Skip to content

P
PaddleOCR

weixin_41840029 / PaddleOCR
与 Fork 源项目一致

Fork自 PaddlePaddle / PaddleOCR

通知 1
Star 1
Fork 0
P
PaddleOCR
提交 56feca01 编写于 作者: qq_25193841's avatar qq_25193841
浏览文件
操作

Merge remote-tracking branch 'PaddlePaddle/dygraph' into dygraph

上级 bfea4a63 c4d20782
隐藏空白更改
内联 并排
Showing with 212 addition and 321 deletion
README_ch.md
浏览文件 @ 56feca01
...@@ -8,7 +8,6 @@ PaddleOCR同时支持动态图与静态图两种编程范式 ...@@ -8,7 +8,6 @@ PaddleOCR同时支持动态图与静态图两种编程范式
- 静态图版本：develop分支 - 静态图版本：develop分支
**近期更新** **近期更新**
- 【预告】 PaddleOCR研发团队对最新发版内容技术深入解读，4月13日晚上19:00，[直播地址](https://live.bilibili.com/21689802)
- 2021.4.8 release 2.1版本，新增AAAI 2021论文[端到端识别算法PGNet](./doc/doc_ch/pgnet.md)开源，[多语言模型](./doc/doc_ch/multi_languages.md)支持种类增加到80+。 - 2021.4.8 release 2.1版本，新增AAAI 2021论文[端到端识别算法PGNet](./doc/doc_ch/pgnet.md)开源，[多语言模型](./doc/doc_ch/multi_languages.md)支持种类增加到80+。
- 2021.2.1 [FAQ](./doc/doc_ch/FAQ.md)新增5个高频问题，总数162个，每周一都会更新，欢迎大家持续关注。 - 2021.2.1 [FAQ](./doc/doc_ch/FAQ.md)新增5个高频问题，总数162个，每周一都会更新，欢迎大家持续关注。
- 2021.1.21 更新多语言识别模型，目前支持语种超过27种，包括中文简体、中文繁体、英文、法文、德文、韩文、日文、意大利文、西班牙文、葡萄牙文、俄罗斯文、阿拉伯文等，后续计划可以参考[多语言研发计划](https://github.com/PaddlePaddle/PaddleOCR/issues/1048) - 2021.1.21 更新多语言识别模型，目前支持语种超过27种，包括中文简体、中文繁体、英文、法文、德文、韩文、日文、意大利文、西班牙文、葡萄牙文、俄罗斯文、阿拉伯文等，后续计划可以参考[多语言研发计划](https://github.com/PaddlePaddle/PaddleOCR/issues/1048)
...@@ -80,7 +79,7 @@ PaddleOCR同时支持动态图与静态图两种编程范式 ...@@ -80,7 +79,7 @@ PaddleOCR同时支持动态图与静态图两种编程范式
- 算法介绍 - 算法介绍
- [文本检测](./doc/doc_ch/algorithm_overview.md) - [文本检测](./doc/doc_ch/algorithm_overview.md)
- [文本识别](./doc/doc_ch/algorithm_overview.md) - [文本识别](./doc/doc_ch/algorithm_overview.md)
- [PP-OCR Pipline](#PP-OCR) - [PP-OCR Pipeline](#PP-OCR)
- [端到端PGNet算法](./doc/doc_ch/pgnet.md) - [端到端PGNet算法](./doc/doc_ch/pgnet.md)
- 模型训练/评估 - 模型训练/评估
- [文本检测](./doc/doc_ch/detection.md) - [文本检测](./doc/doc_ch/detection.md)
...@@ -115,7 +114,7 @@ PaddleOCR同时支持动态图与静态图两种编程范式 ...@@ -115,7 +114,7 @@ PaddleOCR同时支持动态图与静态图两种编程范式
<a name="PP-OCR"></a> <a name="PP-OCR"></a>
## PP-OCR Pipline ## PP-OCR Pipeline
<div align="center"> <div align="center">
<img src="./doc/ppocr_framework.png" width="800"> <img src="./doc/ppocr_framework.png" width="800">
</div> </div>
......
deploy/cpp_infer/include/config.h
浏览文件 @ 56feca01
...@@ -49,6 +49,8 @@ public: ...@@ -49,6 +49,8 @@ public:
this->det_db_unclip_ratio = stod(config_map_["det_db_unclip_ratio"]); this->det_db_unclip_ratio = stod(config_map_["det_db_unclip_ratio"]);
this->use_polygon_score = bool(stoi(config_map_["use_polygon_score"]));
this->det_model_dir.assign(config_map_["det_model_dir"]); this->det_model_dir.assign(config_map_["det_model_dir"]);
this->rec_model_dir.assign(config_map_["rec_model_dir"]); this->rec_model_dir.assign(config_map_["rec_model_dir"]);
...@@ -86,6 +88,8 @@ public: ...@@ -86,6 +88,8 @@ public:
double det_db_unclip_ratio = 2.0; double det_db_unclip_ratio = 2.0;
bool use_polygon_score = false;
std::string det_model_dir; std::string det_model_dir;
std::string rec_model_dir; std::string rec_model_dir;
......
deploy/cpp_infer/include/ocr_det.h
浏览文件 @ 56feca01
...@@ -44,7 +44,8 @@ public: ...@@ -44,7 +44,8 @@ public:
const bool &use_mkldnn, const int &max_side_len, const bool &use_mkldnn, const int &max_side_len,
const double &det_db_thresh, const double &det_db_thresh,
const double &det_db_box_thresh, const double &det_db_box_thresh,
const double &det_db_unclip_ratio, const bool &visualize, const double &det_db_unclip_ratio,
const bool &use_polygon_score, const bool &visualize,
const bool &use_tensorrt, const bool &use_fp16) { const bool &use_tensorrt, const bool &use_fp16) {
this->use_gpu_ = use_gpu; this->use_gpu_ = use_gpu;
this->gpu_id_ = gpu_id; this->gpu_id_ = gpu_id;
...@@ -57,6 +58,7 @@ public: ...@@ -57,6 +58,7 @@ public:
this->det_db_thresh_ = det_db_thresh; this->det_db_thresh_ = det_db_thresh;
this->det_db_box_thresh_ = det_db_box_thresh; this->det_db_box_thresh_ = det_db_box_thresh;
this->det_db_unclip_ratio_ = det_db_unclip_ratio; this->det_db_unclip_ratio_ = det_db_unclip_ratio;
this->use_polygon_score_ = use_polygon_score;
this->visualize_ = visualize; this->visualize_ = visualize;
this->use_tensorrt_ = use_tensorrt; this->use_tensorrt_ = use_tensorrt;
...@@ -85,6 +87,7 @@ private: ...@@ -85,6 +87,7 @@ private:
double det_db_thresh_ = 0.3; double det_db_thresh_ = 0.3;
double det_db_box_thresh_ = 0.5; double det_db_box_thresh_ = 0.5;
double det_db_unclip_ratio_ = 2.0; double det_db_unclip_ratio_ = 2.0;
bool use_polygon_score_ = false;
bool visualize_ = true; bool visualize_ = true;
bool use_tensorrt_ = false; bool use_tensorrt_ = false;
......
