Skip to content

GitLab

Commit f56dcc58 authored by Your Name's avatar Your Name
Browse files

提交Crnn推理C++示例

parent dd02a3eb
Show whitespace changes
Inline Side-by-side
Showing with 593 additions and 14 deletions
.gitignore 0 → 100644
View file @ f56dcc58
# This file is used to ignore files which are generated
# ----------------------------------------------------------------------------
build/*
depend/*
.vscode/*
core.*
# qtcreator generated files
*.pro.user*
# VS
*.sdf
*.opensdf
*.ilk
*.pdb
*.exp
# Compiled Object files
*.slo
*.lo
*.o
*.obj
# Precompiled Headers
*.gch
*.pch
# Compiled Dynamic libraries
*.so
*.dylib
# Fortran module files
*.mod
*.smod
# Compiled Static libraries
*.lai
*.la
*.a
# Executables
*.exe
*.out
*.app
# xemacs temporary files
*.flc
# Vim temporary files
.*.swp
# others
*.avi
*.pyc
*.egg
3rdParty/InstallOpenCVDependences.sh 0 → 100644
View file @ f56dcc58
#! /bin/sh
############### Ubuntu ###############
# 参考：https://docs.opencv.org/3.4.11/d7/d9f/tutorial_linux_install.html
# apt-get install build-essential -y
# apt-get install cmake git libgtk2.0-dev pkg-config libavcodec-dev libavformat-dev libswscale-dev -y
# apt-get install python-dev python-numpy libtbb2 libtbb-dev libjpeg-dev libpng-dev libtiff-dev libjasper-dev libdc1394-22-dev -y # 处理图像所需的包，可选
############### CentOS ###############
yum install gcc gcc-c++ gtk2-devel gimp-devel gimp-devel-tools gimp-help-browser zlib-devel libtiff-devel libjpeg-devel libpng-devel gstreamer-devel libavc1394-devel libraw1394-devel libdc1394-devel jasper-devel jasper-utils swig python libtool nasm -y
\ No newline at end of file
3rdParty/InstallRBuild.sh 0 → 100644
View file @ f56dcc58
############################ 在线安装依赖 ###############################
#cd ./3rdParty
#pip install rbuild-master.tar.gz
############################ 离线安装依赖 ###############################
# 安装依赖
cd ./3rdParty/rbuild_depend
pip install click-6.6-py2.py3-none-any.whl
pip install six-1.15.0-py2.py3-none-any.whl
pip install subprocess32-3.5.4.tar.gz
pip install cget-0.1.9.tar.gz
# 安装rbuild
cd ../
pip install rbuild-master.tar.gz
CMakeLists.txt 0 → 100644
View file @ f56dcc58
# 设置cmake的最低版本
cmake_minimum_required(VERSION 3.5)
# 设置项目名
project(MIGraphX_Samples)
# 设置编译器
set(CMAKE_CXX_COMPILER g++)
set(CMAKE_CXX_FLAGS ${CMAKE_CXX_FLAGS} -std=c++17) # 2.2版本以上需要c++17
set(CMAKE_BUILD_TYPE release)
# 添加头文件路径
set(INCLUDE_PATH ${CMAKE_CURRENT_SOURCE_DIR}/Src/
${CMAKE_CURRENT_SOURCE_DIR}/Src/Utility/
${CMAKE_CURRENT_SOURCE_DIR}/Src/OCR/Crnn/
$ENV{DTKROOT}/include/
${CMAKE_CURRENT_SOURCE_DIR}/depend/include/)
include_directories(${INCLUDE_PATH})
