提交 c1cc81c8 编写于 作者: L LDOUBLEV

Merge branch 'release/2.3' of https://github.com/PaddlePaddle/PaddleOCR into update_doc_2.3

...@@ -4,13 +4,10 @@ ...@@ -4,13 +4,10 @@
- [1. 文本检测模型推理](#文本检测模型推理) - [1. 文本检测模型推理](#文本检测模型推理)
- [2. 文本识别模型推理](#文本识别模型推理) - [2. 文本识别模型推理](#文本识别模型推理)
- [2.1 超轻量中文识别模型推理](#超轻量中文识别模型推理) - [2.1 超轻量中文识别模型推理](#超轻量中文识别模型推理)
- [2.2 多语言模型的推理](#多语言模型的推理) - [2.2 多语言模型的推理](#多语言模型的推理)
- [3. 方向分类模型推理](#方向分类模型推理) - [3. 方向分类模型推理](#方向分类模型推理)
- [4. 文本检测、方向分类和文字识别串联推理](#文本检测、方向分类和文字识别串联推理) - [4. 文本检测、方向分类和文字识别串联推理](#文本检测、方向分类和文字识别串联推理)
<a name="文本检测模型推理"></a> <a name="文本检测模型推理"></a>
......
## OCR模型列表(V2.1,2021年9月6日更新) # OCR模型列表(V2.1,2021年9月6日更新)
> **说明** > **说明**
> 1. 2.1版模型相比2.0版模型,2.1的模型在模型精度上做了提升 > 1. 2.1版模型相比2.0版模型,2.1的模型在模型精度上做了提升
...@@ -6,13 +6,13 @@ ...@@ -6,13 +6,13 @@
> 3. 本文档提供的是PPOCR自研模型列表,更多基于公开数据集的算法介绍与预训练模型可以参考:[算法概览文档](./algorithm_overview.md)。 > 3. 本文档提供的是PPOCR自研模型列表,更多基于公开数据集的算法介绍与预训练模型可以参考:[算法概览文档](./algorithm_overview.md)。
- [一、文本检测模型](#文本检测模型) - [1. 文本检测模型](#文本检测模型)
- [二、文本识别模型](#文本识别模型) - [2. 文本识别模型](#文本识别模型)
- [1. 中文识别模型](#中文识别模型) - [2.1 中文识别模型](#中文识别模型)
- [2. 英文识别模型](#英文识别模型) - [2.2 英文识别模型](#英文识别模型)
- [3. 多语言识别模型](#多语言识别模型) - [2.3 多语言识别模型](#多语言识别模型)
- [三、文本方向分类模型](#文本方向分类模型) - [3. 文本方向分类模型](#文本方向分类模型)
- [四、Paddle-Lite 模型](#Paddle-Lite模型) - [4. Paddle-Lite 模型](#Paddle-Lite模型)
PaddleOCR提供的可下载模型包括`推理模型``训练模型``预训练模型``slim模型`,模型区别说明如下: PaddleOCR提供的可下载模型包括`推理模型``训练模型``预训练模型``slim模型`,模型区别说明如下:
...@@ -29,7 +29,7 @@ PaddleOCR提供的可下载模型包括`推理模型`、`训练模型`、`预训 ...@@ -29,7 +29,7 @@ PaddleOCR提供的可下载模型包括`推理模型`、`训练模型`、`预训
<a name="文本检测模型"></a> <a name="文本检测模型"></a>
### 一、文本检测模型 ## 1. 文本检测模型
|模型名称|模型简介|配置文件|推理模型大小|下载地址| |模型名称|模型简介|配置文件|推理模型大小|下载地址|
| --- | --- | --- | --- | --- | | --- | --- | --- | --- | --- |
...@@ -41,10 +41,11 @@ PaddleOCR提供的可下载模型包括`推理模型`、`训练模型`、`预训 ...@@ -41,10 +41,11 @@ PaddleOCR提供的可下载模型包括`推理模型`、`训练模型`、`预训
<a name="文本识别模型"></a> <a name="文本识别模型"></a>
### 二、文本识别模型 ## 2. 文本识别模型
<a name="中文识别模型"></a> <a name="中文识别模型"></a>
#### 1. 中文识别模型
### 2.1 中文识别模型
|模型名称|模型简介|配置文件|推理模型大小|下载地址| |模型名称|模型简介|配置文件|推理模型大小|下载地址|
| --- | --- | --- | --- | --- | | --- | --- | --- | --- | --- |
...@@ -57,7 +58,7 @@ PaddleOCR提供的可下载模型包括`推理模型`、`训练模型`、`预训 ...@@ -57,7 +58,7 @@ PaddleOCR提供的可下载模型包括`推理模型`、`训练模型`、`预训