deploy/cpp_infer/include/postprocess_op.h
浏览文件 @ 56feca01
...@@ -55,7 +55,8 @@ public: ...@@ -55,7 +55,8 @@ public:
std::vector<std::vector<std::vector<int>>> std::vector<std::vector<std::vector<int>>>
BoxesFromBitmap(const cv::Mat pred, const cv::Mat bitmap, BoxesFromBitmap(const cv::Mat pred, const cv::Mat bitmap,
const float &box_thresh, const float &det_db_unclip_ratio); const float &box_thresh, const float &det_db_unclip_ratio,
const bool &use_polygon_score);
std::vector<std::vector<std::vector<int>>> std::vector<std::vector<std::vector<int>>>
FilterTagDetRes(std::vector<std::vector<std::vector<int>>> boxes, FilterTagDetRes(std::vector<std::vector<std::vector<int>>> boxes,
......
deploy/cpp_infer/readme.md
浏览文件 @ 56feca01
...@@ -183,7 +183,7 @@ cmake .. \ ...@@ -183,7 +183,7 @@ cmake .. \
make -j make -j
``` ```
`OPENCV_DIR`为opencv编译安装的地址；`LIB_DIR`为下载(`paddle_inference`文件夹)或者编译生成的Paddle预测库地址(`build/paddle_inference_install_dir`文件夹)；`CUDA_LIB_DIR`为cuda库文件地址，在docker中；为`/usr/local/cuda/lib64`；`CUDNN_LIB_DIR`为cudnn库文件地址，在docker中为`/usr/lib/x86_64-linux-gnu/`。 `OPENCV_DIR`为opencv编译安装的地址；`LIB_DIR`为下载(`paddle_inference`文件夹)或者编译生成的Paddle预测库地址(`build/paddle_inference_install_dir`文件夹)；`CUDA_LIB_DIR`为cuda库文件地址，在docker中为`/usr/local/cuda/lib64`；`CUDNN_LIB_DIR`为cudnn库文件地址，在docker中为`/usr/lib/x86_64-linux-gnu/`。
* 编译完成之后，会在`build`文件夹下生成一个名为`ocr_system`的可执行文件。 * 编译完成之后，会在`build`文件夹下生成一个名为`ocr_system`的可执行文件。
...@@ -211,6 +211,7 @@ max_side_len 960 # 输入图像长宽大于960时，等比例缩放图像，使 ...@@ -211,6 +211,7 @@ max_side_len 960 # 输入图像长宽大于960时，等比例缩放图像，使
det_db_thresh 0.3 # 用于过滤DB预测的二值化图像，设置为0.-0.3对结果影响不明显 det_db_thresh 0.3 # 用于过滤DB预测的二值化图像，设置为0.-0.3对结果影响不明显
det_db_box_thresh 0.5 # DB后处理过滤box的阈值，如果检测存在漏框情况，可酌情减小 det_db_box_thresh 0.5 # DB后处理过滤box的阈值，如果检测存在漏框情况，可酌情减小
det_db_unclip_ratio 1.6 # 表示文本框的紧致程度，越小则文本框更靠近文本 det_db_unclip_ratio 1.6 # 表示文本框的紧致程度，越小则文本框更靠近文本
use_polygon_score 1 # 是否使用多边形框计算bbox score，0表示使用矩形框计算。矩形框计算速度更快，多边形框对弯曲文本区域计算更准确。
det_model_dir ./inference/det_db # 检测模型inference model地址 det_model_dir ./inference/det_db # 检测模型inference model地址
# cls config # cls config
......
deploy/cpp_infer/readme_en.md
浏览文件 @ 56feca01
...@@ -217,6 +217,7 @@ max_side_len 960 # Limit the maximum image height and width to 960 ...@@ -217,6 +217,7 @@ max_side_len 960 # Limit the maximum image height and width to 960
det_db_thresh 0.3 # Used to filter the binarized image of DB prediction, setting 0.-0.3 has no obvious effect on the result det_db_thresh 0.3 # Used to filter the binarized image of DB prediction, setting 0.-0.3 has no obvious effect on the result
det_db_box_thresh 0.5 # DDB post-processing filter box threshold, if there is a missing box detected, it can be reduced as appropriate det_db_box_thresh 0.5 # DDB post-processing filter box threshold, if there is a missing box detected, it can be reduced as appropriate
det_db_unclip_ratio 1.6 # Indicates the compactness of the text box, the smaller the value, the closer the text box to the text det_db_unclip_ratio 1.6 # Indicates the compactness of the text box, the smaller the value, the closer the text box to the text
use_polygon_score 1 # Whether to use polygon box to calculate bbox score, 0 means to use rectangle box to calculate. Use rectangular box to calculate faster, and polygonal box more accurate for curved text area.
det_model_dir ./inference/det_db # Address of detection inference model det_model_dir ./inference/det_db # Address of detection inference model
# cls config # cls config
......
deploy/cpp_infer/src/main.cpp
浏览文件 @ 56feca01
...@@ -59,7 +59,8 @@ int main(int argc, char **argv) { ...@@ -59,7 +59,8 @@ int main(int argc, char **argv) {
config.gpu_mem, config.cpu_math_library_num_threads, config.gpu_mem, config.cpu_math_library_num_threads,
config.use_mkldnn, config.max_side_len, config.det_db_thresh, config.use_mkldnn, config.max_side_len, config.det_db_thresh,
config.det_db_box_thresh, config.det_db_unclip_ratio, config.det_db_box_thresh, config.det_db_unclip_ratio,
config.visualize, config.use_tensorrt, config.use_fp16); config.use_polygon_score, config.visualize,
config.use_tensorrt, config.use_fp16);
Classifier *cls = nullptr; Classifier *cls = nullptr;
if (config.use_angle_cls == true) { if (config.use_angle_cls == true) {
......
deploy/cpp_infer/src/ocr_det.cpp
浏览文件 @ 56feca01
...@@ -109,9 +109,9 @@ void DBDetector::Run(cv::Mat &img, ...@@ -109,9 +109,9 @@ void DBDetector::Run(cv::Mat &img,
cv::Mat dilation_map; cv::Mat dilation_map;
cv::Mat dila_ele = cv::getStructuringElement(cv::MORPH_RECT, cv::Size(2, 2)); cv::Mat dila_ele = cv::getStructuringElement(cv::MORPH_RECT, cv::Size(2, 2));
cv::dilate(bit_map, dilation_map, dila_ele); cv::dilate(bit_map, dilation_map, dila_ele);
boxes = post_processor_.BoxesFromBitmap(pred_map, dilation_map, boxes = post_processor_.BoxesFromBitmap(
this->det_db_box_thresh_, pred_map, dilation_map, this->det_db_box_thresh_,
this->det_db_unclip_ratio_); this->det_db_unclip_ratio_, this->use_polygon_score_);
boxes = post_processor_.FilterTagDetRes(boxes, ratio_h, ratio_w, srcimg); boxes = post_processor_.FilterTagDetRes(boxes, ratio_h, ratio_w, srcimg);
......
