Merge remote-tracking branch 'origin/android_demo' into android_demo

e296a968 · WenmuZhou · 8601b8e9 · 2f2d152c · e296a968 · e296a968
Commit e296a968 authored May 10, 2021 by WenmuZhou
20 changed files
--- a/doc/doc_ch/multi_languages.md
+++ b/doc/doc_ch/multi_languages.md
@@ -5,6 +5,25 @@
 - 2021.4.9 支持**80种**语言的检测和识别
 - 2021.4.9 支持**轻量高精度**英文模型检测识别
+PaddleOCR 旨在打造一套丰富、领先、且实用的OCR工具库，不仅提供了通用场景下的中英文模型，也提供了专门在英文场景下训练的模型，
+和覆盖[80个语言](#语种缩写)的小语种模型。
+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">
+    <img src="../imgs_results/multi_lang/en_1.jpg" width="400" height="600">
+</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：
+<div align="center">
+    <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">
+</div>
+本文档将简要介绍小语种模型的使用方法。
 - [1 安装](#安装)
    - [1.1 paddle 安装](#paddle安装)
    - [1.2 paddleocr package 安装](#paddleocr_package_安装)  
@@ -40,7 +59,7 @@ pip instll paddlepaddle-gpu
 pip 安装
 ```
-pip install "paddleocr>=2.0.4" # 推荐使用2.0.4版本
+pip install "paddleocr>=2.0.6" # 推荐使用2.0.6版本
 ```
 本地构建并安装
 ```
@@ -68,7 +87,11 @@ Paddleocr目前支持80个语种，可以通过修改--lang参数进行切换，
 paddleocr --image_dir doc/imgs/japan_2.jpg --lang=japan
 ```
-![](https://raw.githubusercontent.com/PaddlePaddle/PaddleOCR/release/2.0/doc/imgs/japan_2.jpg)
+<div align="center">
+    <img src="https://raw.githubusercontent.com/PaddlePaddle/PaddleOCR/release/2.1/doc/imgs/japan_2.jpg" width="800">
+</div>
 结果是一个list，每个item包含了文本框，文字和识别置信度
 ```text
@@ -111,7 +134,7 @@ paddleocr --image_dir PaddleOCR/doc/imgs/11.jpg --rec false
 <a name="python_脚本运行"></a>
 ### 2.2 python 脚本运行
-ppocr 也支持在python脚本中运行，便于嵌入到您自己的代码中：
+ppocr 也支持在python脚本中运行，便于嵌入到您自己的代码中 ：
 * 整图预测（检测+识别）
@@ -132,14 +155,16 @@ image = Image.open(img_path).convert('RGB')
 boxes = [line[0] for line in result]
 txts = [line[1][0] for line in result]
 scores = [line[1][1] for line in result]
-im_show = draw_ocr(image, boxes, txts, scores, font_path='/path/to/PaddleOCR/doc/korean.ttf')
+im_show = draw_ocr(image, boxes, txts, scores, font_path='/path/to/PaddleOCR/doc/fonts/korean.ttf')
 im_show = Image.fromarray(im_show)
 im_show.save('result.jpg')
 ```
 结果可视化:
-![](https://raw.githubusercontent.com/PaddlePaddle/PaddleOCR/release/2.0/doc/imgs_results/korean.jpg)
+<div align="center">
+    <img src="https://raw.githubusercontent.com/PaddlePaddle/PaddleOCR/release/2.1/doc/imgs_results/korean.jpg" width="800">
+</div>
 * 识别预测
@@ -152,7 +177,8 @@ for line in result:
    print(line)
 ```
-![](https://raw.githubusercontent.com/PaddlePaddle/PaddleOCR/release/2.0/doc/imgs_words/german/1.jpg)
+![](../imgs_words/german/1.jpg)
 结果是一个tuple，只包含识别结果和识别置信度
@@ -187,7 +213,10 @@ im_show.save('result.jpg')
 ```
 结果可视化 ：
-![](https://raw.githubusercontent.com/PaddlePaddle/PaddleOCR/release/2.0/doc/imgs_results/whl/12_det.jpg)
+<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)。
@@ -211,7 +240,7 @@ ppocr 支持使用自己的数据进行自定义训练或finetune, 其中识别
 |德文|german|german|
 |日文|japan|japan|
 |韩文|korean|korean|
-|中文繁体|chinese traditional |ch_tra|
+|中文繁体|chinese traditional |chinese_cht|
 |意大利文| Italian |it|
 |西班牙文|Spanish |es|
 |葡萄牙文| Portuguese|pt|
@@ -230,10 +259,9 @@ ppocr 支持使用自己的数据进行自定义训练或finetune, 其中识别
 |乌克兰文|Ukranian|uk|
 |白俄罗斯文|Belarusian|be|
 |泰卢固文|Telugu |te|
-|卡纳达文|Kannada |kn|
 |泰米尔文|Tamil |ta|
 |南非荷兰文 |Afrikaans |af|
-|阿塞拜疆文 |Azerbaijani	|az|
+|阿塞拜疆文 |Azerbaijani  |az|
 |波斯尼亚文|Bosnian|bs|
 |捷克文|Czech|cs|
 |威尔士文 |Welsh |cy|

--- a/doc/doc_ch/pgnet.md
+++ b/doc/doc_ch/pgnet.md
@@ -2,7 +2,7 @@
 - [一、简介](#简介)
 - [二、环境配置](#环境配置)
 - [三、快速使用](#快速使用)
- [四、模型训练、评估、推理](#快速训练)
+- [四、模型训练、评估、推理](#模型训练、评估、推理)
 <a name="简介"></a>
 ## 一、简介
@@ -16,14 +16,31 @@ OCR算法可以分为两阶段算法和端对端的算法。二阶段OCR算法
 - 提出基于图的修正模块（GRM）来进一步提高模型识别性能
 - 精度更高，预测速度更快
-PGNet算法细节详见[论文](https://www.aaai.org/AAAI21Papers/AAAI-2885.WangP.pdf)， 算法原理图如下所示：
+PGNet算法细节详见[论文](https://www.aaai.org/AAAI21Papers/AAAI-2885.WangP.pdf) ,算法原理图如下所示：
 ![](../pgnet_framework.png)
 输入图像经过特征提取送入四个分支，分别是：文本边缘偏移量预测TBO模块，文本中心线预测TCL模块，文本方向偏移量预测TDO模块，以及文本字符分类图预测TCC模块。
 其中TBO以及TCL的输出经过后处理后可以得到文本的检测结果，TCL、TDO、TCC负责文本识别。
 其检测识别效果图如下：
 ![](../imgs_results/e2e_res_img293_pgnet.png)
 ![](../imgs_results/e2e_res_img295_pgnet.png)
+### 性能指标
+测试集: Total Text
+测试环境: NVIDIA Tesla V100-SXM2-16GB
+|PGNetA|det_precision|det_recall|det_f_score|e2e_precision|e2e_recall|e2e_f_score|FPS|下载|
+| --- | --- | --- | --- | --- | --- | --- | --- | --- |
+|Paper|85.30|86.80|86.1|-|-|61.7|38.20 (size=640)|-|
+|Ours|87.03|82.48|84.69|61.71|58.43|60.03|48.73 (size=768)|[下载链接](https://paddleocr.bj.bcebos.com/dygraph_v2.0/pgnet/en_server_pgnetA.tar)|
+*note：PaddleOCR里的PGNet实现针对预测速度做了优化，在精度下降可接受范围内，可以显著提升端对端预测速度*
 <a name="环境配置"></a>
 ## 二、环境配置
 请先参考[快速安装](./installation.md)配置PaddleOCR运行环境。
@@ -49,24 +66,24 @@ wget https://paddleocr.bj.bcebos.com/dygraph_v2.0/pgnet/e2e_server_pgnetA_infer.