# 添加依赖库路径
set(LIBRARY_PATH ${CMAKE_CURRENT_SOURCE_DIR}/depend/lib64/
$ENV{DTKROOT}/lib/)
link_directories(${LIBRARY_PATH})
# 添加依赖库
set(LIBRARY opencv_core
opencv_imgproc
opencv_imgcodecs
opencv_dnn
migraphx_ref
migraphx
migraphx_c
migraphx_device
migraphx_gpu
migraphx_onnx)
link_libraries(${LIBRARY})
# 添加源文件
set(SOURCE_FILES ${CMAKE_CURRENT_SOURCE_DIR}/Src/main.cpp
${CMAKE_CURRENT_SOURCE_DIR}/Src/Sample.cpp
${CMAKE_CURRENT_SOURCE_DIR}/Src/OCR/Crnn/Crnn.cpp
${CMAKE_CURRENT_SOURCE_DIR}/Src/Utility/CommonUtility.cpp
${CMAKE_CURRENT_SOURCE_DIR}/Src/Utility/Filesystem.cpp)
# 添加可执行目标
add_executable(MIGraphX_Samples ${SOURCE_FILES})
Crnn_infer_migraphx.py → Python/Crnn_infer_migraphx.py
View file @ f56dcc58
......@@ -66,8 +66,8 @@ class Crnn:
if __name__ == '__main__':
parser = argparse.ArgumentParser()
parser.add_argument('--imgpath', type=str, default='./data/text.jpg', help="image path")
parser.add_argument('--modelpath', type=str, default='./weights/crnn.onnx', help="onnx filepath")
parser.add_argument('--imgpath', type=str, default='../Resource/Images/text.jpg', help="image path")
parser.add_argument('--modelpath', type=str, default='../Resource/Models/Ocr/CRNN/crnn.onnx', help="onnx filepath")
args = parser.parse_args()
crnn = Crnn(args.modelpath)
......
Python/README.md 0 → 100644
View file @ f56dcc58
# CRNN
## 模型介绍
CRNN是文本识别领域的一种经典算法，该算法的主要思想是认为文本识别需要对序列进行预测，所以采用了预测序列常用的RNN网络。算法通过CNN提取图片特征，然后采用RNN对序列进行预测，最终使用CTC方法得到最终结果。
## 模型结构
CRNN模型的主要结构包括基于CNN的图像特征提取模块以及基于双向LSTM的文字序列特征提取模块。
## 构建安装
在光源可拉取推理的docker镜像，CRNN 模型推理推荐的镜像如下：
```python
docker pull image.sourcefind.cn:5000/dcu/admin/base/custom:ort1.14.0_migraphx3.0.0-dtk22.10.1
```
### 安装Opencv依赖
```python
cd <path_to_migraphx_samples>
sh ./3rdParty/InstallOpenCVDependences.sh
```
### 修改CMakeLists.txt
- 如果使用ubuntu系统，需要修改CMakeLists.txt中依赖库路径：
将"${CMAKE_CURRENT_SOURCE_DIR}/depend/lib64/"修改为"${CMAKE_CURRENT_SOURCE_DIR}/depend/lib/"
- **MIGraphX2.3.0及以上版本需要c++17**
### 安装OpenCV并构建工程
```
rbuild build -d depend
```
### 设置环境变量
将依赖库依赖加入环境变量LD_LIBRARY_PATH，在~/.bashrc中添加如下语句：
**Centos**:
```
export LD_LIBRARY_PATH=<path_to_migraphx_samples>/depend/lib64/:$LD_LIBRARY_PATH
```
**Ubuntu**:
```
export LD_LIBRARY_PATH=<path_to_migraphx_samples>/depend/lib/:$LD_LIBRARY_PATH
```
然后执行:
```
source ~/.bashrc
```
## 推理
### C++版本推理
成功编译CRNN工程后，在build目录下输入如下命令运行该示例：
1. 静态推理
```
./MIGraphX_Samples a
```
静态推理示例输入样本图像为：
![text0](../Resource/Images/CrnnDynamicPic/text0.jpg)
文本识别结果为：
```
a-----v--a-i-l-a-bb-l-e--- => available
```
2. 动态shape推理
```
# 开启环境变量
export MIGRAPHX_DYNAMIC_SHAPE=1
# 执行动态推理
./MIGraphX_Samples b
```
文本识别结果：
```
available
migraphx
waiting
recognition
```