deploy/cpp_infer/src/postprocess_op.cpp
浏览文件 @ 56feca01
...@@ -160,35 +160,49 @@ std::vector<std::vector<float>> PostProcessor::GetMiniBoxes(cv::RotatedRect box, ...@@ -160,35 +160,49 @@ std::vector<std::vector<float>> PostProcessor::GetMiniBoxes(cv::RotatedRect box,
} }
float PostProcessor::PolygonScoreAcc(std::vector<cv::Point> contour, float PostProcessor::PolygonScoreAcc(std::vector<cv::Point> contour,
cv::Mat pred){ cv::Mat pred) {
int width = pred.cols; int width = pred.cols;
int height = pred.rows; int height = pred.rows;
std::vector<float> box_x; std::vector<float> box_x;
std::vector<float> box_y; std::vector<float> box_y;
for(int i=0; i<contour.size(); ++i){ for (int i = 0; i < contour.size(); ++i) {
box_x.push_back(contour[i].x); box_x.push_back(contour[i].x);
box_y.push_back(contour[i].y); box_y.push_back(contour[i].y);
} }
int xmin = clamp(int(std::floor(*(std::min_element(box_x.begin(), box_x.end())))), 0, width - 1); int xmin =
int xmax = clamp(int(std::ceil(*(std::max_element(box_x.begin(), box_x.end())))), 0, width - 1); clamp(int(std::floor(*(std::min_element(box_x.begin(), box_x.end())))), 0,
int ymin = clamp(int(std::floor(*(std::min_element(box_y.begin(), box_y.end())))), 0, height - 1); width - 1);
int ymax = clamp(int(std::ceil(*(std::max_element(box_y.begin(), box_y.end())))), 0, height - 1); int xmax =
clamp(int(std::ceil(*(std::max_element(box_x.begin(), box_x.end())))), 0,
width - 1);
int ymin =
clamp(int(std::floor(*(std::min_element(box_y.begin(), box_y.end())))), 0,
height - 1);
int ymax =
clamp(int(std::ceil(*(std::max_element(box_y.begin(), box_y.end())))), 0,
height - 1);
cv::Mat mask; cv::Mat mask;
mask = cv::Mat::zeros(ymax - ymin + 1, xmax - xmin + 1, CV_8UC1); mask = cv::Mat::zeros(ymax - ymin + 1, xmax - xmin + 1, CV_8UC1);
cv::Point rook_point[contour.size()];
for(int i=0; i<contour.size(); ++i){ cv::Point* rook_point = new cv::Point[contour.size()];
for (int i = 0; i < contour.size(); ++i) {
rook_point[i] = cv::Point(int(box_x[i]) - xmin, int(box_y[i]) - ymin); rook_point[i] = cv::Point(int(box_x[i]) - xmin, int(box_y[i]) - ymin);
} }
const cv::Point *ppt[1] = {rook_point}; const cv::Point *ppt[1] = {rook_point};
int npt[] = {int(contour.size())}; int npt[] = {int(contour.size())};
cv::fillPoly(mask, ppt, npt, 1, cv::Scalar(1)); cv::fillPoly(mask, ppt, npt, 1, cv::Scalar(1));
cv::Mat croppedImg; cv::Mat croppedImg;
pred(cv::Rect(xmin, ymin, xmax - xmin + 1, ymax - ymin + 1)).copyTo(croppedImg); pred(cv::Rect(xmin, ymin, xmax - xmin + 1, ymax - ymin + 1)).copyTo(croppedImg);
float score = cv::mean(croppedImg, mask)[0]; float score = cv::mean(croppedImg, mask)[0];
delete []rook_point;
return score; return score;
} }
...@@ -230,10 +244,9 @@ float PostProcessor::BoxScoreFast(std::vector<std::vector<float>> box_array, ...@@ -230,10 +244,9 @@ float PostProcessor::BoxScoreFast(std::vector<std::vector<float>> box_array,
return score; return score;
} }
std::vector<std::vector<std::vector<int>>> std::vector<std::vector<std::vector<int>>> PostProcessor::BoxesFromBitmap(
PostProcessor::BoxesFromBitmap(const cv::Mat pred, const cv::Mat bitmap, const cv::Mat pred, const cv::Mat bitmap, const float &box_thresh,
const float &box_thresh, const float &det_db_unclip_ratio, const bool &use_polygon_score) {
const float &det_db_unclip_ratio) {
const int min_size = 3; const int min_size = 3;
const int max_candidates = 1000; const int max_candidates = 1000;
...@@ -267,9 +280,12 @@ PostProcessor::BoxesFromBitmap(const cv::Mat pred, const cv::Mat bitmap, ...@@ -267,9 +280,12 @@ PostProcessor::BoxesFromBitmap(const cv::Mat pred, const cv::Mat bitmap,
} }
float score; float score;
score = BoxScoreFast(array, pred); if (use_polygon_score)
/* compute using polygon*/ /* compute using polygon*/
// score = PolygonScoreAcc(contours[_i], pred); score = PolygonScoreAcc(contours[_i], pred);
else
score = BoxScoreFast(array, pred);
if (score < box_thresh) if (score < box_thresh)
continue; continue;
......
deploy/cpp_infer/tools/config.txt
浏览文件 @ 56feca01
...@@ -10,6 +10,7 @@ max_side_len 960 ...@@ -10,6 +10,7 @@ max_side_len 960
det_db_thresh 0.3 det_db_thresh 0.3
det_db_box_thresh 0.5 det_db_box_thresh 0.5
det_db_unclip_ratio 1.6 det_db_unclip_ratio 1.6
use_polygon_score 1
det_model_dir ./inference/ch_ppocr_mobile_v2.0_det_infer/ det_model_dir ./inference/ch_ppocr_mobile_v2.0_det_infer/
# cls config # cls config
......