 ### 单张图像或者图像集合预测
 ```bash
 # 预测image_dir指定的单张图像
-python3 tools/infer/predict_e2e.py --e2e_algorithm="PGNet" --image_dir="./doc/imgs_en/img623.jpg" --e2e_model_dir="./inference/e2e/" --e2e_pgnet_polygon=True
+python3 tools/infer/predict_e2e.py --e2e_algorithm="PGNet" --image_dir="./doc/imgs_en/img623.jpg" --e2e_model_dir="./inference/e2e_server_pgnetA_infer/" --e2e_pgnet_polygon=True
 # 预测image_dir指定的图像集合
-python3 tools/infer/predict_e2e.py --e2e_algorithm="PGNet" --image_dir="./doc/imgs_en/" --e2e_model_dir="./inference/e2e/" --e2e_pgnet_polygon=True
+python3 tools/infer/predict_e2e.py --e2e_algorithm="PGNet" --image_dir="./doc/imgs_en/" --e2e_model_dir="./inference/e2e_server_pgnetA_infer/" --e2e_pgnet_polygon=True
 # 如果想使用CPU进行预测，需设置use_gpu参数为False
-python3 tools/infer/predict_e2e.py --e2e_algorithm="PGNet" --image_dir="./doc/imgs_en/img623.jpg" --e2e_model_dir="./inference/e2e/" --e2e_pgnet_polygon=True --use_gpu=False
+python3 tools/infer/predict_e2e.py --e2e_algorithm="PGNet" --image_dir="./doc/imgs_en/img623.jpg" --e2e_model_dir="./inference/e2e_server_pgnetA_infer/" --e2e_pgnet_polygon=True --use_gpu=False
 ```
 ### 可视化结果
 可视化文本检测结果默认保存到./inference_results文件夹里面，结果文件的名称前缀为'e2e_res'。结果示例如下：
 ![](../imgs_results/e2e_res_img623_pgnet.jpg)
-<a name="快速训练"></a>
+<a name="模型训练、评估、推理"></a>
 ## 四、模型训练、评估、推理
 本节以totaltext数据集为例，介绍PaddleOCR中端到端模型的训练、评估与测试。
 ###  准备数据
-下载解压[totaltext](https://github.com/cs-chan/Total-Text-Dataset/blob/master/Dataset/README.md)数据集到PaddleOCR/train_data/目录，数据集组织结构：
+下载解压[totaltext](https://github.com/cs-chan/Total-Text-Dataset/blob/master/Dataset/README.md) 数据集到PaddleOCR/train_data/目录，数据集组织结构：
 ```
 /PaddleOCR/train_data/total_text/train/
  |- rgb/            # total_text数据集的训练数据
@@ -135,20 +152,20 @@ python3 tools/eval.py -c configs/e2e/e2e_r50_vd_pg.yml  -o Global.checkpoints="{
 ### 模型预测
 测试单张图像的端到端识别效果
 ```shell
-python3 tools/infer_e2e.py -c configs/e2e/e2e_r50_vd_pg.yml -o Global.infer_img="./doc/imgs_en/img_10.jpg" Global.pretrained_model="./output/det_db/best_accuracy" Global.load_static_weights=false
+python3 tools/infer_e2e.py -c configs/e2e/e2e_r50_vd_pg.yml -o Global.infer_img="./doc/imgs_en/img_10.jpg" Global.pretrained_model="./output/e2e_pgnet/best_accuracy" Global.load_static_weights=false
 ```
 测试文件夹下所有图像的端到端识别效果
 ```shell
-python3 tools/infer_e2e.py -c configs/e2e/e2e_r50_vd_pg.yml -o Global.infer_img="./doc/imgs_en/" Global.pretrained_model="./output/det_db/best_accuracy" Global.load_static_weights=false
+python3 tools/infer_e2e.py -c configs/e2e/e2e_r50_vd_pg.yml -o Global.infer_img="./doc/imgs_en/" Global.pretrained_model="./output/e2e_pgnet/best_accuracy" Global.load_static_weights=false
 ```
 ### 预测推理
-#### (1).四边形文本检测模型（ICDAR2015）  
+#### (1). 四边形文本检测模型（ICDAR2015）  
 首先将PGNet端到端训练过程中保存的模型，转换成inference model。以基于Resnet50_vd骨干网络，以英文数据集训练的模型为例[模型下载地址](https://paddleocr.bj.bcebos.com/dygraph_v2.0/pgnet/en_server_pgnetA.tar) ，可以使用如下命令进行转换：
 ```
 wget https://paddleocr.bj.bcebos.com/dygraph_v2.0/pgnet/en_server_pgnetA.tar && tar xf en_server_pgnetA.tar
-python3 tools/export_model.py -c configs/e2e/e2e_r50_vd_pg.yml -o Global.pretrained_model=./en_server_pgnetA/iter_epoch_450 Global.load_static_weights=False Global.save_inference_dir=./inference/e2e
+python3 tools/export_model.py -c configs/e2e/e2e_r50_vd_pg.yml -o Global.pretrained_model=./en_server_pgnetA/best_accuracy Global.load_static_weights=False Global.save_inference_dir=./inference/e2e
 ```
 **PGNet端到端模型推理，需要设置参数`--e2e_algorithm="PGNet"`**，可以执行如下命令：
 ```
@@ -158,7 +175,7 @@ python3 tools/infer/predict_e2e.py --e2e_algorithm="PGNet" --image_dir="./doc/im
 ![](../imgs_results/e2e_res_img_10_pgnet.jpg)
-#### (2).弯曲文本检测模型（Total-Text）
+#### (2). 弯曲文本检测模型（Total-Text）
 对于弯曲文本样例
 **PGNet端到端模型推理，需要设置参数`--e2e_algorithm="PGNet"`，同时，还需要增加参数`--e2e_pgnet_polygon=True`，**可以执行如下命令：

--- a/doc/doc_ch/recognition.md
+++ b/doc/doc_ch/recognition.md
@@ -138,7 +138,7 @@ PaddleOCR内置了一部分字典，可以按需使用。
 `ppocr/utils/dict/german_dict.txt` 是一个包含131个字符的德文字典
-`ppocr/utils/dict/en_dict.txt` 是一个包含63个字符的英文字典
+`ppocr/utils/en_dict.txt` 是一个包含96个字符的英文字典
@@ -285,7 +285,7 @@ Eval:
 <a name="小语种"></a>
 #### 2.3 小语种
-PaddleOCR目前已支持26种（除中文外）语种识别，`configs/rec/multi_languages` 路径下提供了一个多语言的配置文件模版: [rec_multi_language_lite_train.yml](../../configs/rec/multi_language/rec_multi_language_lite_train.yml)。
+PaddleOCR目前已支持80种（除中文外）语种识别，`configs/rec/multi_languages` 路径下提供了一个多语言的配置文件模版: [rec_multi_language_lite_train.yml](../../configs/rec/multi_language/rec_multi_language_lite_train.yml)。
 您有两种方式创建所需的配置文件：
@@ -368,26 +368,12 @@ PaddleOCR目前已支持26种（除中文外）语种识别，`configs/rec/multi
 | rec_ger_lite_train.yml |  CRNN |   Mobilenet_v3 small 0.5 |  None   |  BiLSTM |  ctc  | 德语   | german |
 | rec_japan_lite_train.yml |  CRNN |   Mobilenet_v3 small 0.5 |  None   |  BiLSTM |  ctc  | 日语  | japan |
 | rec_korean_lite_train.yml |  CRNN |   Mobilenet_v3 small 0.5 |  None   |  BiLSTM |  ctc  | 韩语  | korean |
-| rec_it_lite_train.yml |  CRNN |   Mobilenet_v3 small 0.5 |  None   |  BiLSTM |  ctc  | 意大利语  | it |
+| rec_latin_lite_train.yml |  CRNN |   Mobilenet_v3 small 0.5 |  None   |  BiLSTM |  ctc  | 拉丁字母  | latin |
-| rec_xi_lite_train.yml |  CRNN |   Mobilenet_v3 small 0.5 |  None   |  BiLSTM |  ctc  | 西班牙语 |  xi |
+| rec_arabic_lite_train.yml |  CRNN |   Mobilenet_v3 small 0.5 |  None   |  BiLSTM |  ctc  | 阿拉伯字母 |  ar |
-| rec_pu_lite_train.yml |  CRNN |   Mobilenet_v3 small 0.5 |  None   |  BiLSTM |  ctc  | 葡萄牙语   | pu |
+| rec_cyrillic_lite_train.yml |  CRNN |   Mobilenet_v3 small 0.5 |  None   |  BiLSTM |  ctc  | 斯拉夫字母  | cyrillic |
-| rec_ru_lite_train.yml |  CRNN |   Mobilenet_v3 small 0.5 |  None   |  BiLSTM |  ctc  | 俄罗斯语  | ru |
+| rec_devanagari_lite_train.yml |  CRNN |   Mobilenet_v3 small 0.5 |  None   |  BiLSTM |  ctc  | 梵文字母  | devanagari |
-| rec_ar_lite_train.yml |  CRNN |   Mobilenet_v3 small 0.5 |  None   |  BiLSTM |  ctc  | 阿拉伯语  | ar |
-| rec_hi_lite_train.yml |  CRNN |   Mobilenet_v3 small 0.5 |  None   |  BiLSTM |  ctc  | 印地语 |  hi |
+更多支持语种请参考: [多语言模型](https://github.com/PaddlePaddle/PaddleOCR/blob/release/2.1/doc/doc_ch/multi_languages.md#%E8%AF%AD%E7%A7%8D%E7%BC%A9%E5%86%99)
-| rec_ug_lite_train.yml |  CRNN |   Mobilenet_v3 small 0.5 |  None   |  BiLSTM |  ctc  | 维吾尔语  | ug |
-| rec_fa_lite_train.yml |  CRNN |   Mobilenet_v3 small 0.5 |  None   |  BiLSTM |  ctc  | 波斯语  | fa |
-| rec_ur_ite_train.yml |  CRNN |   Mobilenet_v3 small 0.5 |  None   |  BiLSTM |  ctc  | 乌尔都语  | ur |
