EDTformer

Getting Started

This repo follows the framework of CricaVPR, and the Visual Geo-localization Benchmark. We utilize the GSV-Cities dataset for training and you can download it HERE, and refer to VPR-datasets-downloader to prepare test datasets.

The test dataset should be organized in a directory tree as such:

├── datasets_vg
    └── datasets
        └── pitts30k
            └── images
                ├── train
                │   ├── database
                │   └── queries
                ├── val
                │   ├── database
                │   └── queries
                └── test
                    ├── database
                    └── queries

Before training, you should download the pre-trained foundation model DINOv2(ViT-B/14) HERE.

Train

python3 train.py --eval_datasets_folder=/path/to/your/datasets_vg/datasets --eval_dataset_name=pitts30k --foundation_model_path=/path/to/pre-trained/dinov2_vitb14_pretrain.pth --epochs_num=15

Test

To evaluate the trained model:

python3 eval.py --eval_datasets_folder=/path/to/your/datasets_vg/datasets --eval_dataset_name=msls --resume=/path/to/trained/model/your_model.pth

Trained Model

You can directly download the trained model HERE.

Acknowledgements

Parts of this repo are inspired by the following repositories:

CricaVPR

Visual Geo-localization Benchmark

GSV-Cities

Citation

If you find this repo useful for your research, please consider leaving a star⭐️ and citing the paper

@ARTICLE{EDTformer,
  author={Jin, Tong and Lu, Feng and Hu, Shuyu and Yuan, Chun and Liu, Yunpeng},
  journal={IEEE Transactions on Circuits and Systems for Video Technology}, 
  title={EDTformer: An Efficient Decoder Transformer for Visual Place Recognition}, 
  year={2025},
  doi={10.1109/TCSVT.2025.3559084}}}