### python版本推理
CRNN工程的Python推理示例仅提供静态推理，可使用如下命令运行：
```
# 进入python示例目录
cd ./Python
# 安装依赖
pip install -r requirements.txt
# 运行程序
python Crnn_infer_migraphx.py
```
该python示例输入样本图像与C++在静态推理中一致，文本识别结果为：
```
a-----v--a-i-l-a-bb-l-e--- => available
```
## 历史版本
​ https://developer.hpccube.com/codes/modelzoo/crnn_migraphx
## 参考
​ https://github.com/meijieru/crnn.pytorch
Python/requirements.txt 0 → 100644
View file @ f56dcc58
opencv-contrib-python
numpy
argparse
time
\ No newline at end of file
README.md
View file @ f56dcc58
# CRNN
# Convolutional Recurrent Neural Network(CRNN)
## 模型介绍
......@@ -8,33 +8,72 @@ CRNN是文本识别领域的一种经典算法，该算法的主要思想是认
CRNN模型的主要结构包括基于CNN的图像特征提取模块以及基于双向LSTM的文字序列特征提取模块。
## 推理
## 构建安装
在光源可拉取推理的docker镜像，CRNN 模型推理推荐的镜像如下：
```python
docker pull image.sourcefind.cn:5000/dcu/admin/base/custom:ort1.14.0_migraphx3.0.0-dtk22.10.1
```
### 安装Opencv依赖
```python
cd <path_to_migraphx_samples>
sh ./3rdParty/InstallOpenCVDependences.sh
```
### 修改CMakeLists.txt
- 如果使用ubuntu系统，需要修改CMakeLists.txt中依赖库路径：
将"${CMAKE_CURRENT_SOURCE_DIR}/depend/lib64/"修改为"${CMAKE_CURRENT_SOURCE_DIR}/depend/lib/"
### 环境配置
- **MIGraphX2.3.0及以上版本需要c++17**
在[光源](https://www.sourcefind.cn/#/image/dcu/custom)可拉取用于推理的docker镜像，CRNN 模型推理推荐的镜像如下：
### 安装OpenCV并构建工程
```
docker pull image.sourcefind.cn:5000/dcu/admin/base/custom:ort_dcu_1.14.0_migraphx2.5.2_dtk22.10.1
rbuild build -d depend
```
在[光合开发者社区](https://cancon.hpccube.com:65024/4/main/)可下载MIGraphX安装包，python依赖安装：
### 设置环境变量
将依赖库依赖加入环境变量LD_LIBRARY_PATH，在~/.bashrc中添加如下语句：
**Centos**:
```
pip install -r requirements.txt
export LD_LIBRARY_PATH=<path_to_migraphx_samples>/depend/lib64/:$LD_LIBRARY_PATH
```
### 运行示例
**Ubuntu**:
```
export LD_LIBRARY_PATH=<path_to_migraphx_samples>/depend/lib/:$LD_LIBRARY_PATH
```
CRNN模型的推理示例程序是Crnn_infer_migraphx.py，使用如下命令运行该推理示例：
然后执行:
```
python Crnn_infer_migraphx.py
source ~/.bashrc
```
## 推理
### C++版本推理
成功编译CRNN工程后，在build目录下输入如下命令运行该示例：
1. 静态推理
```
./MIGraphX_Samples a
```
该示例输入样本图像为：
静态推理示例输入样本图像为：
<img src="./data/text.jpg" alt="Result" />
![text0](./Resource/Images/CrnnDynamicPic/text0.jpg)
文本识别结果为：
......@@ -42,6 +81,46 @@ python Crnn_infer_migraphx.py
a-----v--a-i-l-a-bb-l-e--- => available
```
2. 动态shape推理
```
# 开启环境变量
export MIGRAPHX_DYNAMIC_SHAPE=1
# 执行动态推理
./MIGraphX_Samples b
```
文本识别结果：
```
available
migraphx
waiting
recognition
```
### python版本推理
CRNN工程的Python推理示例仅提供静态推理，可使用如下命令运行：
```
# 进入python示例目录
cd ./Python
# 安装依赖
pip install -r requirements.txt
# 运行程序
python Crnn_infer_migraphx.py
```
该python示例输入样本图像与C++在静态推理中一致，文本识别结果为：
```
a-----v--a-i-l-a-bb-l-e--- => available
```
## 历史版本
​ https://developer.hpccube.com/codes/modelzoo/crnn_migraphx
......