-| rec_rs_lite_train.yml |  CRNN |   Mobilenet_v3 small 0.5 |  None   |  BiLSTM |  ctc  | 塞尔维亚(latin)语 | rs |
-| rec_oc_lite_train.yml |  CRNN |   Mobilenet_v3 small 0.5 |  None   |  BiLSTM |  ctc  | 欧西坦语   | oc |
-| rec_mr_lite_train.yml |  CRNN |   Mobilenet_v3 small 0.5 |  None   |  BiLSTM |  ctc  | 马拉地语  | mr |
-| rec_ne_lite_train.yml |  CRNN |   Mobilenet_v3 small 0.5 |  None   |  BiLSTM |  ctc  | 尼泊尔语  | ne |
-| rec_rsc_lite_train.yml |  CRNN |   Mobilenet_v3 small 0.5 |  None   |  BiLSTM |  ctc  | 塞尔维亚(cyrillic)语 |  rsc |
-| rec_bg_lite_train.yml |  CRNN |   Mobilenet_v3 small 0.5 |  None   |  BiLSTM |  ctc  | 保加利亚语  | bg |
-| rec_uk_lite_train.yml |  CRNN |   Mobilenet_v3 small 0.5 |  None   |  BiLSTM |  ctc  | 乌克兰语  | uk |
-| rec_be_lite_train.yml |  CRNN |   Mobilenet_v3 small 0.5 |  None   |  BiLSTM |  ctc  | 白俄罗斯语   | be |
-| rec_te_lite_train.yml |  CRNN |   Mobilenet_v3 small 0.5 |  None   |  BiLSTM |  ctc  | 泰卢固语  | te |
-| rec_ka_lite_train.yml |  CRNN |   Mobilenet_v3 small 0.5 |  None   |  BiLSTM |  ctc  | 卡纳达语  | ka |
-| rec_ta_lite_train.yml |  CRNN |   Mobilenet_v3 small 0.5 |  None   |  BiLSTM |  ctc  | 泰米尔语 |  ta |
 多语言模型训练方式与中文模型一致，训练数据集均为100w的合成数据，少量的字体可以在 [百度网盘](https://pan.baidu.com/s/1bS_u207Rm7YbY33wOECKDA) 上下载，提取码：frgi。

--- a/doc/doc_en/models_list_en.md
+++ b/doc/doc_en/models_list_en.md
@@ -102,27 +102,16 @@ python3 generate_multi_language_configs.py -l it \
 | german_mobile_v2.0_rec |Lightweight model for German recognition|[rec_german_lite_train.yml](../../configs/rec/multi_language/rec_german_lite_train.yml)|2.65M|[inference model](https://paddleocr.bj.bcebos.com/dygraph_v2.0/multilingual/german_mobile_v2.0_rec_infer.tar) / [trained model](https://paddleocr.bj.bcebos.com/dygraph_v2.0/multilingual/german_mobile_v2.0_rec_train.tar) |
 | korean_mobile_v2.0_rec |Lightweight model for Korean recognition|[rec_korean_lite_train.yml](../../configs/rec/multi_language/rec_korean_lite_train.yml)|3.9M|[inference model](https://paddleocr.bj.bcebos.com/dygraph_v2.0/multilingual/korean_mobile_v2.0_rec_infer.tar) / [trained model](https://paddleocr.bj.bcebos.com/dygraph_v2.0/multilingual/korean_mobile_v2.0_rec_train.tar) |
 | japan_mobile_v2.0_rec |Lightweight model for Japanese recognition|[rec_japan_lite_train.yml](../../configs/rec/multi_language/rec_japan_lite_train.yml)|4.23M|[inference model](https://paddleocr.bj.bcebos.com/dygraph_v2.0/multilingual/japan_mobile_v2.0_rec_infer.tar) / [trained model](https://paddleocr.bj.bcebos.com/dygraph_v2.0/multilingual/japan_mobile_v2.0_rec_train.tar) |
-| it_mobile_v2.0_rec |Lightweight model for Italian recognition|rec_it_lite_train.yml|2.53M|[inference model](https://paddleocr.bj.bcebos.com/dygraph_v2.0/multilingual/it_mobile_v2.0_rec_infer.tar) / [trained model](https://paddleocr.bj.bcebos.com/dygraph_v2.0/multilingual/it_mobile_v2.0_rec_train.tar) |
+| chinese_cht_mobile_v2.0_rec |Lightweight model for chinese cht recognition|rec_chinese_cht_lite_train.yml|5.63M|[推理模型](https://paddleocr.bj.bcebos.com/dygraph_v2.0/multilingual/chinese_cht_mobile_v2.0_rec_infer.tar) / [训练模型](https://paddleocr.bj.bcebos.com/dygraph_v2.0/multilingual/chinese_cht_mobile_v2.0_rec_train.tar) |
-| xi_mobile_v2.0_rec |Lightweight model for Spanish recognition|rec_xi_lite_train.yml|2.53M|[inference model](https://paddleocr.bj.bcebos.com/dygraph_v2.0/multilingual/xi_mobile_v2.0_rec_infer.tar) / [trained model](https://paddleocr.bj.bcebos.com/dygraph_v2.0/multilingual/xi_mobile_v2.0_rec_train.tar) |
-| pu_mobile_v2.0_rec |Lightweight model for Portuguese recognition|rec_pu_lite_train.yml|2.63M|[inference model](https://paddleocr.bj.bcebos.com/dygraph_v2.0/multilingual/pu_mobile_v2.0_rec_infer.tar) / [trained model](https://paddleocr.bj.bcebos.com/dygraph_v2.0/multilingual/pu_mobile_v2.0_rec_train.tar) |
-| ru_mobile_v2.0_rec |Lightweight model for Russia recognition|rec_ru_lite_train.yml|2.63M|[inference model](https://paddleocr.bj.bcebos.com/dygraph_v2.0/multilingual/ru_mobile_v2.0_rec_infer.tar) / [trained model](https://paddleocr.bj.bcebos.com/dygraph_v2.0/multilingual/ru_mobile_v2.0_rec_train.tar) |
-| ar_mobile_v2.0_rec |Lightweight model for Arabic recognition|rec_ar_lite_train.yml|2.53M|[inference model](https://paddleocr.bj.bcebos.com/dygraph_v2.0/multilingual/ar_mobile_v2.0_rec_infer.tar) / [trained model](https://paddleocr.bj.bcebos.com/dygraph_v2.0/multilingual/ar_mobile_v2.0_rec_train.tar) |
-| hi_mobile_v2.0_rec |Lightweight model for Hindi recognition|rec_hi_lite_train.yml|2.63M|[inference model](https://paddleocr.bj.bcebos.com/dygraph_v2.0/multilingual/hi_mobile_v2.0_rec_infer.tar) / [trained model](https://paddleocr.bj.bcebos.com/dygraph_v2.0/multilingual/hi_mobile_v2.0_rec_train.tar) |
-| chinese_cht_mobile_v2.0_rec |Lightweight model for chinese traditional recognition|rec_chinese_cht_lite_train.yml|5.63M|[inference model](https://paddleocr.bj.bcebos.com/dygraph_v2.0/multilingual/chinese_cht_mobile_v2.0_rec_infer.tar) / [trained model](https://paddleocr.bj.bcebos.com/dygraph_v2.0/multilingual/chinese_cht_mobile_v2.0_rec_train.tar) |
-| ug_mobile_v2.0_rec |Lightweight model for Uyghur recognition|rec_ug_lite_train.yml|2.63M|[inference model](https://paddleocr.bj.bcebos.com/dygraph_v2.0/multilingual/ug_mobile_v2.0_rec_infer.tar) / [trained model](https://paddleocr.bj.bcebos.com/dygraph_v2.0/multilingual/ug_mobile_v2.0_rec_train.tar) |
-| fa_mobile_v2.0_rec |Lightweight model for Persian recognition|rec_fa_lite_train.yml|2.63M|[inference model](https://paddleocr.bj.bcebos.com/dygraph_v2.0/multilingual/fa_mobile_v2.0_rec_infer.tar) / [trained model](https://paddleocr.bj.bcebos.com/dygraph_v2.0/multilingual/fa_mobile_v2.0_rec_train.tar) |
-| ur_mobile_v2.0_rec |Lightweight model for Urdu recognition|rec_ur_lite_train.yml|2.63M|[inference model](https://paddleocr.bj.bcebos.com/dygraph_v2.0/multilingual/ur_mobile_v2.0_rec_infer.tar) / [trained model](https://paddleocr.bj.bcebos.com/dygraph_v2.0/multilingual/ur_mobile_v2.0_rec_train.tar) |
-| rs_mobile_v2.0_rec |Lightweight model for Serbian(latin) recognition|rec_rs_lite_train.yml|2.53M|[inference model](https://paddleocr.bj.bcebos.com/dygraph_v2.0/multilingual/rs_mobile_v2.0_rec_infer.tar) / [trained model](https://paddleocr.bj.bcebos.com/dygraph_v2.0/multilingual/rs_mobile_v2.0_rec_train.tar) |
-| oc_mobile_v2.0_rec |Lightweight model for Occitan recognition|rec_oc_lite_train.yml|2.53M|[inference model](https://paddleocr.bj.bcebos.com/dygraph_v2.0/multilingual/oc_mobile_v2.0_rec_infer.tar) / [trained model](https://paddleocr.bj.bcebos.com/dygraph_v2.0/multilingual/oc_mobile_v2.0_rec_train.tar) |
