Merge pull request #6078 from andyjpaddle/dygraph

update doc for rec

doc/doc_ch/algorithm_rec_sar.md

+# SAR
+
+- [1. 算法简介](#1)
+- [2. 环境配置](#2)
+- [3. 模型训练、评估、预测](#3)
+    - [3.1 训练](#3-1)
+    - [3.2 评估](#3-2)
+    - [3.3 预测](#3-3)
+- [4. 推理部署](#4)
+    - [4.1 Python推理](#4-1)
+    - [4.2 C++推理](#4-2)
+    - [4.3 Serving服务化部署](#4-3)
+    - [4.4 更多推理部署](#4-4)
+- [5. FAQ](#5)
+
+<a name="1"></a>
+## 1. 算法简介
+
+论文信息：
+> [Show, Attend and Read: A Simple and Strong Baseline for Irregular Text Recognition](https://arxiv.org/abs/1811.00751)
+> Hui Li, Peng Wang, Chunhua Shen, Guyu Zhang
+> AAAI, 2019
+
+使用MJSynth和SynthText两个文字识别数据集训练，在IIIT, SVT, IC03, IC13, IC15, SVTP, CUTE数据集上进行评估，算法复现效果如下：
+
+|模型|骨干网络|配置文件|Acc|下载链接|
+| --- | --- | --- | --- | --- | --- | --- |
+|SAR|ResNet31|[rec_r31_sar.yml](../../configs/rec/rec_r31_sar.yml)|87.20%|[训练模型](https://paddleocr.bj.bcebos.com/dygraph_v2.1/rec/rec_r31_sar_train.tar)|
+
+注：除了使用MJSynth和SynthText两个文字识别数据集外，还加入了[SynthAdd](https://pan.baidu.com/share/init?surl=uV0LtoNmcxbO-0YA7Ch4dg)数据（提取码：627x），和部分真实数据，具体数据细节可以参考论文。
+
+<a name="2"></a>
+## 2. 环境配置
+请先参考[《运行环境准备》](./environment.md)配置PaddleOCR运行环境，参考[《项目克隆》](./clone.md)克隆项目代码。
+
+
+<a name="3"></a>
+## 3. 模型训练、评估、预测
+
+请参考[文本识别教程](./recognition.md)。PaddleOCR对代码进行了模块化，训练不同的识别模型只需要**更换配置文件**即可。
+
+训练
+
+具体地，在完成数据准备后，便可以启动训练，训练命令如下：
+
+```
+#单卡训练（训练周期长，不建议）
+python3 tools/train.py -c configs/rec/rec_r31_sar.yml
+
+#多卡训练，通过--gpus参数指定卡号
+python3 -m paddle.distributed.launch --gpus '0,1,2,3'  tools/train.py -c configs/rec/rec_r31_sar.yml
+```
+
+评估
+
+```
+# GPU 评估， Global.pretrained_model 为待测权重
+python3 -m paddle.distributed.launch --gpus '0' tools/eval.py -c configs/rec/rec_r31_sar.yml -o Global.pretrained_model={path/to/weights}/best_accuracy
+```
+
+预测：
+
+```
+# 预测使用的配置文件必须与训练一致
+python3 tools/infer_rec.py -c configs/rec/rec_r31_sar.yml -o Global.pretrained_model={path/to/weights}/best_accuracy Global.infer_img=doc/imgs_words/en/word_1.png
+```
+
+<a name="4"></a>
+## 4. 推理部署
+
+<a name="4-1"></a>
+### 4.1 Python推理
+首先将SAR文本识别训练过程中保存的模型，转换成inference model。（ [模型下载地址](https://paddleocr.bj.bcebos.com/dygraph_v2.1/rec/rec_r31_sar_train.tar) )，可以使用如下命令进行转换：
+
+```
+python3 tools/export_model.py -c configs/rec/rec_r31_sar.yml -o Global.pretrained_model=./rec_r31_sar_train/best_accuracy  Global.save_inference_dir=./inference/rec_sar
+```
+
+SAR文本识别模型推理，可以执行如下命令：
+
+```
+python3 tools/infer/predict_rec.py --image_dir="./doc/imgs_words/en/word_1.png" --rec_model_dir="./inference/rec_sar/" --rec_image_shape="3, 48, 48, 160" --rec_char_type="ch" --rec_algorithm="SAR" --rec_char_dict_path="ppocr/utils/dict90.txt" --max_text_length=30 --use_space_char=False
+```
+
+<a name="4-2"></a>
+### 4.2 C++推理
+
+由于C++预处理后处理还未支持SAR，所以暂未支持
+
+<a name="4-3"></a>
+### 4.3 Serving服务化部署
+
+暂不支持
+
+<a name="4-4"></a>
+### 4.4 更多推理部署
+
+暂不支持
+
+<a name="5"></a>
+## 5. FAQ
+
+
+## 引用
+
+```bibtex
+@article{Li2019ShowAA,
+  title={Show, Attend and Read: A Simple and Strong Baseline for Irregular Text Recognition},
+  author={Hui Li and Peng Wang and Chunhua Shen and Guyu Zhang},
+  journal={ArXiv},
+  year={2019},
+  volume={abs/1811.00751}
+}
+```