**说明:** `训练模型`是基于预训练模型在真实数据与竖排合成文本数据上finetune得到的模型,在真实应用场景中有着更好的表现,`预训练模型`则是直接基于全量真实数据与合成数据训练得到,更适合用于在自己的数据集上finetune。 **说明:** `训练模型`是基于预训练模型在真实数据与竖排合成文本数据上finetune得到的模型,在真实应用场景中有着更好的表现,`预训练模型`则是直接基于全量真实数据与合成数据训练得到,更适合用于在自己的数据集上finetune。
<a name="英文识别模型"></a> <a name="英文识别模型"></a>
#### 2. 英文识别模型 ### 2.2 英文识别模型
|模型名称|模型简介|配置文件|推理模型大小|下载地址| |模型名称|模型简介|配置文件|推理模型大小|下载地址|
| --- | --- | --- | --- | --- | | --- | --- | --- | --- | --- |
...@@ -65,7 +66,7 @@ PaddleOCR提供的可下载模型包括`推理模型`、`训练模型`、`预训 ...@@ -65,7 +66,7 @@ PaddleOCR提供的可下载模型包括`推理模型`、`训练模型`、`预训
|en_number_mobile_v2.0_rec|原始超轻量模型,支持英文、数字识别|[rec_en_number_lite_train.yml](../../configs/rec/multi_language/rec_en_number_lite_train.yml)|2.6M|[推理模型](https://paddleocr.bj.bcebos.com/dygraph_v2.0/multilingual/en_number_mobile_v2.0_rec_infer.tar) / [训练模型](https://paddleocr.bj.bcebos.com/dygraph_v2.0/multilingual/en_number_mobile_v2.0_rec_train.tar) | |en_number_mobile_v2.0_rec|原始超轻量模型,支持英文、数字识别|[rec_en_number_lite_train.yml](../../configs/rec/multi_language/rec_en_number_lite_train.yml)|2.6M|[推理模型](https://paddleocr.bj.bcebos.com/dygraph_v2.0/multilingual/en_number_mobile_v2.0_rec_infer.tar) / [训练模型](https://paddleocr.bj.bcebos.com/dygraph_v2.0/multilingual/en_number_mobile_v2.0_rec_train.tar) |
<a name="多语言识别模型"></a> <a name="多语言识别模型"></a>
#### 3. 多语言识别模型(更多语言持续更新中...) ### 2.3 多语言识别模型(更多语言持续更新中...)
|模型名称|字典文件|模型简介|配置文件|推理模型大小|下载地址| |模型名称|字典文件|模型简介|配置文件|推理模型大小|下载地址|
| --- | --- | --- | --- |--- | --- | | --- | --- | --- | --- |--- | --- |
...@@ -86,7 +87,7 @@ PaddleOCR提供的可下载模型包括`推理模型`、`训练模型`、`预训 ...@@ -86,7 +87,7 @@ PaddleOCR提供的可下载模型包括`推理模型`、`训练模型`、`预训
<a name="文本方向分类模型"></a> <a name="文本方向分类模型"></a>
### 三、文本方向分类模型 ## 3. 文本方向分类模型
|模型名称|模型简介|配置文件|推理模型大小|下载地址| |模型名称|模型简介|配置文件|推理模型大小|下载地址|
| --- | --- | --- | --- | --- | | --- | --- | --- | --- | --- |
...@@ -95,7 +96,7 @@ PaddleOCR提供的可下载模型包括`推理模型`、`训练模型`、`预训 ...@@ -95,7 +96,7 @@ PaddleOCR提供的可下载模型包括`推理模型`、`训练模型`、`预训
<a name="Paddle-Lite模型"></a> <a name="Paddle-Lite模型"></a>
### 四、Paddle-Lite 模型 ## 4. Paddle-Lite 模型
|模型版本|模型简介|模型大小|检测模型|文本方向分类模型|识别模型|Paddle-Lite版本| |模型版本|模型简介|模型大小|检测模型|文本方向分类模型|识别模型|Paddle-Lite版本|
|---|---|---|---|---|---|---| |---|---|---|---|---|---|---|
......