doc/doc_ch/multi_languages.md
浏览文件 @ 56feca01
...@@ -11,7 +11,7 @@ PaddleOCR 旨在打造一套丰富、领先、且实用的OCR工具库，不仅 ...@@ -11,7 +11,7 @@ PaddleOCR 旨在打造一套丰富、领先、且实用的OCR工具库，不仅
其中英文模型支持，大小写字母和常见标点的检测识别，并优化了空格字符的识别： 其中英文模型支持，大小写字母和常见标点的检测识别，并优化了空格字符的识别：
<div align="center"> <div align="center">
<img src="../imgs_results/multi_lang/en_1.jpg" width="400" height="600"> <img src="../imgs_results/multi_lang/img_12.jpg" width="900" height="300">
</div> </div>
小语种模型覆盖了拉丁语系、阿拉伯语系、中文繁体、韩语、日语等等： 小语种模型覆盖了拉丁语系、阿拉伯语系、中文繁体、韩语、日语等等：
...@@ -19,6 +19,8 @@ PaddleOCR 旨在打造一套丰富、领先、且实用的OCR工具库，不仅 ...@@ -19,6 +19,8 @@ PaddleOCR 旨在打造一套丰富、领先、且实用的OCR工具库，不仅
<div align="center"> <div align="center">
<img src="../imgs_results/multi_lang/japan_2.jpg" width="600" height="300"> <img src="../imgs_results/multi_lang/japan_2.jpg" width="600" height="300">
<img src="../imgs_results/multi_lang/french_0.jpg" width="300" height="300"> <img src="../imgs_results/multi_lang/french_0.jpg" width="300" height="300">
<img src="../imgs_results/multi_lang/korean_0.jpg" width="500" height="300">
<img src="../imgs_results/multi_lang/arabic_0.jpg" width="300" height="300">
</div> </div>
...@@ -30,14 +32,9 @@ PaddleOCR 旨在打造一套丰富、领先、且实用的OCR工具库，不仅 ...@@ -30,14 +32,9 @@ PaddleOCR 旨在打造一套丰富、领先、且实用的OCR工具库，不仅
- [2 快速使用](#快速使用) - [2 快速使用](#快速使用)
- [2.1 命令行运行](#命令行运行) - [2.1 命令行运行](#命令行运行)
- [2.1.1 整图预测](#bash_检测+识别)
- [2.1.2 识别预测](#bash_识别)
- [2.1.3 检测预测](#bash_检测)
- [2.2 python 脚本运行](#python_脚本运行) - [2.2 python 脚本运行](#python_脚本运行)
- [2.2.1 整图预测](#python_检测+识别)
- [2.2.2 识别预测](#python_识别)
- [2.2.3 检测预测](#python_检测)
- [3 自定义训练](#自定义训练) - [3 自定义训练](#自定义训练)
- [4 预测部署](#预测部署)
- [4 支持语种及缩写](#语种缩写) - [4 支持语种及缩写](#语种缩写)
<a name="安装"></a> <a name="安装"></a>
...@@ -50,7 +47,7 @@ PaddleOCR 旨在打造一套丰富、领先、且实用的OCR工具库，不仅 ...@@ -50,7 +47,7 @@ PaddleOCR 旨在打造一套丰富、领先、且实用的OCR工具库，不仅
pip install paddlepaddle pip install paddlepaddle
# gpu # gpu
pip instll paddlepaddle-gpu pip install paddlepaddle-gpu
``` ```
<a name="paddleocr_package_安装"></a> <a name="paddleocr_package_安装"></a>
...@@ -108,8 +105,6 @@ paddleocr --image_dir doc/imgs/japan_2.jpg --lang=japan ...@@ -108,8 +105,6 @@ paddleocr --image_dir doc/imgs/japan_2.jpg --lang=japan
paddleocr --image_dir doc/imgs_words/japan/1.jpg --det false --lang=japan paddleocr --image_dir doc/imgs_words/japan/1.jpg --det false --lang=japan
``` ```
![](https://raw.githubusercontent.com/PaddlePaddle/PaddleOCR/release/2.0/doc/imgs_words/japan/1.jpg)
结果是一个tuple，返回识别结果和识别置信度 结果是一个tuple，返回识别结果和识别置信度
```text ```text
...@@ -145,6 +140,9 @@ from paddleocr import PaddleOCR, draw_ocr ...@@ -145,6 +140,9 @@ from paddleocr import PaddleOCR, draw_ocr
ocr = PaddleOCR(lang="korean") # 首次执行会自动下载模型文件 ocr = PaddleOCR(lang="korean") # 首次执行会自动下载模型文件
img_path = 'doc/imgs/korean_1.jpg ' img_path = 'doc/imgs/korean_1.jpg '
result = ocr.ocr(img_path) result = ocr.ocr(img_path)
# 可通过参数控制单独执行识别、检测
# result = ocr.ocr(img_path, det=False) 只执行识别
# result = ocr.ocr(img_path, rec=False) 只执行检测
# 打印检测框和识别结果 # 打印检测框和识别结果
for line in result: for line in result:
print(line) print(line)
...@@ -166,59 +164,7 @@ im_show.save('result.jpg') ...@@ -166,59 +164,7 @@ im_show.save('result.jpg')
<img src="https://raw.githubusercontent.com/PaddlePaddle/PaddleOCR/release/2.1/doc/imgs_results/korean.jpg" width="800"> <img src="https://raw.githubusercontent.com/PaddlePaddle/PaddleOCR/release/2.1/doc/imgs_results/korean.jpg" width="800">
</div> </div>
* 识别预测 ppocr 还支持方向分类， 更多使用方式请参考：[whl包使用说明](https://github.com/PaddlePaddle/PaddleOCR/blob/release/2.0/doc/doc_ch/whl.md)
```
from paddleocr import PaddleOCR
ocr = PaddleOCR(lang="german")
img_path = 'PaddleOCR/doc/imgs_words/german/1.jpg'
result = ocr.ocr(img_path, det=False, cls=True)
for line in result:
print(line)
```
![](../imgs_words/german/1.jpg)
结果是一个tuple，只包含识别结果和识别置信度
```
('leider auch jetzt', 0.97538936)
```
* 检测预测
```python
from paddleocr import PaddleOCR, draw_ocr
ocr = PaddleOCR() # need to run only once to download and load model into memory
img_path = 'PaddleOCR/doc/imgs_en/img_12.jpg'
result = ocr.ocr(img_path, rec=False)
for line in result:
print(line)
# 显示结果
from PIL import Image
image = Image.open(img_path).convert('RGB')
im_show = draw_ocr(image, result, txts=None, scores=None, font_path='/path/to/PaddleOCR/doc/fonts/simfang.ttf')
im_show = Image.fromarray(im_show)
im_show.save('result.jpg')
```
结果是一个list，每个item只包含文本框
```bash
[[26.0, 457.0], [137.0, 457.0], [137.0, 477.0], [26.0, 477.0]]
[[25.0, 425.0], [372.0, 425.0], [372.0, 448.0], [25.0, 448.0]]
[[128.0, 397.0], [273.0, 397.0], [273.0, 414.0], [128.0, 414.0]]
......
```
结果可视化 ：
<div align="center">
<img src="https://raw.githubusercontent.com/PaddlePaddle/PaddleOCR/release/2.1/doc/imgs_results/whl/12_det.jpg" width="800">
</div>
ppocr 还支持方向分类， 更多使用方式请参考：[whl包使用说明](https://github.com/PaddlePaddle/PaddleOCR/blob/release/2.0/doc/doc_ch/whl.md)。
<a name="自定义训练"></a> <a name="自定义训练"></a>
## 3 自定义训练 ## 3 自定义训练
...@@ -229,84 +175,58 @@ ppocr 支持使用自己的数据进行自定义训练或finetune, 其中识别 ...@@ -229,84 +175,58 @@ ppocr 支持使用自己的数据进行自定义训练或finetune, 其中识别
具体数据准备、训练过程可参考：[文本检测](../doc_ch/detection.md)、[文本识别](../doc_ch/recognition.md)，更多功能如预测部署、 具体数据准备、训练过程可参考：[文本检测](../doc_ch/detection.md)、[文本识别](../doc_ch/recognition.md)，更多功能如预测部署、
数据标注等功能可以阅读完整的[文档教程](../../README_ch.md)。 数据标注等功能可以阅读完整的[文档教程](../../README_ch.md)。
