Git 설치
https://irenechoiworld.tistory.com/10
Conda 설치 및 PATH 설정
https://irenechoiworld.tistory.com/11
DETR 다운받기
DETR 프로젝트 Clone 받아오기
git clone https://github.com/facebookresearch/detr.git
pip 업데이트
pip install --upgrade pip
numpy 설치
pip install numpy
필요한 라이브러리 설치
sudo apt-get update
sudo apt-get install build-essential python3-devpip install cython
pip install pycocotools
requirements.txt를 설치
pip install -r requirements.txt# requirements.txt
cython
git+https://github.com/cocodataset/cocoapi.git#subdirectory=PythonAPI&egg=pycocotools
submitit
torch>=1.5.0
torchvision>=0.6.0
git+https://github.com/cocodataset/panopticapi.git#egg=panopticapi
scipy
onnx
onnxruntime
PyTorch 1.5+ / torchvision 0.6+ 설치하기
conda install -c pytorch pytorch torchvision
pycocotools (for evaluation on COCO) and scipy (for training) 설치하기
conda install cython scipy
pip install -U 'git+https://github.com/cocodataset/cocoapi.git#subdirectory=PythonAPI'
(optional) to work with panoptic install panopticapi
pip install git+https://github.com/cocodataset/panopticapi.git
Data preparation
COCO 2017 데이터셋 다운받기
COCO - Common Objects in Context
cocodataset.org
# images
wget http://images.cocodataset.org/zips/train2017.zip # train dataset
wget http://images.cocodataset.org/zips/val2017.zip # validation dataset
wget http://images.cocodataset.org/zips/test2017.zip # test dataset
wget http://images.cocodataset.org/zips/unlabeled2017.zip
# annotations
wget http://images.cocodataset.org/annotations/annotations_trainval2017.zip
wget http://images.cocodataset.org/annotations/stuff_annotations_trainval2017.zip
wget http://images.cocodataset.org/annotations/image_info_test2017.zip
wget http://images.cocodataset.org/annotations/image_info_unlabeled2017.zip