-| mr_mobile_v2.0_rec |Lightweight model for Marathi recognition|rec_mr_lite_train.yml|2.63M|[inference model](https://paddleocr.bj.bcebos.com/dygraph_v2.0/multilingual/mr_mobile_v2.0_rec_infer.tar) / [trained model](https://paddleocr.bj.bcebos.com/dygraph_v2.0/multilingual/mr_mobile_v2.0_rec_train.tar) |
-| ne_mobile_v2.0_rec |Lightweight model for Nepali recognition|rec_ne_lite_train.yml|2.63M|[inference model](https://paddleocr.bj.bcebos.com/dygraph_v2.0/multilingual/ne_mobile_v2.0_rec_infer.tar) / [trained model](https://paddleocr.bj.bcebos.com/dygraph_v2.0/multilingual/ne_mobile_v2.0_rec_train.tar) |
-| rsc_mobile_v2.0_rec |Lightweight model for Serbian(cyrillic) recognition|rec_rsc_lite_train.yml|2.63M|[inference model](https://paddleocr.bj.bcebos.com/dygraph_v2.0/multilingual/rsc_mobile_v2.0_rec_infer.tar) / [trained model](https://paddleocr.bj.bcebos.com/dygraph_v2.0/multilingual/rsc_mobile_v2.0_rec_train.tar) |
-| bg_mobile_v2.0_rec |Lightweight model for Bulgarian recognition|rec_bg_lite_train.yml|2.63M|[inference model](https://paddleocr.bj.bcebos.com/dygraph_v2.0/multilingual/bg_mobile_v2.0_rec_infer.tar) / [trained model](https://paddleocr.bj.bcebos.com/dygraph_v2.0/multilingual/bg_mobile_v2.0_rec_train.tar) |
-| uk_mobile_v2.0_rec |Lightweight model for Ukranian recognition|rec_uk_lite_train.yml|2.63M|[inference model](https://paddleocr.bj.bcebos.com/dygraph_v2.0/multilingual/uk_mobile_v2.0_rec_infer.tar) / [trained model](https://paddleocr.bj.bcebos.com/dygraph_v2.0/multilingual/uk_mobile_v2.0_rec_train.tar) |
-| be_mobile_v2.0_rec |Lightweight model for Belarusian recognition|rec_be_lite_train.yml|2.63M|[inference model](https://paddleocr.bj.bcebos.com/dygraph_v2.0/multilingual/be_mobile_v2.0_rec_infer.tar) / [trained model](https://paddleocr.bj.bcebos.com/dygraph_v2.0/multilingual/be_mobile_v2.0_rec_train.tar) |
 | te_mobile_v2.0_rec |Lightweight model for Telugu recognition|rec_te_lite_train.yml|2.63M|[inference model](https://paddleocr.bj.bcebos.com/dygraph_v2.0/multilingual/te_mobile_v2.0_rec_infer.tar) / [trained model](https://paddleocr.bj.bcebos.com/dygraph_v2.0/multilingual/te_mobile_v2.0_rec_train.tar) |
 | ka_mobile_v2.0_rec |Lightweight model for Kannada recognition|rec_ka_lite_train.yml|2.63M|[inference model](https://paddleocr.bj.bcebos.com/dygraph_v2.0/multilingual/ka_mobile_v2.0_rec_infer.tar) / [trained model](https://paddleocr.bj.bcebos.com/dygraph_v2.0/multilingual/ka_mobile_v2.0_rec_train.tar) |
 | ta_mobile_v2.0_rec |Lightweight model for Tamil recognition|rec_ta_lite_train.yml|2.63M|[inference model](https://paddleocr.bj.bcebos.com/dygraph_v2.0/multilingual/ta_mobile_v2.0_rec_infer.tar) / [trained model](https://paddleocr.bj.bcebos.com/dygraph_v2.0/multilingual/ta_mobile_v2.0_rec_train.tar) |
+| latin_mobile_v2.0_rec |  Lightweight model for latin recognition |  [rec_latin_lite_train.yml](../../configs/rec/multi_language/rec_latin_lite_train.yml) |2.6M|[推理模型](https://paddleocr.bj.bcebos.com/dygraph_v2.0/multilingual/latin_ppocr_mobile_v2.0_rec_infer.tar) / [训练模型](https://paddleocr.bj.bcebos.com/dygraph_v2.0/multilingual/latin_ppocr_mobile_v2.0_rec_train.tar) |
+| arabic_mobile_v2.0_rec |  Lightweight model for arabic recognition | [rec_arabic_lite_train.yml](../../configs/rec/multi_language/rec_arabic_lite_train.yml) |2.6M|[推理模型](https://paddleocr.bj.bcebos.com/dygraph_v2.0/multilingual/arabic_ppocr_mobile_v2.0_rec_infer.tar) / [训练模型](https://paddleocr.bj.bcebos.com/dygraph_v2.0/multilingual/arabic_ppocr_mobile_v2.0_rec_train.tar) |
+| cyrillic_mobile_v2.0_rec | Lightweight model for cyrillic recognition | [rec_cyrillic_lite_train.yml](../../configs/rec/multi_language/rec_cyrillic_lite_train.yml) |2.6M|[推理模型](https://paddleocr.bj.bcebos.com/dygraph_v2.0/multilingual/cyrillic_ppocr_mobile_v2.0_rec_infer.tar) / [训练模型](https://paddleocr.bj.bcebos.com/dygraph_v2.0/multilingual/cyrillic_ppocr_mobile_v2.0_rec_train.tar) |
+| devanagari_mobile_v2.0_rec | Lightweight model for devanagari recognition | [rec_devanagari_lite_train.yml](../../configs/rec/multi_language/rec_devanagari_lite_train.yml) |2.6M|[推理模型](https://paddleocr.bj.bcebos.com/dygraph_v2.0/multilingual/devanagari_ppocr_mobile_v2.0_rec_infer.tar) / [训练模型](https://paddleocr.bj.bcebos.com/dygraph_v2.0/multilingual/devanagari_ppocr_mobile_v2.0_rec_train.tar) |
+For more supported languages, please refer to : [Multi-language model](./multi_languages_en.md)
 <a name="Angle"></a>

--- a/doc/doc_en/multi_languages_en.md
+++ b/doc/doc_en/multi_languages_en.md
@@ -2,24 +2,44 @@
 **Recent Update**
-2021.4.9 supports the detection and recognition of 80 languages
+- 2021.4.9 supports the detection and recognition of 80 languages
-2021.4.9 supports **lightweight high-precision** English model detection and recognition
+- 2021.4.9 supports **lightweight high-precision** English model detection and recognition
-[1 Installation](#Install)
+PaddleOCR aims to create a rich, leading, and practical OCR tool library, which not only provides
-    -[1.1 paddle installation](#paddleinstallation)
+Chinese and English models in general scenarios, but also provides models specifically trained
-    -[1.2 paddleocr package installation](#paddleocr_package_install)
+in English scenarios. And multilingual models covering [80 languages](#language_abbreviations).
-[2 Quick Use](#Quick_Use)
+Among them, the English model supports the detection and recognition of uppercase and lowercase
-    -[2.1 Command line operation](#Command_line_operation)
+letters and common punctuation, and the recognition of space characters is optimized:
-     -[2.1.1 Prediction of the whole image](#bash_detection+recognition)
-     -[2.1.2 Recognition](#bash_Recognition)
+<div align="center">
-     -[2.1.3 Detection](#bash_detection)
+    <img src="../imgs_results/multi_lang/en_1.jpg" width="400" height="600">
-    -[2.2 python script running](#python_Script_running)
+</div>
-     -[2.2.1 Whole image prediction](#python_detection+recognition)
-     -[2.2.2 Recognition](#python_Recognition)
+The multilingual models cover Latin, Arabic, Traditional Chinese, Korean, Japanese, etc.:
-     -[2.2.3 Detection](#python_detection)
-[3 Custom Training](#Custom_Training)
+<div align="center">
-[4 Supported languages and abbreviations](#language_abbreviations)
+    <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">
+</div>
+This document will briefly introduce how to use the multilingual model.