doc/doc_ch/algorithm_rec_srn.md

+# SRN
+
+- [1. 算法简介](#1)
+- [2. 环境配置](#2)
+- [3. 模型训练、评估、预测](#3)
+    - [3.1 训练](#3-1)
+    - [3.2 评估](#3-2)
+    - [3.3 预测](#3-3)
+- [4. 推理部署](#4)
+    - [4.1 Python推理](#4-1)
+    - [4.2 C++推理](#4-2)
+    - [4.3 Serving服务化部署](#4-3)
+    - [4.4 更多推理部署](#4-4)
+- [5. FAQ](#5)
+
+<a name="1"></a>
+## 1. 算法简介
+
+论文信息：
+> [Towards Accurate Scene Text Recognition with Semantic Reasoning Networks](https://arxiv.org/abs/2003.12294#)
+> Deli Yu, Xuan Li, Chengquan Zhang, Junyu Han, Jingtuo Liu, Errui Ding
+> CVPR,2020
+
+使用MJSynth和SynthText两个文字识别数据集训练，在IIIT, SVT, IC03, IC13, IC15, SVTP, CUTE数据集上进行评估，算法复现效果如下：
+
+|模型|骨干网络|配置文件|Acc|下载链接|
+| --- | --- | --- | --- | --- | --- | --- |
+|SRN|Resnet50_vd_fpn|[rec_r50_fpn_srn.yml](../../configs/rec/rec_r50_fpn_srn.yml)|86.31%|[训练模型](https://paddleocr.bj.bcebos.com/dygraph_v2.0/en/rec_r50_vd_srn_train.tar)|
+
+
+<a name="2"></a>
+## 2. 环境配置
+请先参考[《运行环境准备》](./environment.md)配置PaddleOCR运行环境，参考[《项目克隆》](./clone.md)克隆项目代码。
+
+
+<a name="3"></a>
+## 3. 模型训练、评估、预测
+
+请参考[文本识别教程](./recognition.md)。PaddleOCR对代码进行了模块化，训练不同的识别模型只需要**更换配置文件**即可。
+
+训练
+
+具体地，在完成数据准备后，便可以启动训练，训练命令如下：
+
+```
+#单卡训练（训练周期长，不建议）
+python3 tools/train.py -c configs/rec/rec_r50_fpn_srn.yml
+
+#多卡训练，通过--gpus参数指定卡号
+python3 -m paddle.distributed.launch --gpus '0,1,2,3'  tools/train.py -c configs/rec/rec_r50_fpn_srn.yml
+```
+
+评估
+
+```
+# GPU 评估， Global.pretrained_model 为待测权重
+python3 -m paddle.distributed.launch --gpus '0' tools/eval.py -c configs/rec/rec_r50_fpn_srn.yml -o Global.pretrained_model={path/to/weights}/best_accuracy
+```
+
+预测：
+
+```
+# 预测使用的配置文件必须与训练一致
+python3 tools/infer_rec.py -c configs/rec/rec_r50_fpn_srn.yml -o Global.pretrained_model={path/to/weights}/best_accuracy Global.infer_img=doc/imgs_words/en/word_1.png
+```
+
+<a name="4"></a>
+## 4. 推理部署
+
+<a name="4-1"></a>
+### 4.1 Python推理
+首先将SRN文本识别训练过程中保存的模型，转换成inference model。（ [模型下载地址](https://paddleocr.bj.bcebos.com/dygraph_v2.0/en/rec_r50_vd_srn_train.tar) )，可以使用如下命令进行转换：
+
+```
+python3 tools/export_model.py -c configs/rec/rec_r50_fpn_srn.yml -o Global.pretrained_model=./rec_r50_vd_srn_train/best_accuracy  Global.save_inference_dir=./inference/rec_srn
+```
+
+SRN文本识别模型推理，可以执行如下命令：
+
+```
+python3 tools/infer/predict_rec.py --image_dir="./doc/imgs_words/en/word_1.png" --rec_model_dir="./inference/rec_srn/" --rec_image_shape="1,64,256" --rec_char_type="ch" --rec_algorithm="SRN" --rec_char_dict_path=./ppocr/utils/ic15_dict.txt  --use_space_char=False
+```
+
+<a name="4-2"></a>
+### 4.2 C++推理
+
+由于C++预处理后处理还未支持SRN，所以暂未支持
+
+<a name="4-3"></a>
+### 4.3 Serving服务化部署
+
+暂不支持
+
+<a name="4-4"></a>
+### 4.4 更多推理部署
+
+暂不支持
+
+<a name="5"></a>
+## 5. FAQ
+
+
+## 引用
+
+```bibtex
+@article{Yu2020TowardsAS,
+  title={Towards Accurate Scene Text Recognition With Semantic Reasoning Networks},
+  author={Deli Yu and Xuan Li and Chengquan Zhang and Junyu Han and Jingtuo Liu and Errui Ding},
+  journal={2020 IEEE/CVF Conference on Computer Vision and Pattern Recognition (CVPR)},
+  year={2020},
+  pages={12110-12119}
+}
+```