<a name="预测部署"></a>
## 4 预测部署
除了安装whl包进行快速预测，ppocr 也提供了多种预测部署方式，如有需求可阅读相关文档：
- [基于Python脚本预测引擎推理](./inference.md)
- [基于C++预测引擎推理](../../deploy/cpp_infer/readme.md)
- [服务化部署](../../deploy/hubserving/readme.md)
- [端侧部署](https://github.com/PaddlePaddle/PaddleOCR/blob/develop/deploy/lite/readme.md)
- [Benchmark](./benchmark.md)
<a name="语种缩写"></a> <a name="语种缩写"></a>
## 4 支持语种及缩写 ## 5 支持语种及缩写
| 语种 | 描述 | 缩写 | | 语种 | 描述 | 缩写 | | 语种 | 描述 | 缩写 |
| --- | --- | --- | | --- | --- | --- | ---|--- | --- | --- |
|中文|chinese and english|ch| |中文|chinese and english|ch| |保加利亚文|Bulgarian |bg|
|英文|english|en| |英文|english|en| |乌克兰文|Ukranian|uk|
|法文|french|fr| |法文|french|fr| |白俄罗斯文|Belarusian|be|
|德文|german|german| |德文|german|german| |泰卢固文|Telugu |te|
|日文|japan|japan| |日文|japan|japan| | |阿巴扎文|Abaza |abq|
|韩文|korean|korean| |韩文|korean|korean| |泰米尔文|Tamil |ta|
|中文繁体|chinese traditional |chinese_cht| |中文繁体|chinese traditional |ch_tra| |南非荷兰文 |Afrikaans |af|
|意大利文| Italian |it| |意大利文| Italian |it| |阿塞拜疆文 |Azerbaijani |az|
|西班牙文|Spanish |es| |西班牙文|Spanish |es| |波斯尼亚文|Bosnian|bs|
|葡萄牙文| Portuguese|pt| |葡萄牙文| Portuguese|pt| |捷克文|Czech|cs|
|俄罗斯文|Russia|ru| |俄罗斯文|Russia|ru| |威尔士文 |Welsh |cy|
|阿拉伯文|Arabic|ar| |阿拉伯文|Arabic|ar| |丹麦文 |Danish|da|
|印地文|Hindi|hi| |印地文|Hindi|hi| |爱沙尼亚文 |Estonian |et|
|维吾尔|Uyghur|ug| |维吾尔|Uyghur|ug| |爱尔兰文 |Irish |ga|
|波斯文|Persian|fa| |波斯文|Persian|fa| |克罗地亚文|Croatian |hr|
|乌尔都文|Urdu|ur| |乌尔都文|Urdu|ur| |匈牙利文|Hungarian |hu|
|塞尔维亚文（latin)| Serbian(latin) |rs_latin| |塞尔维亚文（latin)| Serbian(latin) |rs_latin| |印尼文|Indonesian|id|
|欧西坦文|Occitan |oc| |欧西坦文|Occitan |oc| |冰岛文 |Icelandic|is|
|马拉地文|Marathi|mr| |马拉地文|Marathi|mr| |库尔德文 |Kurdish|ku|
|尼泊尔文|Nepali|ne| |尼泊尔文|Nepali|ne| |立陶宛文|Lithuanian |lt|
|塞尔维亚文（cyrillic)|Serbian(cyrillic)|rs_cyrillic| |塞尔维亚文（cyrillic)|Serbian(cyrillic)|rs_cyrillic| |拉脱维亚文 |Latvian |lv|
|保加利亚文|Bulgarian |bg| |毛利文|Maori|mi| | 达尔瓦文|Dargwa |dar|
|乌克兰文|Ukranian|uk| |马来文 |Malay|ms| | 因古什文|Ingush |inh|
|白俄罗斯文|Belarusian|be| |马耳他文 |Maltese |mt| | 拉克文|Lak |lbe|
|泰卢固文|Telugu |te| |荷兰文 |Dutch |nl| | 莱兹甘文|Lezghian |lez|
|泰米尔文|Tamil |ta| |挪威文 |Norwegian |no| |塔巴萨兰文 |Tabassaran |tab|
|南非荷兰文 |Afrikaans |af| |波兰文|Polish |pl| | 比尔哈文|Bihari |bh|
|阿塞拜疆文 |Azerbaijani |az| | 罗马尼亚文|Romanian |ro| | 迈蒂利文|Maithili |mai|
|波斯尼亚文|Bosnian|bs| | 斯洛伐克文|Slovak |sk| | 昂加文|Angika |ang|
|捷克文|Czech|cs| | 斯洛文尼亚文|Slovenian |sl| | 孟加拉文|Bhojpuri |bho|
|威尔士文 |Welsh |cy| | 阿尔巴尼亚文|Albanian |sq| | 摩揭陀文 |Magahi |mah|
|丹麦文 |Danish|da| | 瑞典文|Swedish |sv| | 那格浦尔文|Nagpur |sck|
|爱沙尼亚文 |Estonian |et| | 西瓦希里文|Swahili |sw| | 尼瓦尔文|Newari |new|
|爱尔兰文 |Irish |ga| | 塔加洛文|Tagalog |tl| | 保加利亚文 |Goan Konkani|gom|
|克罗地亚文|Croatian |hr| | 土耳其文|Turkish |tr| | 沙特阿拉伯文|Saudi Arabia|sa|
|匈牙利文|Hungarian |hu| | 乌兹别克文|Uzbek |uz| | 阿瓦尔文|Avar |ava|
|印尼文|Indonesian|id| | 越南文|Vietnamese |vi| | 阿瓦尔文|Avar |ava|
|冰岛文 |Icelandic|is| | 蒙古文|Mongolian |mn| | 阿迪赫文|Adyghe |ady|
|库尔德文 |Kurdish|ku|
|立陶宛文|Lithuanian |lt|
|拉脱维亚文 |Latvian |lv|
|毛利文|Maori|mi|
|马来文 |Malay|ms|
|马耳他文 |Maltese |mt|
|荷兰文 |Dutch |nl|
|挪威文 |Norwegian |no|
|波兰文|Polish |pl|
| 罗马尼亚文|Romanian |ro|
| 斯洛伐克文|Slovak |sk|
| 斯洛文尼亚文|Slovenian |sl|
| 阿尔巴尼亚文|Albanian |sq|
| 瑞典文|Swedish |sv|
| 西瓦希里文|Swahili |sw|
| 塔加洛文|Tagalog |tl|
| 土耳其文|Turkish |tr|
| 乌兹别克文|Uzbek |uz|
| 越南文|Vietnamese |vi|
| 蒙古文|Mongolian |mn|
| 阿巴扎文|Abaza |abq|
| 阿迪赫文|Adyghe |ady|
| 卡巴丹文|Kabardian |kbd|
| 阿瓦尔文|Avar |ava|
| 达尔瓦文|Dargwa |dar|
| 因古什文|Ingush |inh|
| 拉克文|Lak |lbe|
| 莱兹甘文|Lezghian |lez|
|塔巴萨兰文 |Tabassaran |tab|
| 比尔哈文|Bihari |bh|
| 迈蒂利文|Maithili |mai|
| 昂加文|Angika |ang|
| 孟加拉文|Bhojpuri |bho|
| 摩揭陀文 |Magahi |mah|
| 那格浦尔文|Nagpur |sck|
| 尼瓦尔文|Newari |new|
| 保加利亚文 |Goan Konkani|gom|
| 沙特阿拉伯文|Saudi Arabia|sa|
doc/doc_en/multi_languages_en.md
浏览文件 @ 56feca01
...@@ -13,7 +13,7 @@ Among them, the English model supports the detection and recognition of uppercas ...@@ -13,7 +13,7 @@ Among them, the English model supports the detection and recognition of uppercas
letters and common punctuation, and the recognition of space characters is optimized: letters and common punctuation, and the recognition of space characters is optimized:
<div align="center"> <div align="center">
<img src="../imgs_results/multi_lang/en_1.jpg" width="400" height="600"> <img src="../imgs_results/multi_lang/img_12.jpg" width="900" height="300">
</div> </div>
The multilingual models cover Latin, Arabic, Traditional Chinese, Korean, Japanese, etc.: The multilingual models cover Latin, Arabic, Traditional Chinese, Korean, Japanese, etc.:
...@@ -21,6 +21,8 @@ The multilingual models cover Latin, Arabic, Traditional Chinese, Korean, Japane ...@@ -21,6 +21,8 @@ The multilingual models cover Latin, Arabic, Traditional Chinese, Korean, Japane
<div align="center"> <div align="center">
<img src="../imgs_results/multi_lang/japan_2.jpg" width="600" height="300"> <img src="../imgs_results/multi_lang/japan_2.jpg" width="600" height="300">
<img src="../imgs_results/multi_lang/french_0.jpg" width="300" height="300"> <img src="../imgs_results/multi_lang/french_0.jpg" width="300" height="300">
<img src="../imgs_results/multi_lang/korean_0.jpg" width="500" height="300">
<img src="../imgs_results/multi_lang/arabic_0.jpg" width="300" height="300">
</div> </div>
This document will briefly introduce how to use the multilingual model. This document will briefly introduce how to use the multilingual model.
