PaddlePaddle是百度开源的深度学习平台,官网地址:https://www.paddlepaddle.org.cn
PaddleOCR依赖 PaddlePaddle,根据不同硬件(CPU、GPU)安装不同版本
我是基于docker安装的,移植性好,安装文档
安装步骤
1.下载PaddleOCR的源码,地址:https://github.com/PaddlePaddle/PaddleOCR.git
1.1.源码目录中的 doc、deploy 有大量文档和部署文档,可以多看看
2.在PaddleOCR源码文件夹中新建inference目录,把3个模型下载好放入inference目录
模型名称 | 下载地址 |
---|---|
检测:原始超轻量模型,支持中英文、多语种文本检测 | https://paddleocr.bj.bcebos.com/dygraph_v2.0/ch/ch_ppocr_mobile_v2.0_det_infer.tar |
分类:原始分类器模型,对检测到的文本行文字角度分类 | https://paddleocr.bj.bcebos.com/dygraph_v2.0/ch/ch_ppocr_mobile_v2.0_cls_infer.tar |
识别:通用推理模型,支持中英文、数字识别 | https://paddleocr.bj.bcebos.com/dygraph_v2.0/ch/ch_ppocr_server_v2.0_rec_infer.tar |
3.把PaddleOCR文件夹压缩为:PaddleOCR.tar.gz,编译Docker镜像时会用到此文件
4.准备一个启动脚本:start.sh,复制到镜像的/PaddleOCR目录,用于docker容器启动时执行该脚本来启动PaddleOCR服务
#!/bin/bash
echo $(date +%F%n%T) >> /PaddleOCR/paddle_log.log
#通过config.json配置文件启动ocr_system可以更好的控制启动参数
hub install /PaddleOCR/deploy/hubserving/ocr_system/ && hub serving start -c /PaddleOCR/deploy/hubserving/ocr_system/config.json
4.Dockerfile文件内容:
# Version: 2.4.2
FROM registry.baidubce.com/paddlepaddle/paddle:2.4.2
COPY ["PaddleOCR.tar.gz", "start.sh", "/"]
# PaddleOCR base on Python3.7
RUN pip install --upgrade pip -i https://mirror.baidu.com/pypi/simple \
&& pip install paddlehub --upgrade -i https://mirror.baidu.com/pypi/simple \
&& pip install PyMuPdf==1.18.14 \
&& tar -xvzf /PaddleOCR.tar.gz -C / \
&& mv /start.sh /PaddleOCR \
&& chmod 777 /PaddleOCR/start.sh
WORKDIR /PaddleOCR
RUN pip install -r requirements.txt -i https://mirror.baidu.com/pypi/simple
# Download orc detect model(light version). if you want to change normal version, you can change ch_ppocr_mobile_v2.0_det_infer to ch_ppocr_server_v2.0_det_infer, also remember change det_model_dir in deploy/hubserving/ocr_system/params.py)
#ADD {link} /PaddleOCR/inference/
#RUN tar xf /PaddleOCR/inference/{file} -C /PaddleOCR/inference/
# Download direction classifier(light version). If you want to change normal version, you can change ch_ppocr_mobile_v2.0_cls_infer to ch_ppocr_mobile_v2.0_cls_infer, also remember change cls_model_dir in deploy/hubserving/ocr_system/params.py)
#ADD {link} /PaddleOCR/inference/
#RUN tar xf /PaddleOCR/inference/{file}.tar -C /PaddleOCR/inference/
# Download orc recognition model(light version). If you want to change normal version, you can change ch_ppocr_mobile_v2.0_rec_infer to ch_ppocr_server_v2.0_rec_infer, also remember change rec_model_dir in deploy/hubserving/ocr_system/params.py)
#ADD {link} /PaddleOCR/inference/
#RUN tar xf /PaddleOCR/inference/{file}.tar -C /PaddleOCR/inference/
EXPOSE 8868
#CMD ["/bin/bash","-c","hub install deploy/hubserving/ocr_system/ && hub serving start -m ocr_system"]
CMD ["/bin/bash","-c","/PaddleOCR/start.sh"]
5.编译镜像:docker build -t paddleocr_cpu:2.4.2 .
6.启动参数:/PaddleOCR/deploy/hubserving/ocr_system/config.json ,需要在容器中改一下
{
"modules_info": {
"ocr_system": {
"init_args": {
"version": "1.0.0",
"use_gpu": true #改为false
},
"predict_args": {
}
}
},
"port": 8868,
"use_multiprocess": false, #如果要多进程,则改为true
"workers": 2 #可以根据服务器核心数调整此参数
}
参数详解:
use_gpu :是否启用GPU,默认true,cpu模式需要设置为false,否则容器会启动失败
use_multiprocess :是否多进程模式,默认false单进程
workers : 工作线程数,默认 2
6.启动容器:docker run -itd --name paddleocr -p 8868:8866 -v /data/paddle:/home/paddle paddleocr_cpu:2.4.2
查看启动日志: docker logs paddleocr
2023-07-07 14:54:08
[2023-07-07 14:54:13,126] [ INFO] - Successfully installed ocr_system-1.0.0
[2023-07-07 14:54:17,372] [ WARNING] - The _initialize method in HubModule will soon be deprecated, you can use the __init__() to handle the initialization of the object
* Serving Flask app 'paddlehub.serving.app_compat'
* Debug mode: off
WARNING: This is a development server. Do not use it in a production deployment. Use a production WSGI server instead.
* Running on all addresses (0.0.0.0)
* Running on http://127.0.0.1:8866
* Running on http://172.17.0.2:8866
Press CTRL+C to quit
7.发送rest请求,识别图片:curl -H “Content-Type:application/json” -X POST --data “{“images”: [“图片的Base64编码”]}” http://localhost:8868/predict/ocr_system
Base64编码不要 “data:image/png;base64,” 下面的请求是已经把一张图片转换为base64编码后发起识别
curl -H "Content-Type:application/json" -X POST --data "{\"images\": [\"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\"]}" http://localhost:8868/predict/ocr_system
得到识别的结果:
{"msg":"","results":[[{"confidence":0.7192772626876831,"text":"韩国小馆","text_region":[[8,2],[348,2],[348,68],[8,68]]}]],"status":"000"}