yangjun dfa27afb39 提交PaddleDetection develop 分支 d56cf3f7c294a7138013dac21f87da4ea6bee829 пре 1 година
..
_base_ dfa27afb39 提交PaddleDetection develop 分支 d56cf3f7c294a7138013dac21f87da4ea6bee829 пре 1 година
README.md dfa27afb39 提交PaddleDetection develop 分支 d56cf3f7c294a7138013dac21f87da4ea6bee829 пре 1 година
detr_r50_1x_coco.yml dfa27afb39 提交PaddleDetection develop 分支 d56cf3f7c294a7138013dac21f87da4ea6bee829 пре 1 година

README.md

DETR

Introduction

DETR is an object detection model based on transformer. We reproduced the model of the paper.

Model Zoo

Backbone Model Images/GPU Inf time (fps) Box AP Config Download
R-50 DETR 4 --- 42.3 config model

Notes:

  • DETR is trained on COCO train2017 dataset and evaluated on val2017 results of mAP(IoU=0.5:0.95).
  • DETR uses 8GPU to train 500 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/detr/detr_r50_1x_coco.yml --fleet

Citations

@inproceedings{detr,
  author    = {Nicolas Carion and
               Francisco Massa and
               Gabriel Synnaeve and
               Nicolas Usunier and
               Alexander Kirillov and
               Sergey Zagoruyko},
  title     = {End-to-End Object Detection with Transformers},
  booktitle = {ECCV},
  year      = {2020}
}