72 Star 778 Fork 286

GVPPaddlePaddle/PaddleDetection

加入 Gitee
与超过 1200万 开发者一起发现、参与优秀开源项目,私有仓库也完全免费 :)
免费加入
文件
克隆/下载
贡献代码
同步代码
取消
提示: 由于 Git 不支持空文件夾,创建文件夹后会生成空的 .keep 文件
Loading...
README

TOOD

Introduction

TOOD: Task-aligned One-stage Object Detection

TOOD is an object detection model. We reproduced the model of the paper.

Model Zoo

Backbone Model Images/GPU Inf time (fps) Box AP Config Download
R-50 TOOD 4 --- 42.5 config model

Notes:

  • TOOD is trained on COCO train2017 dataset and evaluated on val2017 results of mAP(IoU=0.5:0.95).
  • TOOD uses 8GPU to train 12 epochs.

GPU multi-card training

export CUDA_VISIBLE_DEVICES=0,1,2,3,4,5,6,7
python -m paddle.distributed.launch --gpus 0,1,2,3,4,5,6,7 tools/train.py -c configs/tood/tood_r50_fpn_1x_coco.yml --fleet

Citations

@inproceedings{feng2021tood,
    title={TOOD: Task-aligned One-stage Object Detection},
    author={Feng, Chengjian and Zhong, Yujie and Gao, Yu and Scott, Matthew R and Huang, Weilin},
    booktitle={ICCV},
    year={2021}
}
马建仓 AI 助手
尝试更多
代码解读
代码找茬
代码优化
Python
1
https://gitee.com/paddlepaddle/PaddleDetection.git
git@gitee.com:paddlepaddle/PaddleDetection.git
paddlepaddle
PaddleDetection
PaddleDetection
release/2.8.1

搜索帮助