+- [1 Installation](#Install)
+    - [1.1 paddle installation](#paddleinstallation)
+    - [1.2 paddleocr package installation](#paddleocr_package_install)
+- [2 Quick Use](#Quick_Use)
+    - [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)
+     - [2.2.1 Whole image prediction](#python_detection+recognition)
+     - [2.2.2 Recognition](#python_Recognition)
+     - [2.2.3 Detection](#python_detection)
+- [3 Custom Training](#Custom_Training)
+- [4 Supported languages and abbreviations](#language_abbreviations)
 <a name="Install"></a>
 ## 1 Installation
@@ -40,7 +60,7 @@ pip instll paddlepaddle-gpu
 pip install
 ```
-pip install "paddleocr>=2.0.4" # 2.0.4 version is recommended
+pip install "paddleocr>=2.0.6" # 2.0.6 version is recommended
 ```
 Build and install locally
 ```
@@ -69,7 +89,7 @@ The specific supported [language] (#language_abbreviations) can be viewed in the
 paddleocr --image_dir doc/imgs/japan_2.jpg --lang=japan
 ```
-![](https://raw.githubusercontent.com/PaddlePaddle/PaddleOCR/release/2.0/doc/imgs/japan_2.jpg)
+![](https://raw.githubusercontent.com/PaddlePaddle/PaddleOCR/release/2.1/doc/imgs/japan_2.jpg)
 The result is a list, each item contains a text box, text and recognition confidence
 ```text
@@ -86,7 +106,7 @@ 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
 ```
-![](https://raw.githubusercontent.com/PaddlePaddle/PaddleOCR/release/2.0/doc/imgs_words/japan/1.jpg)
+![](https://raw.githubusercontent.com/PaddlePaddle/PaddleOCR/release/2.1/doc/imgs_words/japan/1.jpg)
 The result is a tuple, which returns the recognition result and recognition confidence
@@ -133,13 +153,13 @@ image = Image.open(img_path).convert('RGB')
 boxes = [line[0] for line in result]
 txts = [line[1][0] for line in result]
 scores = [line[1][1] for line in result]
-im_show = draw_ocr(image, boxes, txts, scores, font_path='/path/to/PaddleOCR/doc/korean.ttf')
+im_show = draw_ocr(image, boxes, txts, scores, font_path='/path/to/PaddleOCR/doc/fonts/korean.ttf')
 im_show = Image.fromarray(im_show)
 im_show.save('result.jpg')
 ```
 Visualization of results:
-![](https://raw.githubusercontent.com/PaddlePaddle/PaddleOCR/release/2.0/doc/imgs_results/korean.jpg)
+![](https://raw.githubusercontent.com/PaddlePaddle/PaddleOCR/release/2.1/doc/imgs_results/korean.jpg)
 * Recognition
@@ -153,7 +173,7 @@ for line in result:
    print(line)
 ```
-![](https://raw.githubusercontent.com/PaddlePaddle/PaddleOCR/release/2.0/doc/imgs_words/german/1.jpg)
+![](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
@@ -188,7 +208,7 @@ The result is a list, each item contains only text boxes
 ```
 Visualization of results:
-![](https://raw.githubusercontent.com/PaddlePaddle/PaddleOCR/release/2.0/doc/imgs_results/whl/12_det.jpg)
+![](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).
@@ -212,7 +232,7 @@ For functions such as data annotation, you can read the complete [Document Tutor
 |german|german|
 |japan|japan|
 |korean|korean|
-|chinese traditional |ch_tra|
+|chinese traditional |chinese_cht|
 | Italian |it|
 |Spanish |es|
 | Portuguese|pt|
@@ -231,7 +251,6 @@ For functions such as data annotation, you can read the complete [Document Tutor
 |Ukranian|uk|
 |Belarusian|be|
 |Telugu |te|
-|Kannada |kn|
 |Tamil |ta|
 |Afrikaans |af|
 |Azerbaijani    |az|

--- a/doc/doc_en/pgnet_en.md
+++ b/doc/doc_en/pgnet_en.md
@@ -15,7 +15,7 @@ In recent years, the end-to-end OCR algorithm has been well developed, including
 - A graph based modification module (GRM) is proposed to further improve the performance of model recognition
 - Higher accuracy and faster prediction speed
-For details of PGNet algorithm, please refer to [paper](https://www.aaai.org/AAAI21Papers/AAAI-2885.WangP.pdf), The schematic diagram of the algorithm is as follows:
+For details of PGNet algorithm, please refer to [paper](https://www.aaai.org/AAAI21Papers/AAAI-2885.WangP.pdf) ,The schematic diagram of the algorithm is as follows:
 ![](../pgnet_framework.png)
 After feature extraction, the input image is sent to four branches: TBO module for text edge offset prediction, TCL module for text centerline prediction, TDO module for text direction offset prediction, and TCC module for text character classification graph prediction.
 The output of TBO and TCL can get text detection results after post-processing, and TCL, TDO and TCC are responsible for text recognition.
@@ -23,6 +23,16 @@ The output of TBO and TCL can get text detection results after post-processing,
 The results of detection and recognition are as follows:
 ![](../imgs_results/e2e_res_img293_pgnet.png)
 ![](../imgs_results/e2e_res_img295_pgnet.png)
+### Performance
+####Test set: Total Text
+####Test environment: NVIDIA Tesla V100-SXM2-16GB
+|PGNetA|det_precision|det_recall|det_f_score|e2e_precision|e2e_recall|e2e_f_score|FPS|download|
+| --- | --- | --- | --- | --- | --- | --- | --- | --- |
+|Paper|85.30|86.80|86.1|-|-|61.7|38.20 (size=640)|-|
+|Ours|87.03|82.48|84.69|61.71|58.43|60.03|48.73 (size=768)|[download link](https://paddleocr.bj.bcebos.com/dygraph_v2.0/pgnet/en_server_pgnetA.tar)|
+*note：PGNet in PaddleOCR optimizes the prediction speed, and can significantly improve the end-to-end prediction speed within the acceptable range of accuracy reduction*
 <a name="Environment_Configuration"></a>
 ## 2. Environment Configuration
@@ -49,13 +59,13 @@ After decompression, there should be the following file structure:
 ### Single image or image set prediction
 ```bash
 # Prediction single image specified by image_dir
-python3 tools/infer/predict_e2e.py --e2e_algorithm="PGNet" --image_dir="./doc/imgs_en/img623.jpg" --e2e_model_dir="./inference/e2e/" --e2e_pgnet_polygon=True
+python3 tools/infer/predict_e2e.py --e2e_algorithm="PGNet" --image_dir="./doc/imgs_en/img623.jpg" --e2e_model_dir="./inference/e2e_server_pgnetA_infer/" --e2e_pgnet_polygon=True
 # Prediction the collection of images specified by image_dir
-python3 tools/infer/predict_e2e.py --e2e_algorithm="PGNet" --image_dir="./doc/imgs_en/" --e2e_model_dir="./inference/e2e/" --e2e_pgnet_polygon=True
+python3 tools/infer/predict_e2e.py --e2e_algorithm="PGNet" --image_dir="./doc/imgs_en/" --e2e_model_dir="./inference/e2e_server_pgnetA_infer/" --e2e_pgnet_polygon=True
 # If you want to use CPU for prediction, you need to set use_gpu parameter is false
-python3 tools/infer/predict_e2e.py --e2e_algorithm="PGNet" --image_dir="./doc/imgs_en/img623.jpg" --e2e_model_dir="./inference/e2e/" --e2e_pgnet_polygon=True --use_gpu=False
+python3 tools/infer/predict_e2e.py --e2e_algorithm="PGNet" --image_dir="./doc/imgs_en/img623.jpg" --e2e_model_dir="./inference/e2e_server_pgnetA_infer/" --e2e_pgnet_polygon=True --use_gpu=False
 ```
 ### Visualization results
 The visualized end-to-end results are saved to the `./inference_results` folder by default, and the name of the result file is prefixed with 'e2e_res'. Examples of results are as follows:
@@ -141,12 +151,12 @@ python3 tools/eval.py -c configs/e2e/e2e_r50_vd_pg.yml  -o Global.checkpoints="{
 ### Model Test
 Test the end-to-end result on a single image:
 ```shell
-python3 tools/infer_e2e.py -c configs/e2e/e2e_r50_vd_pg.yml -o Global.infer_img="./doc/imgs_en/img_10.jpg" Global.pretrained_model="./output/det_db/best_accuracy" Global.load_static_weights=false
+python3 tools/infer_e2e.py -c configs/e2e/e2e_r50_vd_pg.yml -o Global.infer_img="./doc/imgs_en/img_10.jpg" Global.pretrained_model="./output/e2e_pgnet/best_accuracy" Global.load_static_weights=false
 ```
 Test the end-to-end result on all images in the folder:
 ```shell
-python3 tools/infer_e2e.py -c configs/e2e/e2e_r50_vd_pg.yml -o Global.infer_img="./doc/imgs_en/" Global.pretrained_model="./output/det_db/best_accuracy" Global.load_static_weights=false
+python3 tools/infer_e2e.py -c configs/e2e/e2e_r50_vd_pg.yml -o Global.infer_img="./doc/imgs_en/" Global.pretrained_model="./output/e2e_pgnet/best_accuracy" Global.load_static_weights=false
 ```
 ### Model inference
@@ -154,7 +164,7 @@ python3 tools/infer_e2e.py -c configs/e2e/e2e_r50_vd_pg.yml -o Global.infer_img=
 First, convert the model saved in the PGNet end-to-end training process into an inference model. In the first stage of training based on composite dataset, the model of English data set training is taken as an example[model download link](https://paddleocr.bj.bcebos.com/dygraph_v2.0/pgnet/en_server_pgnetA.tar), you can use the following command to convert:
 ```
 wget https://paddleocr.bj.bcebos.com/dygraph_v2.0/pgnet/en_server_pgnetA.tar && tar xf en_server_pgnetA.tar
-python3 tools/export_model.py -c configs/e2e/e2e_r50_vd_pg.yml -o Global.pretrained_model=./en_server_pgnetA/iter_epoch_450 Global.load_static_weights=False Global.save_inference_dir=./inference/e2e
+python3 tools/export_model.py -c configs/e2e/e2e_r50_vd_pg.yml -o Global.pretrained_model=./en_server_pgnetA/best_accuracy Global.load_static_weights=False Global.save_inference_dir=./inference/e2e
 ```
 **For PGNet quadrangle end-to-end model inference, you need to set the parameter `--e2e_algorithm="PGNet"`**, run the following command:
 ```

--- a/doc/doc_en/recognition_en.md
+++ b/doc/doc_en/recognition_en.md
@@ -131,7 +131,7 @@ PaddleOCR has built-in dictionaries, which can be used on demand.
