https://github.com/LeapLabTHU/Rank-DETR
[NeurIPS 2023] Rank-DETR for High Quality Object Detection
https://github.com/LeapLabTHU/Rank-DETR
Last synced: about 2 months ago
JSON representation
[NeurIPS 2023] Rank-DETR for High Quality Object Detection
- Host: GitHub
- URL: https://github.com/LeapLabTHU/Rank-DETR
- Owner: LeapLabTHU
- License: apache-2.0
- Created: 2023-10-12T03:02:25.000Z (over 1 year ago)
- Default Branch: main
- Last Pushed: 2023-10-19T07:34:15.000Z (over 1 year ago)
- Last Synced: 2024-11-05T12:33:38.025Z (7 months ago)
- Language: Python
- Homepage:
- Size: 8.88 MB
- Stars: 87
- Watchers: 2
- Forks: 6
- Open Issues: 1
-
Metadata Files:
- Readme: README.md
- License: LICENSE
Awesome Lists containing this project
README
# Rank-DETR for High Quality Object Detection (NeurIPS 2023)
Yifan Pu, Weicong Liang, Yiduo Hao, Yuhui Yuan, Yukang Yang, Chao Zhang, Han Hu, and Gao Huang
[[`arXiv`](https://arxiv.org/abs/2310.08854)] [[`BibTeX`](#citing-rank-detr)]
## Table of Contents
- [Installation](#installation)
- [Pretrained Models](#pretrained-models)
- [Run](#run)
- [Training](#training)
- [Evaluation](#evaluation)
- [Citation](#citing-rank-detr)
## Installation
Please refer to the [installation document](https://detrex.readthedocs.io/en/latest/tutorials/Installation.html) of detrex.
## Pretrained Models
Here we provide the Rank-DETR model pretrained weights based on detrex:
Name
Backbone
Query Num
Epochs
AP
download
Rank-DETR
R50
300
12
50.2
model
Rank-DETR
R50
300
36
51.2
model
Rank-DETR
Swin Tiny
300
12
52.7
model
Rank-DETR
Swin Tiny
300
36
54.7
model
Rank-DETR
Swin Large
300
12
57.3
model
Rank-DETR
Swin Large
300
36
58.2
model
## Run
### Training
All configs can be trained with:
```bash
cd detrex
python projects/rank_detr/train_net.py --config-file projects/rank_detr/configs/path/to/config.py --num-gpus 8
```
* By default, we use 8 GPUs with total batch size as 16 for training.
* To train/eval a model with the swin transformer backbone, you need to download the backbone from the [offical repo](https://github.com/microsoft/Swin-Transformer#main-results-on-imagenet-with-pretrained-models) frist and specify argument `train.init_checkpoint` like [our configs](./configs/rank_detr_swin_tiny_two_stage_12ep.py).
### Evaluation
Model evaluation can be done as follows:
```bash
cd detrex
python projects/rank_detr/train_net.py --config-file projects/rank_detr/configs/path/to/config.py --eval-only train.init_checkpoint=/path/to/model_checkpoint
```
## Citing Rank-DETR
If you find Rank-DETR useful in your research, please consider citing:
```bibtex
@inproceedings{pu2023rank,
title={Rank-DETR for High Quality Object Detection},
author={Pu, Yifan and Liang, Weicong and Hao, Yiduo and Yuan, Yuhui and Yang, Yukang and Zhang, Chao and Hu, Han and Huang, Gao},
booktitle={NeurIPS},
year={2023}
}
```