...@@ -31,14 +33,9 @@ This document will briefly introduce how to use the multilingual model. ...@@ -31,14 +33,9 @@ This document will briefly introduce how to use the multilingual model.
- [2 Quick Use](#Quick_Use) - [2 Quick Use](#Quick_Use)
- [2.1 Command line operation](#Command_line_operation) - [2.1 Command line operation](#Command_line_operation)
- [2.1.1 Prediction of the whole image](#bash_detection+recognition)
- [2.1.2 Recognition](#bash_Recognition)
- [2.1.3 Detection](#bash_detection)
- [2.2 python script running](#python_Script_running) - [2.2 python script running](#python_Script_running)
- [2.2.1 Whole image prediction](#python_detection+recognition)
- [2.2.2 Recognition](#python_Recognition)
- [2.2.3 Detection](#python_detection)
- [3 Custom Training](#Custom_Training) - [3 Custom Training](#Custom_Training)
- [4 Inference and Deployment](#inference)
- [4 Supported languages and abbreviations](#language_abbreviations) - [4 Supported languages and abbreviations](#language_abbreviations)
<a name="Install"></a> <a name="Install"></a>
...@@ -51,7 +48,7 @@ This document will briefly introduce how to use the multilingual model. ...@@ -51,7 +48,7 @@ This document will briefly introduce how to use the multilingual model.
pip install paddlepaddle pip install paddlepaddle
# gpu # gpu
pip instll paddlepaddle-gpu pip install paddlepaddle-gpu
``` ```
<a name="paddleocr_package_install"></a> <a name="paddleocr_package_install"></a>
...@@ -89,7 +86,7 @@ The specific supported [language] (#language_abbreviations) can be viewed in the ...@@ -89,7 +86,7 @@ The specific supported [language] (#language_abbreviations) can be viewed in the
paddleocr --image_dir doc/imgs/japan_2.jpg --lang=japan paddleocr --image_dir doc/imgs/japan_2.jpg --lang=japan
``` ```
![](https://raw.githubusercontent.com/PaddlePaddle/PaddleOCR/release/2.1/doc/imgs/japan_2.jpg) ![](https://raw.githubusercontent.com/PaddlePaddle/PaddleOCR/release/2.0/doc/imgs/japan_2.jpg)
The result is a list, each item contains a text box, text and recognition confidence The result is a list, each item contains a text box, text and recognition confidence
```text ```text
...@@ -106,7 +103,7 @@ The result is a list, each item contains a text box, text and recognition confid ...@@ -106,7 +103,7 @@ The result is a list, each item contains a text box, text and recognition confid
paddleocr --image_dir doc/imgs_words/japan/1.jpg --det false --lang=japan paddleocr --image_dir doc/imgs_words/japan/1.jpg --det false --lang=japan
``` ```
![](https://raw.githubusercontent.com/PaddlePaddle/PaddleOCR/release/2.1/doc/imgs_words/japan/1.jpg) ![](https://raw.githubusercontent.com/PaddlePaddle/PaddleOCR/release/2.0/doc/imgs_words/japan/1.jpg)
The result is a tuple, which returns the recognition result and recognition confidence The result is a tuple, which returns the recognition result and recognition confidence
...@@ -143,6 +140,9 @@ from paddleocr import PaddleOCR, draw_ocr ...@@ -143,6 +140,9 @@ from paddleocr import PaddleOCR, draw_ocr
ocr = PaddleOCR(lang="korean") # The model file will be downloaded automatically when executed for the first time ocr = PaddleOCR(lang="korean") # The model file will be downloaded automatically when executed for the first time
img_path ='doc/imgs/korean_1.jpg' img_path ='doc/imgs/korean_1.jpg'
result = ocr.ocr(img_path) result = ocr.ocr(img_path)
# Recognition and detection can be performed separately through parameter control
# result = ocr.ocr(img_path, det=False) Only perform recognition
# result = ocr.ocr(img_path, rec=False) Only perform detection
# Print detection frame and recognition result # Print detection frame and recognition result
for line in result: for line in result:
print(line) print(line)
...@@ -162,54 +162,6 @@ Visualization of results: ...@@ -162,54 +162,6 @@ Visualization of results:
![](https://raw.githubusercontent.com/PaddlePaddle/PaddleOCR/release/2.1/doc/imgs_results/korean.jpg) ![](https://raw.githubusercontent.com/PaddlePaddle/PaddleOCR/release/2.1/doc/imgs_results/korean.jpg)
* Recognition
```
from paddleocr import PaddleOCR
ocr = PaddleOCR(lang="german")
img_path ='PaddleOCR/doc/imgs_words/german/1.jpg'
result = ocr.ocr(img_path, det=False, cls=True)
for line in result:
print(line)
```
![](https://raw.githubusercontent.com/PaddlePaddle/PaddleOCR/release/2.1/doc/imgs_words/german/1.jpg)
The result is a tuple, which only contains the recognition result and recognition confidence
```
('leider auch jetzt', 0.97538936)
```
* Detection
```python
from paddleocr import PaddleOCR, draw_ocr
ocr = PaddleOCR() # need to run only once to download and load model into memory
img_path ='PaddleOCR/doc/imgs_en/img_12.jpg'
result = ocr.ocr(img_path, rec=False)
for line in result:
print(line)
# show result
from PIL import Image
image = Image.open(img_path).convert('RGB')
im_show = draw_ocr(image, result, txts=None, scores=None, font_path='/path/to/PaddleOCR/doc/fonts/simfang.ttf')
im_show = Image.fromarray(im_show)
im_show.save('result.jpg')
```
The result is a list, each item contains only text boxes
```bash
[[26.0, 457.0], [137.0, 457.0], [137.0, 477.0], [26.0, 477.0]]
[[25.0, 425.0], [372.0, 425.0], [372.0, 448.0], [25.0, 448.0]]
[[128.0, 397.0], [273.0, 397.0], [273.0, 414.0], [128.0, 414.0]]
......
```
Visualization of results:
![](https://raw.githubusercontent.com/PaddlePaddle/PaddleOCR/release/2.1/doc/imgs_results/whl/12_det.jpg)
ppocr also supports direction classification. For more usage methods, please refer to: [whl package instructions](https://github.com/PaddlePaddle/PaddleOCR/blob/release/2.0/doc/doc_ch/whl.md). ppocr also supports direction classification. For more usage methods, please refer to: [whl package instructions](https://github.com/PaddlePaddle/PaddleOCR/blob/release/2.0/doc/doc_ch/whl.md).