doc/doc_ch/detection.md

@@ -167,11 +167,15 @@ PaddleOCR支持了基于知识蒸馏的检测模型训练过程，更多内容
 ## 2.7 其他训练环境
 
 - Windows GPU/CPU
+在Windows平台上与Linux平台略有不同:
+Windows平台只支持`单卡`的训练与预测，指定GPU进行训练`set CUDA_VISIBLE_DEVICES=0`
+在Windows平台，DataLoader只支持单进程模式，因此需要设置 `num_workers` 为0;
 
 - macOS
+不支持GPU模式，需要在配置文件中设置`use_gpu`为False，其余训练评估预测命令与Linux GPU完全相同。
 
 - Linux DCU
-
+DCU设备上运行需要设置环境变量 `export HIP_VISIBLE_DEVICES=0,1,2,3`，其余训练评估预测命令与Linux GPU完全相同。
 
 <a name="3--------"></a>
 # 3. 模型评估与预测

doc/doc_ch/recognition.md

@@ -3,30 +3,138 @@
 本文提供了PaddleOCR文本识别任务的全流程指南，包括数据准备、模型训练、调优、评估、预测，各个阶段的详细说明：
 
 - [1. 数据准备](#1-数据准备)
-  - [1.1 准备数据集](#11-准备数据集)
-  - [1.2 字典](#12-字典)
-  - [1.3 添加空格类别](#13-添加空格类别)
-- [2. 启动训练](#2-启动训练)
-  - [2.1 数据增强](#21-数据增强)
-  - [2.2 通用模型训练](#22-通用模型训练)
-  - [2.3 多语言模型训练](#23-多语言模型训练)
-  - [2.4 知识蒸馏训练](#24-知识蒸馏训练)
-- [3 评估](#3-评估)
-- [4 预测](#4-预测)
-- [5. 转Inference模型测试](#5-转inference模型测试)
-
-
-<a name="数据准备"></a>
-## 1. 数据准备
+  * [1.1 自定义数据集](#11-自定义数据集)
+  * [1.2 数据下载](#12-数据下载)
+  * [1.3 字典](#13-字典)
+  * [1.4 添加空格类别](#14-添加空格类别)
+  * [1.5 数据增强](#15-数据增强)
+- [2. 开始训练](#2-开始训练)
+  * [2.1 启动训练](#21-----)
+  * [2.2 断点训练](#22-----)
+  * [2.3 更换Backbone 训练](#23---backbone---)
+  * [2.4 混合精度训练](#24---amp---)
+  * [2.5 分布式训练](#25---fleet---)
+  * [2.6 知识蒸馏训练](#26---distill---)
+  * [2.7 多语言模型训练](#27-多语言模型训练)
+  * [2.8 其他训练环境（Windows/macOS/Linux DCU）](#28---other---)
+- [3. 模型评估与预测](#3--------)
+  * [3.1 指标评估](#31-----)
+  * [3.2 测试识别效果](#32-------)
+- [4. 模型导出与预测](#4--------)
+- [5. FAQ](#5-faq)
+
+
+<a name="1-数据准备"></a>
+# 1. 数据准备
 