 `ppocr/utils/dict/german_dict.txt` is a German dictionary with 131 characters
-`ppocr/utils/dict/en_dict.txt` is a English dictionary with 63 characters
+`ppocr/utils/en_dict.txt` is a English dictionary with 96 characters
 The current multi-language model is still in the demo stage and will continue to optimize the model and add languages. **You are very welcome to provide us with dictionaries and fonts in other languages**,
@@ -279,7 +279,7 @@ Eval:
 <a name="Multi_language"></a>
 #### 2.3 Multi-language
-PaddleOCR currently supports 26 (except Chinese) language recognition. A multi-language configuration file template is
+PaddleOCR currently supports 80 (except Chinese) language recognition. A multi-language configuration file template is
 provided under the path `configs/rec/multi_languages`: [rec_multi_language_lite_train.yml](../../configs/rec/multi_language/rec_multi_language_lite_train.yml)。
 There are two ways to create the required configuration file:：
@@ -368,27 +368,12 @@ Currently, the multi-language algorithms supported by PaddleOCR are:
 | rec_ger_lite_train.yml |  CRNN |   Mobilenet_v3 small 0.5 |  None   |  BiLSTM |  ctc  | German   | german |
 | rec_japan_lite_train.yml |  CRNN |   Mobilenet_v3 small 0.5 |  None   |  BiLSTM |  ctc  | Japanese | japan |
 | rec_korean_lite_train.yml |  CRNN |   Mobilenet_v3 small 0.5 |  None   |  BiLSTM |  ctc  | Korean  | korean |
-| rec_it_lite_train.yml |  CRNN |   Mobilenet_v3 small 0.5 |  None   |  BiLSTM |  ctc  | Italian  | it |
+| rec_latin_lite_train.yml |  CRNN |   Mobilenet_v3 small 0.5 |  None   |  BiLSTM |  ctc  | Latin  | latin |
-| rec_xi_lite_train.yml |  CRNN |   Mobilenet_v3 small 0.5 |  None   |  BiLSTM |  ctc  | Spanish |  xi |
+| rec_arabic_lite_train.yml |  CRNN |   Mobilenet_v3 small 0.5 |  None   |  BiLSTM |  ctc  | arabic |  ar |
-| rec_pu_lite_train.yml |  CRNN |   Mobilenet_v3 small 0.5 |  None   |  BiLSTM |  ctc  | Portuguese   | pu |
+| rec_cyrillic_lite_train.yml |  CRNN |   Mobilenet_v3 small 0.5 |  None   |  BiLSTM |  ctc  | cyrillic   | cyrillic |
-| rec_ru_lite_train.yml |  CRNN |   Mobilenet_v3 small 0.5 |  None   |  BiLSTM |  ctc  | Russia  | ru |
+| rec_devanagari_lite_train.yml |  CRNN |   Mobilenet_v3 small 0.5 |  None   |  BiLSTM |  ctc  | devanagari  | devanagari |
-| rec_ar_lite_train.yml |  CRNN |   Mobilenet_v3 small 0.5 |  None   |  BiLSTM |  ctc  | Arabic  | ar |
-| rec_hi_lite_train.yml |  CRNN |   Mobilenet_v3 small 0.5 |  None   |  BiLSTM |  ctc  | Hindi |  hi |
-| rec_ug_lite_train.yml |  CRNN |   Mobilenet_v3 small 0.5 |  None   |  BiLSTM |  ctc  | Uyghur  | ug |
-| rec_fa_lite_train.yml |  CRNN |   Mobilenet_v3 small 0.5 |  None   |  BiLSTM |  ctc  | Persian(Farsi)  | fa |
-| rec_ur_ite_train.yml |  CRNN |   Mobilenet_v3 small 0.5 |  None   |  BiLSTM |  ctc  | Urdu  | ur |
-| rec_rs_lite_train.yml |  CRNN |   Mobilenet_v3 small 0.5 |  None   |  BiLSTM |  ctc  | Serbian(latin) | rs |
-| rec_oc_lite_train.yml |  CRNN |   Mobilenet_v3 small 0.5 |  None   |  BiLSTM |  ctc  | Occitan  | oc |
-| rec_mr_lite_train.yml |  CRNN |   Mobilenet_v3 small 0.5 |  None   |  BiLSTM |  ctc  | Marathi  | mr |
-| rec_ne_lite_train.yml |  CRNN |   Mobilenet_v3 small 0.5 |  None   |  BiLSTM |  ctc  | Nepali  | ne |
-| rec_rsc_lite_train.yml |  CRNN |   Mobilenet_v3 small 0.5 |  None   |  BiLSTM |  ctc  | Serbian(cyrillic) |  rsc |
-| rec_bg_lite_train.yml |  CRNN |   Mobilenet_v3 small 0.5 |  None   |  BiLSTM |  ctc  | Bulgarian  | bg |
-| rec_uk_lite_train.yml |  CRNN |   Mobilenet_v3 small 0.5 |  None   |  BiLSTM |  ctc  | Ukranian  | uk |
-| rec_be_lite_train.yml |  CRNN |   Mobilenet_v3 small 0.5 |  None   |  BiLSTM |  ctc  | Belarusian   | be |
-| rec_te_lite_train.yml |  CRNN |   Mobilenet_v3 small 0.5 |  None   |  BiLSTM |  ctc  | Telugu  | te |
-| rec_ka_lite_train.yml |  CRNN |   Mobilenet_v3 small 0.5 |  None   |  BiLSTM |  ctc  | Kannada  | ka |
-| rec_ta_lite_train.yml |  CRNN |   Mobilenet_v3 small 0.5 |  None   |  BiLSTM |  ctc  | Tamil |  ta |
+For more supported languages, please refer to : [Multi-language model](https://github.com/PaddlePaddle/PaddleOCR/blob/release/2.1/doc/doc_en/multi_languages_en.md#4-support-languages-and-abbreviations)
 The multi-language model training method is the same as the Chinese model. The training data set is 100w synthetic data. A small amount of fonts and test data can be downloaded on [Baidu Netdisk](https://pan.baidu.com/s/1bS_u207Rm7YbY33wOECKDA),Extraction code:frgi.

