开源易用的中文离线OCR,识别率媲美大厂,并且提供了易用的web页面及web的接口,方便人类日常工作使用或者其他程序来调用~
APACHE-2.0 License
TrWebOCR Tr http web
tornado
WindowsMacOSDocker Linux
python3.7 miniconda
pip install -r requirements.txt
8089gpu
python backend/main.py [--port=8089][--open_gpu=0]
# --port 8089
# --open_gpu gpu 0(1
tr 2.3.0 https://github.com/myhub/tr
Server is running: http://192.168.31.95:8089
Now version is: cpu
Dockerfile Pull
# dockerfile
docker build -t trwebocr:latest .
#
docker run -itd --rm -p 8089:8089 --name trwebocr trwebocr:latest
# dockerhub pull
docker pull mmmz/trwebocr:latest
#
docker run -itd --rm -p 8089:8089 --name trwebocr mmmz/trwebocr:latest
80898089run-p 8089:8089
dockerIP8089
import requests
url = 'http://192.168.31.108:8089/api/tr-run/'
img1_file = {
'file': open('img1.png', 'rb')
}
res = requests.post(url=url, data={'compress': 0}, files=img1_file)
import requests
import base64
def img_to_base64(img_path):
with open(img_path, 'rb')as read:
b64 = base64.b64encode(read.read())
return b64
url = 'http://192.168.31.108:8089/api/tr-run/'
img_b64 = img_to_base64('./img1.png')
res = requests.post(url=url, data={'img': img_b64})
20220123 tr2.3.1
~~CPU
20220116
20200817 Dockerfiledockertr2.3
20200730 GPU/CPU
Apache 2.0
GitHub clone~
star (^.^)