...@@ -5,13 +5,10 @@ This article introduces the use of the Python inference engine for the PP-OCR mo ...@@ -5,13 +5,10 @@ This article introduces the use of the Python inference engine for the PP-OCR mo
- [Text Detection Model Inference](#DETECTION_MODEL_INFERENCE) - [Text Detection Model Inference](#DETECTION_MODEL_INFERENCE)
- [Text Recognition Model Inference](#RECOGNITION_MODEL_INFERENCE) - [Text Recognition Model Inference](#RECOGNITION_MODEL_INFERENCE)
- [1. Lightweight Chinese Recognition Model Inference](#LIGHTWEIGHT_RECOGNITION) - [1. Lightweight Chinese Recognition Model Inference](#LIGHTWEIGHT_RECOGNITION)
- [2. Multilingaul Model Inference](#MULTILINGUAL_MODEL_INFERENCE) - [2. Multilingaul Model Inference](#MULTILINGUAL_MODEL_INFERENCE)
- [Angle Classification Model Inference](#ANGLE_CLASS_MODEL_INFERENCE) - [Angle Classification Model Inference](#ANGLE_CLASS_MODEL_INFERENCE)
- [Text Detection Angle Classification and Recognition Inference Concatenation](#CONCATENATION) - [Text Detection Angle Classification and Recognition Inference Concatenation](#CONCATENATION)
<a name="DETECTION_MODEL_INFERENCE"></a> <a name="DETECTION_MODEL_INFERENCE"></a>
......
## OCR model list(V2.1, updated on 2021.9.6) # OCR Model List(V2.1, updated on 2021.9.6)
> **Note** > **Note**
> 1. Compared with the model v2.0, the 2.1 version of the detection model has a improvement in accuracy, and the 2.1 version of the recognition model is optimized in accuracy and CPU speed. > 1. Compared with the model v2.0, the 2.1 version of the detection model has a improvement in accuracy, and the 2.1 version of the recognition model is optimized in accuracy and CPU speed.
> 2. Compared with [models 1.1](https://github.com/PaddlePaddle/PaddleOCR/blob/develop/doc/doc_en/models_list_en.md), which are trained with static graph programming paradigm, models 2.0 are the dynamic graph trained version and achieve close performance. > 2. Compared with [models 1.1](https://github.com/PaddlePaddle/PaddleOCR/blob/develop/doc/doc_en/models_list_en.md), which are trained with static graph programming paradigm, models 2.0 are the dynamic graph trained version and achieve close performance.
...@@ -6,9 +6,9 @@ ...@@ -6,9 +6,9 @@
- [1. Text Detection Model](#Detection) - [1. Text Detection Model](#Detection)
- [2. Text Recognition Model](#Recognition) - [2. Text Recognition Model](#Recognition)
- [Chinese Recognition Model](#Chinese) - [2.1 Chinese Recognition Model](#Chinese)
- [English Recognition Model](#English) - [2.2 English Recognition Model](#English)
- [Multilingual Recognition Model](#Multilingual) - [2.3 Multilingual Recognition Model](#Multilingual)
- [3. Text Angle Classification Model](#Angle) - [3. Text Angle Classification Model](#Angle)
- [4. Paddle-Lite Model](#Paddle-Lite) - [4. Paddle-Lite Model](#Paddle-Lite)
...@@ -25,7 +25,7 @@ Relationship of the above models is as follows. ...@@ -25,7 +25,7 @@ Relationship of the above models is as follows.