--- a/doc/imgs_results/multi_lang/en_1.jpg
+++ b/doc/imgs_results/multi_lang/en_1.jpg
--- a/doc/imgs_results/multi_lang/en_2.jpg
+++ b/doc/imgs_results/multi_lang/en_2.jpg
--- a/doc/imgs_results/multi_lang/en_3.jpg
+++ b/doc/imgs_results/multi_lang/en_3.jpg
--- a/doc/imgs_results/multi_lang/french_0.jpg
+++ b/doc/imgs_results/multi_lang/french_0.jpg
--- a/doc/imgs_results/multi_lang/img_01.jpg
+++ b/doc/imgs_results/multi_lang/img_01.jpg
--- a/doc/imgs_results/multi_lang/img_02.jpg
+++ b/doc/imgs_results/multi_lang/img_02.jpg
--- a/doc/imgs_results/multi_lang/japan_2.jpg
+++ b/doc/imgs_results/multi_lang/japan_2.jpg
--- a/paddleocr.py
+++ b/paddleocr.py
@@ -30,12 +30,17 @@ from ppocr.utils.logging import get_logger
 logger = get_logger()
 from ppocr.utils.utility import check_and_read_gif, get_image_file_list
+from tools.infer.utility import draw_ocr
 __all__ = ['PaddleOCR']
 model_urls = {
-    'det':
+    'det': {
-    'https://paddleocr.bj.bcebos.com/dygraph_v2.0/ch/ch_ppocr_mobile_v2.0_det_infer.tar',
+        'ch':
+        'https://paddleocr.bj.bcebos.com/dygraph_v2.0/ch/ch_ppocr_mobile_v2.0_det_infer.tar',
+        'en':
+        'https://paddleocr.bj.bcebos.com/dygraph_v2.0/multilingual/en_ppocr_mobile_v2.0_det_infer.tar'
+    },
    'rec': {
        'ch': {
            'url':
@@ -45,7 +50,7 @@ model_urls = {
        'en': {
            'url':
            'https://paddleocr.bj.bcebos.com/dygraph_v2.0/multilingual/en_number_mobile_v2.0_rec_infer.tar',
-            'dict_path': './ppocr/utils/dict/en_dict.txt'
+            'dict_path': './ppocr/utils/en_dict.txt'
        },
        'french': {
            'url':
@@ -113,7 +118,7 @@ model_urls = {
 }
 SUPPORT_DET_MODEL = ['DB']
-VERSION = 2.0
+VERSION = '2.1'
 SUPPORT_REC_MODEL = ['CRNN']
 BASE_DIR = os.path.expanduser("~/.paddleocr/")
@@ -199,7 +204,7 @@ def parse_args(mMain=True, add_help=True):
        parser.add_argument("--rec_model_dir", type=str, default=None)
        parser.add_argument("--rec_image_shape", type=str, default="3, 32, 320")
        parser.add_argument("--rec_char_type", type=str, default='ch')
-        parser.add_argument("--rec_batch_num", type=int, default=30)
+        parser.add_argument("--rec_batch_num", type=int, default=6)
        parser.add_argument("--max_text_length", type=int, default=25)
        parser.add_argument("--rec_char_dict_path", type=str, default=None)
        parser.add_argument("--use_space_char", type=bool, default=True)
@@ -209,7 +214,7 @@ def parse_args(mMain=True, add_help=True):
        parser.add_argument("--cls_model_dir", type=str, default=None)
        parser.add_argument("--cls_image_shape", type=str, default="3, 48, 192")
        parser.add_argument("--label_list", type=list, default=['0', '180'])
-        parser.add_argument("--cls_batch_num", type=int, default=30)
+        parser.add_argument("--cls_batch_num", type=int, default=6)
        parser.add_argument("--cls_thresh", type=float, default=0.9)
        parser.add_argument("--enable_mkldnn", type=bool, default=False)
@@ -243,7 +248,7 @@ def parse_args(mMain=True, add_help=True):
            rec_model_dir=None,
            rec_image_shape="3, 32, 320",
            rec_char_type='ch',
-            rec_batch_num=30,
+            rec_batch_num=6,
            max_text_length=25,
            rec_char_dict_path=None,
            use_space_char=True,
@@ -251,7 +256,7 @@ def parse_args(mMain=True, add_help=True):
            cls_model_dir=None,
            cls_image_shape="3, 48, 192",
            label_list=['0', '180'],
-            cls_batch_num=30,
+            cls_batch_num=6,
            cls_thresh=0.9,
            enable_mkldnn=False,
            use_zero_copy_run=False,
@@ -274,10 +279,10 @@ class PaddleOCR(predict_system.TextSystem):
        self.use_angle_cls = postprocess_params.use_angle_cls
        lang = postprocess_params.lang
        latin_lang = [
-            'af', 'az', 'bs', 'cs', 'cy', 'da', 'de', 'en', 'es', 'et', 'fr',
+            'af', 'az', 'bs', 'cs', 'cy', 'da', 'de', 'es', 'et', 'fr', 'ga',
-            'ga', 'hr', 'hu', 'id', 'is', 'it', 'ku', 'la', 'lt', 'lv', 'mi',
+            'hr', 'hu', 'id', 'is', 'it', 'ku', 'la', 'lt', 'lv', 'mi', 'ms',
-            'ms', 'mt', 'nl', 'no', 'oc', 'pi', 'pl', 'pt', 'ro', 'rs_latin',
+            'mt', 'nl', 'no', 'oc', 'pi', 'pl', 'pt', 'ro', 'rs_latin', 'sk',
-            'sk', 'sl', 'sq', 'sv', 'sw', 'tl', 'tr', 'uz', 'vi'
+            'sl', 'sq', 'sv', 'sw', 'tl', 'tr', 'uz', 'vi'
        ]
        arabic_lang = ['ar', 'fa', 'ug', 'ur']
        cyrillic_lang = [
@@ -299,6 +304,10 @@ class PaddleOCR(predict_system.TextSystem):
        assert lang in model_urls[
            'rec'], 'param lang must in {}, but got {}'.format(
                model_urls['rec'].keys(), lang)
+        if lang == "ch":
+            det_lang = "ch"
+        else:
+            det_lang = "en"
        use_inner_dict = False
        if postprocess_params.rec_char_dict_path is None:
            use_inner_dict = True
@@ -307,17 +316,17 @@ class PaddleOCR(predict_system.TextSystem):
        # init model dir
        if postprocess_params.det_model_dir is None:
-            postprocess_params.det_model_dir = os.path.join(
+            postprocess_params.det_model_dir = os.path.join(BASE_DIR, VERSION,
-                BASE_DIR, '{}/det'.format(VERSION))
+                                                            'det', det_lang)
        if postprocess_params.rec_model_dir is None:
-            postprocess_params.rec_model_dir = os.path.join(
+            postprocess_params.rec_model_dir = os.path.join(BASE_DIR, VERSION,
-                BASE_DIR, '{}/rec/{}'.format(VERSION, lang))
+                                                            'rec', lang)
        if postprocess_params.cls_model_dir is None:
-            postprocess_params.cls_model_dir = os.path.join(
+            postprocess_params.cls_model_dir = os.path.join(BASE_DIR, 'cls')
-                BASE_DIR, '{}/cls'.format(VERSION))
        print(postprocess_params)
        # download model
-        maybe_download(postprocess_params.det_model_dir, model_urls['det'])
+        maybe_download(postprocess_params.det_model_dir,
+                       model_urls['det'][det_lang])
        maybe_download(postprocess_params.rec_model_dir,
                       model_urls['rec'][lang]['url'])
        maybe_download(postprocess_params.cls_model_dir, model_urls['cls'])

--- a/ppocr/data/imaug/label_ops.py
+++ b/ppocr/data/imaug/label_ops.py
@@ -96,7 +96,7 @@ class BaseRecLabelEncode(object):
            'ch', 'en', 'EN_symbol', 'french', 'german', 'japan', 'korean',
            'EN', 'it', 'xi', 'pu', 'ru', 'ar', 'ta', 'ug', 'fa', 'ur', 'rs',
            'oc', 'rsc', 'bg', 'uk', 'be', 'te', 'ka', 'chinese_cht', 'hi',
-            'mr', 'ne'
+            'mr', 'ne', 'latin', 'arabic', 'cyrillic', 'devanagari'
        ]
        assert character_type in support_character_type, "Only {} are supported now but get {}".format(
            support_character_type, character_type)
@@ -200,18 +200,16 @@ class E2ELabelEncode(BaseRecLabelEncode):
        self.pad_num = len(self.dict)  # the length to pad
    def __call__(self, data):
-        text_label_index_list, temp_text = [], []
        texts = data['strs']
+        temp_texts = []
        for text in texts:
            text = text.lower()
-            temp_text = []
+            text = self.encode(text)
-            for c_ in text:
+            if text is None:
-                if c_ in self.dict:
+                return None
-                    temp_text.append(self.dict[c_])
+            text = text + [self.pad_num] * (self.max_text_len - len(text))
-            temp_text = temp_text + [self.pad_num] * (self.max_text_len -
+            temp_texts.append(text)
-                                                      len(temp_text))
+        data['strs'] = np.array(temp_texts)
-            text_label_index_list.append(temp_text)
-        data['strs'] = np.array(text_label_index_list)
        return data

--- a/ppocr/data/pgnet_dataset.py
+++ b/ppocr/data/pgnet_dataset.py
@@ -64,9 +64,6 @@ class PGDataSet(Dataset):
            for line in f.readlines():
                poly_str, txt = line.strip().split('\t')
                poly = list(map(float, poly_str.split(',')))
-                if self.mode.lower() == "eval":
-                    while len(poly) < 100:
-                        poly.append(-1)
                text_polys.append(
                    np.array(
                        poly, dtype=np.float32).reshape(-1, 2))
@@ -139,23 +136,21 @@ class PGDataSet(Dataset):
        try:
            if self.data_format == 'icdar':
                im_path = os.path.join(data_path, 'rgb', data_line)
-                if self.mode.lower() == "eval":
+                poly_path = os.path.join(data_path, 'poly',
-                    poly_path = os.path.join(data_path, 'poly_gt',
+                                         data_line.split('.')[0] + '.txt')
-                                             data_line.split('.')[0] + '.txt')
-                else:
-                    poly_path = os.path.join(data_path, 'poly',
-                                             data_line.split('.')[0] + '.txt')
                text_polys, text_tags, text_strs = self.extract_polys(poly_path)
            else:
                image_dir = os.path.join(os.path.dirname(data_path), 'image')
                im_path, text_polys, text_tags, text_strs = self.extract_info_textnet(
                    data_line, image_dir)
+            img_id = int(data_line.split(".")[0][3:])
            data = {
                'img_path': im_path,
                'polys': text_polys,
                'tags': text_tags,
-                'strs': text_strs
+                'strs': text_strs,
+                'img_id': img_id
            }
            with open(data['img_path'], 'rb') as f:
                img = f.read()

--- a/ppocr/metrics/e2e_metric.py
+++ b/ppocr/metrics/e2e_metric.py
@@ -19,58 +19,29 @@ from __future__ import print_function
 __all__ = ['E2EMetric']
 from ppocr.utils.e2e_metric.Deteval import get_socre, combine_results
-from ppocr.utils.e2e_utils.extract_textpoint import get_dict
+from ppocr.utils.e2e_utils.extract_textpoint_slow import get_dict
 class E2EMetric(object):
    def __init__(self,
+                 gt_mat_dir,
                 character_dict_path,
                 main_indicator='f_score_e2e',
                 **kwargs):
+        self.gt_mat_dir = gt_mat_dir
        self.label_list = get_dict(character_dict_path)
        self.max_index = len(self.label_list)
        self.main_indicator = main_indicator
        self.reset()
    def __call__(self, preds, batch, **kwargs):
-        temp_gt_polyons_batch = batch[2]
+        img_id = batch[5][0]
-        temp_gt_strs_batch = batch[3]
+        e2e_info_list = [{
-        ignore_tags_batch = batch[4]
+            'points': det_polyon,
-        gt_polyons_batch = []
+            'text': pred_str
-        gt_strs_batch = []
+        } for det_polyon, pred_str in zip(preds['points'], preds['strs'])]
+        result = get_socre(self.gt_mat_dir, img_id, e2e_info_list)
-        temp_gt_polyons_batch = temp_gt_polyons_batch[0].tolist()
+        self.results.append(result)
-        for temp_list in temp_gt_polyons_batch:
-            t = []
-            for index in temp_list:
-                if index[0] != -1 and index[1] != -1:
-                    t.append(index)
-            gt_polyons_batch.append(t)
-        temp_gt_strs_batch = temp_gt_strs_batch[0].tolist()
-        for temp_list in temp_gt_strs_batch:
-            t = ""
-            for index in temp_list:
-                if index < self.max_index:
-                    t += self.label_list[index]
-            gt_strs_batch.append(t)
-        for pred, gt_polyons, gt_strs, ignore_tags in zip(
-            [preds], [gt_polyons_batch], [gt_strs_batch], ignore_tags_batch):
-            # prepare gt
-            gt_info_list = [{
-                'points': gt_polyon,
-                'text': gt_str,
-                'ignore': ignore_tag
-            } for gt_polyon, gt_str, ignore_tag in
-                            zip(gt_polyons, gt_strs, ignore_tags)]
-            # prepare det
-            e2e_info_list = [{
-                'points': det_polyon,
-                'text': pred_str
-            } for det_polyon, pred_str in zip(pred['points'], pred['strs'])]
-            result = get_socre(gt_info_list, e2e_info_list)
-            self.results.append(result)
    def get_metric(self):
        metircs = combine_results(self.results)

--- a/ppocr/postprocess/pg_postprocess.py
+++ b/ppocr/postprocess/pg_postprocess.py
@@ -22,10 +22,7 @@ import sys
 __dir__ = os.path.dirname(__file__)
 sys.path.append(__dir__)
 sys.path.append(os.path.join(__dir__, '..'))