Resource/Configuration.xml 0 → 100644
View file @ f56dcc58
<?xml version="1.0" encoding="GB2312"?>
<opencv_storage>
<!--分类器-->
<Classifier>
<ModelPath>"../Resource/Models/Classifier/mnist-12.onnx"</ModelPath>
<Scale>0.003922</Scale><!--缩放尺度-->
<MeanValue1>0.0</MeanValue1><!--均值-->
<MeanValue2>0.0</MeanValue2>
<MeanValue3>0.0</MeanValue3>
<SwapRB>0</SwapRB>
<Crop>0</Crop>
<UseInt8>0</UseInt8><!--是否使用int8,不支持-->
<UseFP16>0</UseFP16><!--是否使用FP16-->
<AddSoftmax>1</AddSoftmax><!--是否需要添加Softmax计算(如果onnx模型中包含了softmax，则设置为0)-->
</Classifier>
<!--超分辨率重建-->
<Espcn>
<ModelPath>"../Resource/Models/Super_Resolution/super.onnx"</ModelPath>
</Espcn>
<!--Unet-->
<Unet>
<ModelPath>"../Resource/Models/Segmentation/unet_13_256.onnx"</ModelPath>
</Unet>
<!--Bert-->
<Bert>
<ModelPath>"../Resource/Models/NLP/Bert/bertsquad-10.onnx"</ModelPath>
</Bert>
<!--GPT2-->
<GPT2>
<ModelPath>"../Resource/Models/NLP/GPT2/GPT2_shici.onnx"</ModelPath>
</GPT2>
<!--SSD检测器-->
<DetectorSSD>
<ModelPath>"../Resource/Models/Detector/SSD/yufacedetectnet-open-v2.onnx"</ModelPath>
<Scale>1.0</Scale><!--缩放尺度-->
<MeanValue1>0</MeanValue1><!--均值，顺序为bgr-->
<MeanValue2>0</MeanValue2>
<MeanValue3>0</MeanValue3>
<SwapRB>0</SwapRB>
<Crop>0</Crop>
<UseInt8>0</UseInt8><!--是否使用int8,不支持-->
<UseFP16>0</UseFP16><!--是否使用FP16-->
<!--////////////////// SSD网络结构参数 ////////////////// -->
<!--priorbox层的个数-->
<PriorBoxLayerNumber>4</PriorBoxLayerNumber>
<!--每个priorbox层的minisize和maxSize(需要与输出检测层顺序保持一致，下面涉及每个priorbox层参数的都需要保持顺序一致)-->
<MinSize11>10</MinSize11>
<MinSize12>16</MinSize12>
<MinSize13>24</MinSize13>
<MinSize21>32</MinSize21>
<MinSize22>48</MinSize22>
<MinSize31>64</MinSize31>
<MinSize32>96</MinSize32>
<MinSize41>128</MinSize41>
<MinSize42>192</MinSize42>
<MinSize43>256</MinSize43>
<!--每个priorbox层的Flip和Clip(使用0,1表示)-->
<Flip1>0</Flip1>
<Flip2>0</Flip2>
<Flip3>0</Flip3>
<Flip4>0</Flip4>
<Clip1>0</Clip1>
<Clip2>0</Clip2>
<Clip3>0</Clip3>
<Clip4>0</Clip4>
<!--每个priorbox层的宽高比(不包括1,且忽略flip,比如宽高比设置为0.3333和0.25且flip为true,则只需要写0.3333和0.25,如果宽高比只有1,则不用填写该项)-->
<!-- <AspectRatio11>0.3333</AspectRatio11>
<AspectRatio12>0.25</AspectRatio12>
<AspectRatio21>0.3333</AspectRatio21>
<AspectRatio22>0.25</AspectRatio22>
<AspectRatio31>0.3333</AspectRatio31>
<AspectRatio32>0.25</AspectRatio32>
<AspectRatio41>0.3333</AspectRatio41>
<AspectRatio42>0.25</AspectRatio42> -->
<!--每个priorbox层的step-->
