jbnu-vclab/pnid_recognition_project

P&ID Recognition Project

pnid_recognition_project

P&ID Recognition Project

Training example

python ./tools/training/pnid_train.py ./options_example/pnid_train_example.yaml

Installation

Our project depends on MMRotate. Please refer to installation guidance below. For additional, you can refer MMRotate Install Guide for more detailed instruction.
It requires Python 3.7+, CUDA 9.2+ and PyTorch 1.6+.

1. create conda environment

conda create --n pnid python=3.8 -y
conda activate pnid

2. install pytorch (official page)

conda install pytorch==1.8.0 torchvision==0.9.0 torchaudio==0.8.0 cudatoolkit=11.1 -c pytorch -c conda-forge

3. install mmrotate

pip install -U openmim
mim install mmcv-full
mim install mmdet<3.0.0

cd mmrotate
pip install -v -e .

(Optional) verify mmrotate installation

mim download mmrotate --config oriented_rcnn_r50_fpn_1x_dota_le90 --dest .
python demo/image_demo.py demo/demo.jpg oriented_rcnn_r50_fpn_1x_dota_le90.py oriented_rcnn_r50_fpn_1x_dota_le90-6d2b2ce0.pth --out-file result.jpg

4. install our project

cd ..
pip install -v -e .

pnid_recognition_project's Issues

exe 입력 인자 제한

yaml은 python 코드를 실행할 때만 사용하고, exe로 빌드된 코드는 꼭 필요한 argument (예를 들어 image path, 텍스트 인식의 경우 image_path & xml path 등) 만 집어넣어 실행되도록 할 것

Inference config symbol/text 주석처리하지 않아도 동작하도록 수정

pnid_inference_test.yaml에서 symbol/text를 주석 처리하지 않아도 동작하도록 수정

common/fileutil 모듈명 수정

fileutil? file_util?

MMOCR config 에서 albumentation 제거

Dataset gen config에서 저장되는 yaml을 입력 yaml과 동일하도록 수정