 ### 1.1 准备数据集
 
-准备数据集可参考 [ocr_datasets](./dataset/ocr_datasets.md) 。
+PaddleOCR 支持两种数据格式:
+ - `lmdb` 用于训练以lmdb格式存储的数据集(LMDBDataSet);
+ - `通用数据` 用于训练以文本文件存储的数据集(SimpleDataSet);
+
+训练数据的默认存储路径是 `PaddleOCR/train_data`,如果您的磁盘上已有数据集，只需创建软链接至数据集目录：
+
+```
+# linux and mac os
+ln -sf <path/to/dataset> <path/to/paddle_ocr>/train_data/dataset
+# windows
+mklink /d <path/to/paddle_ocr>/train_data/dataset <path/to/dataset>
+```
+
+<a name="11-自定义数据集"></a>
+## 1.1 自定义数据集
+下面以通用数据集为例， 介绍如何准备数据集：
+
+* 训练集
+
+建议将训练图片放入同一个文件夹，并用一个txt文件（rec_gt_train.txt）记录图片路径和标签，txt文件里的内容如下:
+
+**注意：** txt文件中默认请将图片路径和图片标签用 \t 分割，如用其他方式分割将造成训练报错。
+
+```
+" 图像文件名                 图像标注信息 "
+
+train_data/rec/train/word_001.jpg   简单可依赖
+train_data/rec/train/word_002.jpg   用科技让复杂的世界更简单
+...
+```
+
+最终训练集应有如下文件结构：
+```
+|-train_data
+  |-rec
+    |- rec_gt_train.txt
+    |- train
+        |- word_001.png
+        |- word_002.jpg
+        |- word_003.jpg
+        | ...
+```
+
+除上述单张图像为一行格式之外，PaddleOCR也支持对离线增广后的数据进行训练，为了防止相同样本在同一个batch中被多次采样，我们可以将相同标签对应的图片路径写在一行中，以列表的形式给出，在训练中，PaddleOCR会随机选择列表中的一张图片进行训练。对应地，标注文件的格式如下。
+
+```
+["11.jpg", "12.jpg"]   简单可依赖
+["21.jpg", "22.jpg", "23.jpg"]   用科技让复杂的世界更简单
+3.jpg   ocr
+```
+
+上述示例标注文件中，"11.jpg"和"12.jpg"的标签相同，都是`简单可依赖`，在训练的时候，对于该行标注，会随机选择其中的一张图片进行训练。
+
+
+- 验证集
+
+同训练集类似，验证集也需要提供一个包含所有图片的文件夹（test）和一个rec_gt_test.txt，验证集的结构如下所示：
+
+```
+|-train_data
+  |-rec
+    |- rec_gt_test.txt
+    |- test
+        |- word_001.jpg
+        |- word_002.jpg
+        |- word_003.jpg
+        | ...
+```
+
+<a name="12-数据下载"></a>
+## 1.2 数据下载
+
+- ICDAR2015
+
+若您本地没有数据集，可以在官网下载 [ICDAR2015](http://rrc.cvc.uab.es/?ch=4&com=downloads) 数据，用于快速验证。也可以参考[DTRB](https://github.com/clovaai/deep-text-recognition-benchmark#download-lmdb-dataset-for-traininig-and-evaluation-from-here) ，下载 benchmark 所需的lmdb格式数据集。
 
 如果希望复现SAR的论文指标，需要下载[SynthAdd](https://pan.baidu.com/share/init?surl=uV0LtoNmcxbO-0YA7Ch4dg), 提取码：627x。此外，真实数据集icdar2013, icdar2015, cocotext, IIIT5也作为训练数据的一部分。具体数据细节可以参考论文SAR。
 