<PriorBoxStepWidth1>8</PriorBoxStepWidth1><!--第一个priorbox层的step的width-->
<PriorBoxStepWidth2>16</PriorBoxStepWidth2>
<PriorBoxStepWidth3>32</PriorBoxStepWidth3>
<PriorBoxStepWidth4>64</PriorBoxStepWidth4>
<PriorBoxStepHeight1>8</PriorBoxStepHeight1><!--第一个priorbox层的step的height-->
<PriorBoxStepHeight2>16</PriorBoxStepHeight2>
<PriorBoxStepHeight3>32</PriorBoxStepHeight3>
<PriorBoxStepHeight4>64</PriorBoxStepHeight4>
<!--priorbox层中的offset-->
<Offset>0.5</Offset>
<!--DetectionOutput参数-->
<ClassNumber>2</ClassNumber>
<TopK>400</TopK>
<KeepTopK>200</KeepTopK>
<NMSThreshold>0.3</NMSThreshold>
<ConfidenceThreshold>0.9</ConfidenceThreshold>
</DetectorSSD>
<!--RetinaFace检测器-->
<DetectorRetinaFace>
<ModelPath>"../Resource/Models/Detector/RetinaFace/mobilenet0.25_Final.onnx"</ModelPath>
<Scale>1.0</Scale><!--缩放尺度-->
<MeanValue1>104</MeanValue1><!--均值，顺序为bgr-->
<MeanValue2>117</MeanValue2>
<MeanValue3>123</MeanValue3>
<SwapRB>0</SwapRB>
<Crop>0</Crop>
<UseInt8>0</UseInt8><!--是否使用int8,不支持-->
<UseFP16>0</UseFP16><!--是否使用FP16-->
<!--////////////////// RetinaFace检测器参数 ////////////////// -->
<!--priorbox层的个数-->
<PriorBoxLayerNumber>3</PriorBoxLayerNumber>
<!--每个priorbox层的minisize和maxSize(需要与输出检测层顺序保持一致，下面涉及每个priorbox层参数的都需要保持顺序一致)-->
<MinSize11>16</MinSize11>
<MinSize12>32</MinSize12>
<MinSize21>64</MinSize21>
<MinSize22>128</MinSize22>
<MinSize31>256</MinSize31>
<MinSize32>512</MinSize32>
<!--每个priorbox层的Flip和Clip(使用0,1表示)-->
<Flip1>0</Flip1>
<Flip2>0</Flip2>
<Flip3>0</Flip3>
<Clip1>0</Clip1>
<Clip2>0</Clip2>
<Clip3>0</Clip3>
<!--每个priorbox层的宽高比(由于RetinaFace只包含宽高比为1的anchor，所以这里不需要设置宽高比)-->
<!-- <AspectRatio11>0.3333</AspectRatio11>
<AspectRatio12>0.25</AspectRatio12>
<AspectRatio21>0.3333</AspectRatio21>
<AspectRatio22>0.25</AspectRatio22>
<AspectRatio31>0.3333</AspectRatio31>
<AspectRatio32>0.25</AspectRatio32>
<AspectRatio41>0.3333</AspectRatio41>
<AspectRatio42>0.25</AspectRatio42> -->
<!--每个priorbox层的step-->
<PriorBoxStepWidth1>8</PriorBoxStepWidth1><!--第一个priorbox层的step的width-->
<PriorBoxStepWidth2>16</PriorBoxStepWidth2>
<PriorBoxStepWidth3>32</PriorBoxStepWidth3>
<PriorBoxStepHeight1>8</PriorBoxStepHeight1><!--第一个priorbox层的step的height-->
<PriorBoxStepHeight2>16</PriorBoxStepHeight2>
<PriorBoxStepHeight3>32</PriorBoxStepHeight3>
<!--priorbox层中的offset-->
<Offset>0.5</Offset>
<!--DetectionOutput参数-->
<ClassNumber>2</ClassNumber>
<TopK>400</TopK>
<KeepTopK>200</KeepTopK>
<NMSThreshold>0.3</NMSThreshold>