<a name="Custom_training"></a> <a name="Custom_training"></a>
...@@ -221,84 +173,61 @@ Modify the training data path, dictionary and other parameters. ...@@ -221,84 +173,61 @@ Modify the training data path, dictionary and other parameters.
For specific data preparation and training process, please refer to: [Text Detection](../doc_en/detection_en.md), [Text Recognition](../doc_en/recognition_en.md), more functions such as predictive deployment, For specific data preparation and training process, please refer to: [Text Detection](../doc_en/detection_en.md), [Text Recognition](../doc_en/recognition_en.md), more functions such as predictive deployment,
For functions such as data annotation, you can read the complete [Document Tutorial](../../README.md). For functions such as data annotation, you can read the complete [Document Tutorial](../../README.md).
<a name="language_abbreviation"></a>
## 4 Support languages and abbreviations <a name="inference"></a>
## 4 Inference and Deployment
| Language | Abbreviation |
| --- | --- | In addition to installing the whl package for quick forecasting,
|chinese and english|ch| ppocr also provides a variety of forecasting deployment methods.
|english|en| If necessary, you can read related documents:
|french|fr|
|german|german| - [Python Inference](./inference_en.md)
|japan|japan| - [C++ Inference](../../deploy/cpp_infer/readme_en.md)
|korean|korean| - [Serving](../../deploy/hubserving/readme_en.md)
|chinese traditional |chinese_cht| - [Mobile](https://github.com/PaddlePaddle/PaddleOCR/blob/develop/deploy/lite/readme_en.md)
| Italian |it| - [Benchmark](./benchmark_en.md)
|Spanish |es|
| Portuguese|pt|
|Russia|ru| <a name="language_abbreviations"></a>
|Arabic|ar| ## 5 Support languages and abbreviations
|Hindi|hi|
|Uyghur|ug| | Language | Abbreviation | | Language | Abbreviation |
|Persian|fa| | --- | --- | --- | --- | --- |
|Urdu|ur| |chinese and english|ch| |Arabic|ar|
| Serbian(latin) |rs_latin| |english|en| |Hindi|hi|
|Occitan |oc| |french|fr| |Uyghur|ug|
|Marathi|mr| |german|german| |Persian|fa|
|Nepali|ne| |japan|japan| |Urdu|ur|
|Serbian(cyrillic)|rs_cyrillic| |korean|korean| | Serbian(latin) |rs_latin|
|Bulgarian |bg| |chinese traditional |ch_tra| |Occitan |oc|
|Ukranian|uk| | Italian |it| |Marathi|mr|
|Belarusian|be| |Spanish |es| |Nepali|ne|
|Telugu |te| | Portuguese|pt| |Serbian(cyrillic)|rs_cyrillic|
|Tamil |ta| |Russia|ru||Bulgarian |bg|
|Afrikaans |af| |Ukranian|uk| |Estonian |et|
|Azerbaijani |az| |Belarusian|be| |Irish |ga|
|Bosnian|bs| |Telugu |te| |Croatian |hr|
|Czech|cs| |Saudi Arabia|sa| |Hungarian |hu|
|Welsh |cy| |Tamil |ta| |Indonesian|id|
|Danish|da| |Afrikaans |af| |Icelandic|is|
|Estonian |et| |Azerbaijani |az||Kurdish|ku|
|Irish |ga| |Bosnian|bs| |Lithuanian |lt|
|Croatian |hr| |Czech|cs| |Latvian |lv|
|Hungarian |hu| |Welsh |cy| |Maori|mi|
|Indonesian|id| |Danish|da| |Malay|ms|
|Icelandic|is| |Maltese |mt| |Adyghe |ady|
|Kurdish|ku| |Dutch |nl| |Kabardian |kbd|
|Lithuanian |lt| |Norwegian |no| |Avar |ava|
|Latvian |lv| |Polish |pl| |Dargwa |dar|
|Maori|mi| |Romanian |ro| |Ingush |inh|
|Malay|ms| |Slovak |sk| |Lak |lbe|
|Maltese |mt| |Slovenian |sl| |Lezghian |lez|
|Dutch |nl| |Albanian |sq| |Tabassaran |tab|
|Norwegian |no| |Swedish |sv| |Bihari |bh|
|Polish |pl| |Swahili |sw| |Maithili |mai|
|Romanian |ro| |Tagalog |tl| |Angika |ang|
|Slovak |sk| |Turkish |tr| |Bhojpuri |bho|
|Slovenian |sl| |Uzbek |uz| |Magahi |mah|
|Albanian |sq| |Vietnamese |vi| |Nagpur |sck|
|Swedish |sv| |Mongolian |mn| |Newari |new|
|Swahili |sw| |Abaza |abq| |Goan Konkani|gom|
|Tagalog |tl|
|Turkish |tr|
|Uzbek |uz|
|Vietnamese |vi|
|Mongolian |mn|
|Abaza |abq|
|Adyghe |ady|
|Kabardian |kbd|
|Avar |ava|
|Dargwa |dar|
|Ingush |inh|
|Lak |lbe|
|Lezghian |lez|
|Tabassaran |tab|
|Bihari |bh|
|Maithili |mai|
|Angika |ang|
|Bhojpuri |bho|
|Magahi |mah|
|Nagpur |sck|
|Newari |new|
|Goan Konkani|gom|
|Saudi Arabia|sa|
ppocr/modeling/backbones/rec_resnet_vd.py
浏览文件 @ 56feca01
...@@ -249,7 +249,7 @@ class ResNet(nn.Layer): ...@@ -249,7 +249,7 @@ class ResNet(nn.Layer):
name=conv_name)) name=conv_name))
shortcut = True shortcut = True
self.block_list.append(bottleneck_block) self.block_list.append(bottleneck_block)
self.out_channels = num_filters[block] self.out_channels = num_filters[block] * 4
else: else:
for block in range(len(depth)): for block in range(len(depth)):
shortcut = False shortcut = False
......
ppocr/postprocess/rec_postprocess.py
浏览文件 @ 56feca01
...@@ -218,6 +218,7 @@ class SRNLabelDecode(BaseRecLabelDecode): ...@@ -218,6 +218,7 @@ class SRNLabelDecode(BaseRecLabelDecode):
**kwargs): **kwargs):
super(SRNLabelDecode, self).__init__(character_dict_path, super(SRNLabelDecode, self).__init__(character_dict_path,
character_type, use_space_char) character_type, use_space_char)
self.max_text_length = kwargs.get('max_text_length', 25)
def __call__(self, preds, label=None, *args, **kwargs): def __call__(self, preds, label=None, *args, **kwargs):
pred = preds['predict'] pred = preds['predict']
...@@ -229,9 +230,9 @@ class SRNLabelDecode(BaseRecLabelDecode): ...@@ -229,9 +230,9 @@ class SRNLabelDecode(BaseRecLabelDecode):
preds_idx = np.argmax(pred, axis=1) preds_idx = np.argmax(pred, axis=1)
preds_prob = np.max(pred, axis=1) preds_prob = np.max(pred, axis=1)
preds_idx = np.reshape(preds_idx, [-1, 25]) preds_idx = np.reshape(preds_idx, [-1, self.max_text_length])
preds_prob = np.reshape(preds_prob, [-1, 25]) preds_prob = np.reshape(preds_prob, [-1, self.max_text_length])
text = self.decode(preds_idx, preds_prob) text = self.decode(preds_idx, preds_prob)
......