-<a name="字典"></a>
-### 1.2 字典
+如果你使用的是icdar2015的公开数据集，PaddleOCR 提供了一份用于训练 ICDAR2015 数据集的标签文件，通过以下方式下载：
+
+```
+# 训练集标签
+wget -P ./train_data/ic15_data  https://paddleocr.bj.bcebos.com/dataset/rec_gt_train.txt
+# 测试集标签
+wget -P ./train_data/ic15_data  https://paddleocr.bj.bcebos.com/dataset/rec_gt_test.txt
+```
+
+PaddleOCR 也提供了数据格式转换脚本，可以将ICDAR官网 label 转换为PaddleOCR支持的数据格式。 数据转换工具在 `ppocr/utils/gen_label.py`, 这里以训练集为例：
+
+```
+# 将官网下载的标签文件转换为 rec_gt_label.txt
+python gen_label.py --mode="rec" --input_path="{path/of/origin/label}" --output_label="rec_gt_label.txt"
+```
+
+数据样式格式如下，(a)为原始图片,(b)为每张图片对应的 Ground Truth 文本文件：
+![](../datasets/icdar_rec.png)
+
+- 多语言数据集
+
+多语言模型的训练数据集均为100w的合成数据，使用了开源合成工具 [text_renderer](https://github.com/Sanster/text_renderer) ，少量的字体可以通过下面两种方式下载。
+* [百度网盘](https://pan.baidu.com/s/1bS_u207Rm7YbY33wOECKDA) 提取码：frgi
+* [google drive](https://drive.google.com/file/d/18cSWX7wXSy4G0tbKJ0d9PuIaiwRLHpjA/view)
+
+
+<a name="13-字典"></a>
+## 1.3 字典
 
 最后需要提供一个字典（{word_dict_name}.txt），使模型在训练时，可以将所有出现的字符映射为字典的索引。
 
@@ -61,9 +169,6 @@ PaddleOCR内置了一部分字典，可以按需使用。
 
 `ppocr/utils/en_dict.txt` 是一个包含96个字符的英文字典
 
-
-
-
 目前的多语言模型仍处在demo阶段，会持续优化模型并补充语种，**非常欢迎您为我们提供其他语言的字典和字体**，
 如您愿意可将字典文件提交至 [dict](../../ppocr/utils/dict)，我们会在Repo中感谢您。
 
@@ -72,16 +177,12 @@ PaddleOCR内置了一部分字典，可以按需使用。
 如需自定义dic文件，请在 `configs/rec/rec_icdar15_train.yml` 中添加 `character_dict_path` 字段, 指向您的字典路径。
 
 <a name="支持空格"></a>
-### 1.3 添加空格类别
+## 1.4 添加空格类别
 
 如果希望支持识别"空格"类别, 请将yml文件中的 `use_space_char` 字段设置为 `True`。
 
-
-<a name="启动训练"></a>
-## 2. 启动训练
-
 <a name="数据增强"></a>
-### 2.1 数据增强
+## 1.5 数据增强
 
 PaddleOCR提供了多种数据增强方式，默认配置文件中已经添加了数据增广。
 
@@ -91,11 +192,14 @@ PaddleOCR提供了多种数据增强方式，默认配置文件中已经添加
 
 *由于OpenCV的兼容性问题，扰动操作暂时只支持Linux*
 
-<a name="通用模型训练"></a>
-### 2.2 通用模型训练
+<a name="开始训练"></a>
+# 2. 开始训练
 
 PaddleOCR提供了训练脚本、评估脚本和预测脚本，本节将以 CRNN 识别模型为例：
 
+<a name="启动训练"></a>
+## 2.1 启动训练
+
 首先下载pretrain model，您可以下载训练好的模型在 icdar2015 数据上进行finetune
 
 ```
@@ -215,8 +319,96 @@ Eval:
 ```
 **注意，预测/评估时的配置文件请务必与训练一致。**
 