![](../imgs_en/model_prod_flow_en.png) ![](../imgs_en/model_prod_flow_en.png)
<a name="Detection"></a> <a name="Detection"></a>
### 1. Text Detection Model ## 1. Text Detection Model
|model name|description|config|model size|download| |model name|description|config|model size|download|
| --- | --- | --- | --- | --- | | --- | --- | --- | --- | --- |
...@@ -36,10 +36,10 @@ Relationship of the above models is as follows. ...@@ -36,10 +36,10 @@ Relationship of the above models is as follows.
|ch_ppocr_server_v2.0_det|General model, which is larger than the lightweight model, but achieved better performance|[ch_det_res18_db_v2.0.yml](../../configs/det/ch_ppocr_v2.0/ch_det_res18_db_v2.0.yml)|47M|[inference model](https://paddleocr.bj.bcebos.com/dygraph_v2.0/ch/ch_ppocr_server_v2.0_det_infer.tar) / [trained model](https://paddleocr.bj.bcebos.com/dygraph_v2.0/ch/ch_ppocr_server_v2.0_det_train.tar)| |ch_ppocr_server_v2.0_det|General model, which is larger than the lightweight model, but achieved better performance|[ch_det_res18_db_v2.0.yml](../../configs/det/ch_ppocr_v2.0/ch_det_res18_db_v2.0.yml)|47M|[inference model](https://paddleocr.bj.bcebos.com/dygraph_v2.0/ch/ch_ppocr_server_v2.0_det_infer.tar) / [trained model](https://paddleocr.bj.bcebos.com/dygraph_v2.0/ch/ch_ppocr_server_v2.0_det_train.tar)|
<a name="Recognition"></a> <a name="Recognition"></a>
### 2. Text Recognition Model ## 2. Text Recognition Model
<a name="Chinese"></a> <a name="Chinese"></a>
#### Chinese Recognition Model ### 2.1 Chinese Recognition Model
|model name|description|config|model size|download| |model name|description|config|model size|download|
| --- | --- | --- | --- | --- | | --- | --- | --- | --- | --- |
...@@ -53,7 +53,7 @@ Relationship of the above models is as follows. ...@@ -53,7 +53,7 @@ Relationship of the above models is as follows.
**Note:** The `trained model` is finetuned on the `pre-trained model` with real data and synthsized vertical text data, which achieved better performance in real scene. The `pre-trained model` is directly trained on the full amount of real data and synthsized data, which is more suitable for finetune on your own dataset. **Note:** The `trained model` is finetuned on the `pre-trained model` with real data and synthsized vertical text data, which achieved better performance in real scene. The `pre-trained model` is directly trained on the full amount of real data and synthsized data, which is more suitable for finetune on your own dataset.
<a name="English"></a> <a name="English"></a>
#### English Recognition Model ### 2.2 English Recognition Model
|model name|description|config|model size|download| |model name|description|config|model size|download|
| --- | --- | --- | --- | --- | | --- | --- | --- | --- | --- |
...@@ -61,7 +61,7 @@ Relationship of the above models is as follows. ...@@ -61,7 +61,7 @@ Relationship of the above models is as follows.
|en_number_mobile_v2.0_rec|Original lightweight model, supporting English and number recognition|[rec_en_number_lite_train.yml](../../configs/rec/multi_language/rec_en_number_lite_train.yml)|2.6M|[inference model](https://paddleocr.bj.bcebos.com/dygraph_v2.0/multilingual/en_number_mobile_v2.0_rec_infer.tar) / [trained model](https://paddleocr.bj.bcebos.com/dygraph_v2.0/multilingual/en_number_mobile_v2.0_rec_train.tar) | |en_number_mobile_v2.0_rec|Original lightweight model, supporting English and number recognition|[rec_en_number_lite_train.yml](../../configs/rec/multi_language/rec_en_number_lite_train.yml)|2.6M|[inference model](https://paddleocr.bj.bcebos.com/dygraph_v2.0/multilingual/en_number_mobile_v2.0_rec_infer.tar) / [trained model](https://paddleocr.bj.bcebos.com/dygraph_v2.0/multilingual/en_number_mobile_v2.0_rec_train.tar) |
