Skip to content

Latest commit

 

History

History
89 lines (72 loc) · 2.79 KB

README_en.md

File metadata and controls

89 lines (72 loc) · 2.79 KB
Raw

urban_seg

JPG Image GIF Image JPG Image GIF Image

This project is a beginner-friendly semantic segmentation project based on remote sensing images. We use the Unicom model pretrained on 400 million images, which is highly efficient and performs exceptionally well on remote sensing segmentation tasks. Surprisingly, we achieve excellent results by training the model with only 4 remote sensing images. If you want to get started quickly, you can use train_one_gpu.py for training, which is a simple code with only 200 lines. However, if you aim for better performance, you can try a slightly more complex code, train_multi_gpus.py, which supports multi-GPU training.

Please note that train_multi_gpus.py may require some additional configurations and settings to run multi-GPU training correctly. Make sure to carefully read the instructions and documentation in the code to ensure proper setup and configuration.

Installation

git clone https://github.com/anxiangsir/urban_seg.git

Usage

Install Dependencies

pip install -r requirements.txt

Data and Pretrained Models

The dataset provided by CCF Satellite Image AI Classification and Recognition competition contains a total of five satellite remote sensing images.

Baidu Cloud, Password: 3ih2

dataset
├── origin // 5 remote sensing images with labels
├── test   // 3 remote sensing images without labels, not used in this task
└── train  // empty, generated by random sampling using `python preprocess.py`
    ├── images
    └── labels
FP16-ViT-B-32.pt
FP16-ViT-B-16.pt
FP16-ViT-L-14.pt
FP16-ViT-L-14-336px.pt

Training on a Single GPU

  1. Download the dataset to the current directory.
  2. Preprocess the data.
python preprocess.py
  1. Train the model.
python train_one_gpu.py

Training on Eight GPUs

  1. Download the dataset to the current directory.
  2. Preprocess the data.
python preprocess.py
  1. Train the model.
torchrun --nproc_per_node 8 train_multi_gpus.py

Discussion and Feedback

QQ Group: 679897048

Cite Us

If you find this project useful, please consider citing our paper:

@inproceedings{an2023unicom,
  title={Unicom: Universal and Compact Representation Learning for Image Retrieval},
  author={An, Xiang and Deng, Jiankang and Yang, Kaicheng and Li, Jiawei and Feng, Ziyong and Guo, Jia and Yang, Jing and Liu, Tongliang},
  booktitle={ICLR},
  year={2023}
}