<ConfidenceThreshold>0.9</ConfidenceThreshold>
</DetectorRetinaFace>
<!--YOLOV3检测器 -->
<DetectorYOLOV3>
<ModelPath>"../Resource/Models/Detector/YOLOV3/yolov3-tiny.onnx"</ModelPath>
<ClassNameFile>"../Resource/Models/Detector/YOLOV3/coco.names"</ClassNameFile>
<UseFP16>0</UseFP16><!--是否使用FP16-->
<NumberOfClasses>80</NumberOfClasses><!--类别数(不包括背景类)，COCO:80,VOC:20-->
<ConfidenceThreshold>0.2</ConfidenceThreshold>
<NMSThreshold>0.4</NMSThreshold>
<ObjectThreshold>0.4</ObjectThreshold>
</DetectorYOLOV3>
<!--YOLOV5检测器 -->
<DetectorYOLOV5>
<ModelPath>"../Resource/Models/Detector/YOLOV5/yolov5s.onnx"</ModelPath>
<ClassNameFile>"../Resource/Models/Detector/YOLOV5/coco.names"</ClassNameFile>
<UseFP16>0</UseFP16><!--是否使用FP16-->
<NumberOfClasses>80</NumberOfClasses><!--类别数(不包括背景类)，COCO:80,VOC:20-->
<ConfidenceThreshold>0.25</ConfidenceThreshold>
<NMSThreshold>0.5</NMSThreshold>
<ObjectThreshold>0.5</ObjectThreshold>
</DetectorYOLOV5>
<!--MTCNN检测器 -->
<DetectorMTCNN>
<PNet>
<ModelPath>"../Resource/Models/Detector/MTCNN/PNet.onnx"</ModelPath>
<MaxHeight>512</MaxHeight>
<MaxWidth>512</MaxWidth>
<ConfidenceThreshold>0.90</ConfidenceThreshold>
<UseFP16>0</UseFP16>
</PNet>
</DetectorMTCNN>
<!--YOLOV7检测器 -->
<DetectorYOLOV7>
<ModelPath>"../Resource/Models/Detector/YOLOV7/yolov7-tiny.onnx"</ModelPath>
<ClassNameFile>"../Resource/Models/Detector/YOLOV7/coco.names"</ClassNameFile>
<UseFP16>0</UseFP16><!--是否使用FP16-->
<NumberOfClasses>80</NumberOfClasses><!--类别数(不包括背景类)，COCO:80,VOC:20-->
<ConfidenceThreshold>0.25</ConfidenceThreshold>
<NMSThreshold>0.5</NMSThreshold>
<ObjectThreshold>0.5</ObjectThreshold>
</DetectorYOLOV7>
<!--CRNN动态文本识别 -->
<CrnnDynamic>
<ModelPath>"../Resource/Models/Ocr/CRNN/crnn_dynamic.onnx"</ModelPath>
</CrnnDynamic>
<!--PaddleOCR车牌检测 -->
<OcrDB>
<ModelPath>"../Resource/Models/PaddleOCR/VLPR/db.onnx"</ModelPath>
<BinaryThreshold>0.3</BinaryThreshold>
<BoxThreshold>0.5</BoxThreshold>
<UnclipRatio>1.6</UnclipRatio>
<LimitSideLen>2500</LimitSideLen>
<ScoreMode>"fast"</ScoreMode>
</OcrDB>
<!--PaddleOCR车牌识别 -->
<OcrSVTR>
<ModelPath>"../Resource/Models/PaddleOCR/VLPR/svtr.onnx"</ModelPath>
<DictPath>"../Resource/Models/PaddleOCR/VLPR/ppocr_keys_v1.txt"</DictPath>
</OcrSVTR>
</opencv_storage>
Markdown is supported
0% or .
You are about to add 0 people to the discussion. Proceed with caution.
Finish editing this message first!
Please register or to comment