<a name="Multilingual"></a> <a name="Multilingual"></a>
#### Multilingual Recognition Model(Updating...) ### 2.3 Multilingual Recognition Model(Updating...)
|model name| dict file | description|config|model size|download| |model name| dict file | description|config|model size|download|
| --- | --- | --- |--- | --- | --- | | --- | --- | --- |--- | --- | --- |
...@@ -82,7 +82,7 @@ For more supported languages, please refer to : [Multi-language model](./multi_l ...@@ -82,7 +82,7 @@ For more supported languages, please refer to : [Multi-language model](./multi_l
<a name="Angle"></a> <a name="Angle"></a>
### 3. Text Angle Classification Model ## 3. Text Angle Classification Model
|model name|description|config|model size|download| |model name|description|config|model size|download|
| --- | --- | --- | --- | --- | | --- | --- | --- | --- | --- |
...@@ -90,7 +90,7 @@ For more supported languages, please refer to : [Multi-language model](./multi_l ...@@ -90,7 +90,7 @@ For more supported languages, please refer to : [Multi-language model](./multi_l
|ch_ppocr_mobile_v2.0_cls|Original model for text angle classification|[cls_mv3.yml](../../configs/cls/cls_mv3.yml)|1.38M|[inference model](https://paddleocr.bj.bcebos.com/dygraph_v2.0/ch/ch_ppocr_mobile_v2.0_cls_infer.tar) / [trained model](https://paddleocr.bj.bcebos.com/dygraph_v2.0/ch/ch_ppocr_mobile_v2.0_cls_train.tar) | |ch_ppocr_mobile_v2.0_cls|Original model for text angle classification|[cls_mv3.yml](../../configs/cls/cls_mv3.yml)|1.38M|[inference model](https://paddleocr.bj.bcebos.com/dygraph_v2.0/ch/ch_ppocr_mobile_v2.0_cls_infer.tar) / [trained model](https://paddleocr.bj.bcebos.com/dygraph_v2.0/ch/ch_ppocr_mobile_v2.0_cls_train.tar) |
<a name="Paddle-Lite"></a> <a name="Paddle-Lite"></a>
### 4. Paddle-Lite Model ## 4. Paddle-Lite Model
|Version|Introduction|Model size|Detection model|Text Direction model|Recognition model|Paddle-Lite branch| |Version|Introduction|Model size|Detection model|Text Direction model|Recognition model|Paddle-Lite branch|
|---|---|---|---|---|---|---| |---|---|---|---|---|---|---|
|PP-OCRv2|extra-lightweight chinese OCR optimized model|11M|[download link](https://paddleocr.bj.bcebos.com/PP-OCRv2/chinese/ch_PP-OCRv2_det_infer_opt.nb)|[download link](https://paddleocr.bj.bcebos.com/dygraph_v2.0/lite/ch_ppocr_mobile_v2.0_cls_opt.nb)|[download link](https://paddleocr.bj.bcebos.com/PP-OCRv2/chinese/ch_PP-OCRv2_rec_infer_opt.nb)|v2.9| |PP-OCRv2|extra-lightweight chinese OCR optimized model|11M|[download link](https://paddleocr.bj.bcebos.com/PP-OCRv2/chinese/ch_PP-OCRv2_det_infer_opt.nb)|[download link](https://paddleocr.bj.bcebos.com/dygraph_v2.0/lite/ch_ppocr_mobile_v2.0_cls_opt.nb)|[download link](https://paddleocr.bj.bcebos.com/PP-OCRv2/chinese/ch_PP-OCRv2_rec_infer_opt.nb)|v2.9|
......
Markdown is supported
0% .
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