+from ppocr.utils.e2e_utils.pgnet_pp_utils import PGNet_PostProcess
-from ppocr.utils.e2e_utils.extract_textpoint import *
-from ppocr.utils.e2e_utils.visual import *
-import paddle
 class PGPostProcess(object):
@@ -33,10 +30,12 @@ class PGPostProcess(object):
    The post process for PGNet.
    """
-    def __init__(self, character_dict_path, valid_set, score_thresh, **kwargs):
+    def __init__(self, character_dict_path, valid_set, score_thresh, mode,
-        self.Lexicon_Table = get_dict(character_dict_path)
+                 **kwargs):
+        self.character_dict_path = character_dict_path
        self.valid_set = valid_set
        self.score_thresh = score_thresh
+        self.mode = mode
        # c++ la-nms is faster, but only support python 3.5
        self.is_python35 = False
@@ -44,112 +43,10 @@ class PGPostProcess(object):
            self.is_python35 = True
    def __call__(self, outs_dict, shape_list):
-        p_score = outs_dict['f_score']
+        post = PGNet_PostProcess(self.character_dict_path, self.valid_set,
-        p_border = outs_dict['f_border']
+                                 self.score_thresh, outs_dict, shape_list)
-        p_char = outs_dict['f_char']
+        if self.mode == 'fast':
-        p_direction = outs_dict['f_direction']
+            data = post.pg_postprocess_fast()
-        if isinstance(p_score, paddle.Tensor):
-            p_score = p_score[0].numpy()
-            p_border = p_border[0].numpy()
-            p_direction = p_direction[0].numpy()
-            p_char = p_char[0].numpy()
        else:
-            p_score = p_score[0]
+            data = post.pg_postprocess_slow()
-            p_border = p_border[0]
-            p_direction = p_direction[0]
-            p_char = p_char[0]
-        src_h, src_w, ratio_h, ratio_w = shape_list[0]
-        is_curved = self.valid_set == "totaltext"
-        instance_yxs_list = generate_pivot_list(
-            p_score,
-            p_char,
-            p_direction,
-            score_thresh=self.score_thresh,
-            is_backbone=True,
-            is_curved=is_curved)
-        p_char = paddle.to_tensor(np.expand_dims(p_char, axis=0))
-        char_seq_idx_set = []
-        for i in range(len(instance_yxs_list)):
-            gather_info_lod = paddle.to_tensor(instance_yxs_list[i])
-            f_char_map = paddle.transpose(p_char, [0, 2, 3, 1])
-            feature_seq = paddle.gather_nd(f_char_map, gather_info_lod)
-            feature_seq = np.expand_dims(feature_seq.numpy(), axis=0)
-            feature_len = [len(feature_seq[0])]
-            featyre_seq = paddle.to_tensor(feature_seq)
-            feature_len = np.array([feature_len]).astype(np.int64)
-            length = paddle.to_tensor(feature_len)
-            seq_pred = paddle.fluid.layers.ctc_greedy_decoder(
-                input=featyre_seq, blank=36, input_length=length)
-            seq_pred_str = seq_pred[0].numpy().tolist()[0]
-            seq_len = seq_pred[1].numpy()[0][0]
-            temp_t = []
-            for c in seq_pred_str[:seq_len]:
-                temp_t.append(c)
-            char_seq_idx_set.append(temp_t)
-        seq_strs = []
-        for char_idx_set in char_seq_idx_set:
-            pr_str = ''.join([self.Lexicon_Table[pos] for pos in char_idx_set])
-            seq_strs.append(pr_str)
-        poly_list = []
-        keep_str_list = []
-        all_point_list = []
-        all_point_pair_list = []
-        for yx_center_line, keep_str in zip(instance_yxs_list, seq_strs):
-            if len(yx_center_line) == 1:
-                yx_center_line.append(yx_center_line[-1])
-            offset_expand = 1.0
-            if self.valid_set == 'totaltext':
-                offset_expand = 1.2
-            point_pair_list = []
-            for batch_id, y, x in yx_center_line:
-                offset = p_border[:, y, x].reshape(2, 2)
-                if offset_expand != 1.0:
-                    offset_length = np.linalg.norm(
-                        offset, axis=1, keepdims=True)
-                    expand_length = np.clip(
-                        offset_length * (offset_expand - 1),
-                        a_min=0.5,
-                        a_max=3.0)
-                    offset_detal = offset / offset_length * expand_length
-                    offset = offset + offset_detal
-                ori_yx = np.array([y, x], dtype=np.float32)
-                point_pair = (ori_yx + offset)[:, ::-1] * 4.0 / np.array(
-                    [ratio_w, ratio_h]).reshape(-1, 2)
-                point_pair_list.append(point_pair)
-                all_point_list.append([
-                    int(round(x * 4.0 / ratio_w)),
-                    int(round(y * 4.0 / ratio_h))
-                ])
-                all_point_pair_list.append(point_pair.round().astype(np.int32)
-                                           .tolist())
-            detected_poly, pair_length_info = point_pair2poly(point_pair_list)
-            detected_poly = expand_poly_along_width(
-                detected_poly, shrink_ratio_of_width=0.2)
-            detected_poly[:, 0] = np.clip(
-                detected_poly[:, 0], a_min=0, a_max=src_w)
-            detected_poly[:, 1] = np.clip(
-                detected_poly[:, 1], a_min=0, a_max=src_h)
-            if len(keep_str) < 2:
-                continue
-            keep_str_list.append(keep_str)
-            if self.valid_set == 'partvgg':
-                middle_point = len(detected_poly) // 2
-                detected_poly = detected_poly[
-                    [0, middle_point - 1, middle_point, -1], :]
-                poly_list.append(detected_poly)
-            elif self.valid_set == 'totaltext':
-                poly_list.append(detected_poly)
-            else:
-                print('--> Not supported format.')
-                exit(-1)
-        data = {
-            'points': poly_list,
-            'strs': keep_str_list,
-        }
        return data
--- a/ppocr/postprocess/rec_postprocess.py
+++ b/ppocr/postprocess/rec_postprocess.py
@@ -28,7 +28,7 @@ class BaseRecLabelDecode(object):
            'ch', 'en', 'EN_symbol', 'french', 'german', 'japan', 'korean',
            'it', 'xi', 'pu', 'ru', 'ar', 'ta', 'ug', 'fa', 'ur', 'rs', 'oc',
            'rsc', 'bg', 'uk', 'be', 'te', 'ka', 'chinese_cht', 'hi', 'mr',
-            'ne', 'EN'
+            'ne', 'EN', 'latin', 'arabic', 'cyrillic', 'devanagari'
        ]
        assert character_type in support_character_type, "Only {} are supported now but get {}".format(
            support_character_type, character_type)