Semantic Segmentation in PyTorch

This repository contains the implementation of semantic segmentation models in PyTorch.

Prerequisites

• python >= 3.6
• torch >= 1.8.1
• torchvision >= 0.9.1

Usage

1. Clone the repository and install the required dependencies with the following command:

$ git clone https://github.com/woohyun-jeon/pytorch-segmentation.git
$ cd pytorch-segmentation
$ pip install -r requirements.txt

2. Download CityScapes into datasets directory

The directory structure should be as follows:

  datasets/
    CityScapes/      
      gtFine/
        train/
            aachen/
                *_gtFine_labelIds.png
                ...
            bochum/
                *_gtFine_labelIds.png
                ...
            ...
        val/
        test/
      leftImg8bit/
        train/
            aachen/
                *_leftImg8bit.png
                ...
            bochum/
                *_leftImg8bit.png
                ...
            ...
        val/
        test/
      

3. Run python train.py for training

Supported Models

• UNet
• UNet++
• UNet3+
• AttentionUNet
• Swin-UNet
• SegNet
• HRNet
• DeepLabv3
• DeepLabv3+
• PSPNet
• PAN