptsegnet is a PyTorch research implementation of a method for adapting semantic segmentation from synthetic source-domain images to real target-domain images.

Explore the practical guide ↓

Independent OpenViST project profile, based on linked sources.

Adaptsegnet — OpenViST project profile in AI Models.
OpenViST project illustration. See the linked sources for product details.
Understand the project

Practical overview

When it fits

Use it as research code when investigating whether a semantic-segmentation model trained with synthetic imagery can transfer to a real-image target domain.

Who it is for

  • Computer-vision researchers studying synthetic-to-real domain adaptation.
  • Researchers comparing semantic-segmentation adaptation methods on GTA5-to-Cityscapes or related documented settings.

Documented uses

  • Train a semantic-segmentation model with synthetic source data and real target-domain data.
  • Evaluate documented pretrained models on Cityscapes.
  • Compare single-level and multi-level adaptation configurations, including an LS-GAN training example.
Before you choose

Requirements & limits

The repository is public. Its README states that the code and model are available for non-commercial research purposes only; no license is supplied in the repository evidence.

  • The documented environment is based on older PyTorch, Python and CUDA versions.
  • The README warns that multi-level training may require smaller adversarial-loss weights.
  • The project supplies evaluation and training examples, but it is research code rather than a hosted segmentation service.

Check current requirements in the official documentation. Source-based guidance does not mean OpenViST has run or independently tested this project.

Source-based practical guide

Evaluate a pretrained GTA5-to-Cityscapes model

Run the repository's documented Cityscapes evaluation using its pretrained multi-level model.

Before you start

  • The AdaptSegNet repository and its evaluation code.
  • PyTorch with the documented Python and CUDA setup.
  • Cityscapes data arranged under data/Cityscapes.
  • The pretrained GTA5-to-Cityscapes multi-level model placed in the model folder.

Steps

  1. Clone the repository and change into its directory.

  2. Download the documented pretrained multi-level model and place it in the model folder.

  3. Run the command below.

    Documented command
    python evaluate_cityscapes.py --restore-from ./model/GTA2Cityscapes_multi-ed35151c.pth

  4. Find the evaluation results in the result folder.

Expected outcome

The evaluation script writes results to the result folder.

Guide limitations

  • The README describes the setup with older Python, PyTorch and CUDA versions.
  • The instructions depend on the specified model file and Cityscapes directory layout.
  • The repository states that its code and model are for non-commercial research purposes only.
Documentation 1 ↗
Source-based practical guide

Prepare the GTA5-to-Cityscapes training data

Place the documented source and target datasets in the repository's expected directories.

Before you start

  • The AdaptSegNet repository.
  • The GTA5 dataset as the source domain.
  • The Cityscapes dataset as the target domain.

Steps

  1. Download GTA5 and place it in data/GTA5.

  2. Download Cityscapes and place it in data/Cityscapes.

Expected outcome

The two datasets are arranged in the locations specified for the documented benchmark setup.

Guide limitations

  • This guide covers dataset placement only, not model training or evaluation.
  • Dataset access and any separate dataset terms are outside the repository instructions.
Documentation 1 ↗
Evidence and transparency

Sources & review

Official repository ↗
Source record details
Source availability
Active
Sources checked
18 October 2026
Original website title
wasidennis/AdaptSegNet · Official homepage title · retained in its source language
How we review sourcesSuggest a correction