-<a name="多语言模型训练"></a>
-### 2.3 多语言模型训练
+
+<a name="断点训练"></a>
+## 2.2 断点训练
+
+如果训练程序中断，如果希望加载训练中断的模型从而恢复训练，可以通过指定Global.checkpoints指定要加载的模型路径：
+```shell
+python3 tools/train.py -c configs/rec/rec_icdar15_train.yml -o Global.checkpoints=./your/trained/model
+```
+
+**注意**：`Global.checkpoints`的优先级高于`Global.pretrained_model`的优先级，即同时指定两个参数时，优先加载`Global.checkpoints`指定的模型，如果`Global.checkpoints`指定的模型路径有误，会加载`Global.pretrained_model`指定的模型。
+
+<a name="23---backbone---"></a>
+## 2.3 更换Backbone 训练
+
+PaddleOCR将网络划分为四部分，分别在[ppocr/modeling](../../ppocr/modeling)下。 进入网络的数据将按照顺序(transforms->backbones->necks->heads)依次通过这四个部分。
+
+```bash
+├── architectures # 网络的组网代码
+├── transforms    # 网络的图像变换模块
+├── backbones     # 网络的特征提取模块
+├── necks         # 网络的特征增强模块
+└── heads         # 网络的输出模块
+```
+如果要更换的Backbone 在PaddleOCR中有对应实现，直接修改配置yml文件中`Backbone`部分的参数即可。
+
+如果要使用新的Backbone，更换backbones的例子如下:
+
+1. 在 [ppocr/modeling/backbones](../../ppocr/modeling/backbones) 文件夹下新建文件，如my_backbone.py。
+2. 在 my_backbone.py 文件内添加相关代码，示例代码如下:
+
+```python
+import paddle
+import paddle.nn as nn
+import paddle.nn.functional as F
+
+
+class MyBackbone(nn.Layer):
+    def __init__(self, *args, **kwargs):
+        super(MyBackbone, self).__init__()
+        # your init code
+        self.conv = nn.xxxx
+
+    def forward(self, inputs):
+        # your network forward
+        y = self.conv(inputs)
+        return y
+```
+
+3. 在 [ppocr/modeling/backbones/\__init\__.py](../../ppocr/modeling/backbones/__init__.py)文件内导入添加的`MyBackbone`模块，然后修改配置文件中Backbone进行配置即可使用，格式如下:
+
+```yaml
+Backbone:
+name: MyBackbone
+args1: args1
+```
+
+**注意**：如果要更换网络的其他模块，可以参考[文档](./add_new_algorithm.md)。
+
+<a name="24---amp---"></a>
+## 2.4 混合精度训练
+
+如果您想进一步加快训练速度，可以使用[自动混合精度训练](https://www.paddlepaddle.org.cn/documentation/docs/zh/guides/01_paddle2.0_introduction/basic_concept/amp_cn.html)， 以单机单卡为例，命令如下：
+
+```shell
+python3 tools/train.py -c configs/rec/rec_icdar15_train.yml \
+     -o Global.pretrained_model=./pretrain_models/rec_mv3_none_bilstm_ctc_v2.0_train \
+     Global.use_amp=True Global.scale_loss=1024.0 Global.use_dynamic_loss_scaling=True
+ ```
+
+<a name="26---fleet---"></a>
+## 2.5 分布式训练
+
+多机多卡训练时，通过 `--ips` 参数设置使用的机器IP地址，通过 `--gpus` 参数设置使用的GPU ID：
+
+```bash
+python3 -m paddle.distributed.launch --ips="xx.xx.xx.xx,xx.xx.xx.xx" --gpus '0,1,2,3' tools/train.py -c configs/rec/rec_icdar15_train.yml \
+     -o Global.pretrained_model=./pretrain_models/rec_mv3_none_bilstm_ctc_v2.0_train
+```
+
+**注意:** 采用多机多卡训练时，需要替换上面命令中的ips值为您机器的地址，机器之间需要能够相互ping通。另外，训练时需要在多个机器上分别启动命令。查看机器ip地址的命令为`ifconfig`。
+
+
+<a name="26---distill---"></a>
+## 2.6 知识蒸馏训练
+
+PaddleOCR支持了基于知识蒸馏的文本识别模型训练过程，更多内容可以参考[知识蒸馏说明文档](./knowledge_distillation.md)。
+
+
+<a name="27-多语言模型训练"></a>
+## 2.7 多语言模型训练
 
 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)。
 
@@ -272,24 +464,36 @@ Eval:
     ...
 ```
 
-<a name="知识蒸馏训练"></a>
+<a name="28---other---"></a>
+## 2.8 其他训练环境
 
-### 2.4 知识蒸馏训练
+- Windows GPU/CPU
+在Windows平台上与Linux平台略有不同:
+Windows平台只支持`单卡`的训练与预测，指定GPU进行训练`set CUDA_VISIBLE_DEVICES=0`
+在Windows平台，DataLoader只支持单进程模式，因此需要设置 `num_workers` 为0;
 