ppocr/utils/save_load.py
浏览文件 @ 56feca01
...@@ -121,7 +121,7 @@ def init_model(config, model, logger, optimizer=None, lr_scheduler=None): ...@@ -121,7 +121,7 @@ def init_model(config, model, logger, optimizer=None, lr_scheduler=None):
return best_model_dict return best_model_dict
def save_model(net, def save_model(model,
optimizer, optimizer,
model_path, model_path,
logger, logger,
...@@ -133,7 +133,7 @@ def save_model(net, ...@@ -133,7 +133,7 @@ def save_model(net,
""" """
_mkdir_if_not_exist(model_path, logger) _mkdir_if_not_exist(model_path, logger)
model_prefix = os.path.join(model_path, prefix) model_prefix = os.path.join(model_path, prefix)
paddle.save(net.state_dict(), model_prefix + '.pdparams') paddle.save(model.state_dict(), model_prefix + '.pdparams')
paddle.save(optimizer.state_dict(), model_prefix + '.pdopt') paddle.save(optimizer.state_dict(), model_prefix + '.pdopt')
# save metric and config # save metric and config
......
tools/export_model.py
浏览文件 @ 56feca01
...@@ -53,17 +53,19 @@ def main(): ...@@ -53,17 +53,19 @@ def main():
save_path = '{}/inference'.format(config['Global']['save_inference_dir']) save_path = '{}/inference'.format(config['Global']['save_inference_dir'])
if config['Architecture']['algorithm'] == "SRN": if config['Architecture']['algorithm'] == "SRN":
max_text_length = config['Architecture']['Head']['max_text_length']
other_shape = [ other_shape = [
paddle.static.InputSpec( paddle.static.InputSpec(
shape=[None, 1, 64, 256], dtype='float32'), [ shape=[None, 1, 64, 256], dtype='float32'), [
paddle.static.InputSpec( paddle.static.InputSpec(
shape=[None, 256, 1], shape=[None, 256, 1],
dtype="int64"), paddle.static.InputSpec( dtype="int64"), paddle.static.InputSpec(
shape=[None, 25, 1], shape=[None, max_text_length, 1], dtype="int64"),
dtype="int64"), paddle.static.InputSpec(
shape=[None, 8, 25, 25], dtype="int64"),
paddle.static.InputSpec( paddle.static.InputSpec(
shape=[None, 8, 25, 25], dtype="int64") shape=[None, 8, max_text_length, max_text_length],
dtype="int64"), paddle.static.InputSpec(
shape=[None, 8, max_text_length, max_text_length],
dtype="int64")
] ]
] ]
model = to_static(model, input_spec=other_shape) model = to_static(model, input_spec=other_shape)
......
tools/program.py
浏览文件 @ 56feca01
...@@ -18,6 +18,7 @@ from __future__ import print_function ...@@ -18,6 +18,7 @@ from __future__ import print_function
import os import os
import sys import sys
import platform
import yaml import yaml
import time import time
import shutil import shutil
...@@ -159,6 +160,8 @@ def train(config, ...@@ -159,6 +160,8 @@ def train(config,
eval_batch_step = config['Global']['eval_batch_step'] eval_batch_step = config['Global']['eval_batch_step']
global_step = 0 global_step = 0
if 'global_step' in pre_best_model_dict:
global_step = pre_best_model_dict['global_step']
start_eval_step = 0 start_eval_step = 0
if type(eval_batch_step) == list and len(eval_batch_step) >= 2: if type(eval_batch_step) == list and len(eval_batch_step) >= 2:
start_eval_step = eval_batch_step[0] start_eval_step = eval_batch_step[0]
...@@ -196,8 +199,12 @@ def train(config, ...@@ -196,8 +199,12 @@ def train(config,
train_reader_cost = 0.0 train_reader_cost = 0.0
batch_sum = 0 batch_sum = 0
batch_start = time.time() batch_start = time.time()
for idx, batch in enumerate(train_dataloader()): max_iter = len(train_dataloader) - 1 if platform.system(
) == "Windows" else len(train_dataloader)
for idx, batch in enumerate(train_dataloader):
train_reader_cost += time.time() - batch_start train_reader_cost += time.time() - batch_start
if idx >= max_iter:
break
lr = optimizer.get_lr() lr = optimizer.get_lr()
images = batch[0] images = batch[0]
if use_srn: if use_srn:
...@@ -285,7 +292,8 @@ def train(config, ...@@ -285,7 +292,8 @@ def train(config,
is_best=True, is_best=True,
prefix='best_accuracy', prefix='best_accuracy',
best_model_dict=best_model_dict, best_model_dict=best_model_dict,
epoch=epoch) epoch=epoch,
global_step=global_step)
best_str = 'best metric, {}'.format(', '.join([ best_str = 'best metric, {}'.format(', '.join([
'{}: {}'.format(k, v) for k, v in best_model_dict.items() '{}: {}'.format(k, v) for k, v in best_model_dict.items()
])) ]))
...@@ -307,7 +315,8 @@ def train(config, ...@@ -307,7 +315,8 @@ def train(config,
is_best=False, is_best=False,
prefix='latest', prefix='latest',
best_model_dict=best_model_dict, best_model_dict=best_model_dict,
epoch=epoch) epoch=epoch,
global_step=global_step)
if dist.get_rank() == 0 and epoch > 0 and epoch % save_epoch_step == 0: if dist.get_rank() == 0 and epoch > 0 and epoch % save_epoch_step == 0:
save_model( save_model(
model, model,
...@@ -317,7 +326,8 @@ def train(config, ...@@ -317,7 +326,8 @@ def train(config,
is_best=False, is_best=False,
prefix='iter_epoch_{}'.format(epoch), prefix='iter_epoch_{}'.format(epoch),
best_model_dict=best_model_dict, best_model_dict=best_model_dict,
epoch=epoch) epoch=epoch,
global_step=global_step)
best_str = 'best metric, {}'.format(', '.join( best_str = 'best metric, {}'.format(', '.join(
['{}: {}'.format(k, v) for k, v in best_model_dict.items()])) ['{}: {}'.format(k, v) for k, v in best_model_dict.items()]))
logger.info(best_str) logger.info(best_str)
...@@ -333,8 +343,10 @@ def eval(model, valid_dataloader, post_process_class, eval_class, ...@@ -333,8 +343,10 @@ def eval(model, valid_dataloader, post_process_class, eval_class,
total_frame = 0.0 total_frame = 0.0
total_time = 0.0 total_time = 0.0
pbar = tqdm(total=len(valid_dataloader), desc='eval model:') pbar = tqdm(total=len(valid_dataloader), desc='eval model:')
max_iter = len(valid_dataloader) - 1 if platform.system(
) == "Windows" else len(valid_dataloader)
for idx, batch in enumerate(valid_dataloader): for idx, batch in enumerate(valid_dataloader):
if idx >= len(valid_dataloader): if idx >= max_iter:
break break
images = batch[0] images = batch[0]
start = time.time() start = time.time()
......
Markdown is supported
0% 或 .
You are about to add 0 people to the discussion. Proceed with caution.
先完成此消息的编辑！
想要评论请 注册 或