-PaddleOCR支持了基于知识蒸馏的文本识别模型训练过程，更多内容可以参考[知识蒸馏说明文档](./knowledge_distillation.md)。
+- macOS
+不支持GPU模式，需要在配置文件中设置`use_gpu`为False，其余训练评估预测命令与Linux GPU完全相同。
+
+- Linux DCU
+DCU设备上运行需要设置环境变量 `export HIP_VISIBLE_DEVICES=0,1,2,3`，其余训练评估预测命令与Linux GPU完全相同。
+
+
+<a name="3--------"></a>
+# 3. 模型评估与预测
 
-<a name="评估"></a>
-## 3 评估
+<a name="31-----"></a>
+## 3.1 指标评估
 
-评估数据集可以通过 `configs/rec/rec_icdar15_train.yml`  修改Eval中的 `label_file_path` 设置。
+训练中模型参数默认保存在`Global.save_model_dir`目录下。在评估指标时，需要设置`Global.checkpoints`指向保存的参数文件。评估数据集可以通过 `configs/rec/rec_icdar15_train.yml`  修改Eval中的 `label_file_path` 设置。
 
 ```
 # GPU 评估， Global.checkpoints 为待测权重
 python3 -m paddle.distributed.launch --gpus '0' tools/eval.py -c configs/rec/rec_icdar15_train.yml -o Global.checkpoints={path/to/weights}/best_accuracy
 ```
 
-<a name="预测"></a>
-## 4 预测
+<a name="32-------"></a>
+## 3.2 测试识别效果
 
 使用 PaddleOCR 训练好的模型，可以通过以下脚本进行快速预测。
 
@@ -348,9 +552,14 @@ infer_img: doc/imgs_words/ch/word_1.jpg
         result: ('韩国小馆', 0.997218)
 ```
 
-<a name="Inference"></a>
 
-## 5. 转Inference模型测试
+<a name="4--------"></a>
+# 4. 模型导出与预测
+
+inference 模型（`paddle.jit.save`保存的模型）
+一般是模型训练，把模型结构和模型参数保存在文件中的固化模型，多用于预测部署场景。
+训练过程中保存的模型是checkpoints模型，保存的只有模型的参数，多用于恢复训练等。
+与checkpoints模型相比，inference 模型会额外保存模型的结构信息，在预测部署、加速推理上性能优越，灵活方便，适合于实际系统集成。
 
 识别模型转inference模型与检测的方式相同，如下：
 
@@ -381,3 +590,11 @@ python3 tools/export_model.py -c configs/rec/ch_ppocr_v2.0/rec_chinese_lite_trai
   ```
   python3 tools/infer/predict_rec.py --image_dir="./doc/imgs_words_en/word_336.png" --rec_model_dir="./your inference model" --rec_image_shape="3, 32, 100" --rec_char_dict_path="your text dict path"
   ```
+
+
+<a name="5-faq"></a>
+# 5. FAQ
+
+Q1: 训练模型转inference 模型之后预测效果不一致？
+
+**A**：此类问题出现较多，问题多是trained model预测时候的预处理、后处理参数和inference model预测的时候的预处理、后处理参数不一致导致的。可以对比训练使用的配置文件中的预处理、后处理和预测时是否存在差异。

doc/doc_en/algorithm_rec_sar_en.md

+# SAR
+
+- [1. Introduction](#1)
+- [2. Environment](#2)
+- [3. Model Training / Evaluation / Prediction](#3)
+    - [3.1 Training](#3-1)
+    - [3.2 Evaluation](#3-2)
+    - [3.3 Prediction](#3-3)
+- [4. Inference and Deployment](#4)
+    - [4.1 Python Inference](#4-1)
+    - [4.2 C++ Inference](#4-2)
+    - [4.3 Serving](#4-3)
+    - [4.4 More](#4-4)
+- [5. FAQ](#5)
+
+<a name="1"></a>
+## 1. Introduction
+
+Paper:
+> [Show, Attend and Read: A Simple and Strong Baseline for Irregular Text Recognition](https://arxiv.org/abs/1811.00751)
+> Hui Li, Peng Wang, Chunhua Shen, Guyu Zhang
+> AAAI, 2019
+
+Using MJSynth and SynthText two text recognition datasets for training, and evaluating on IIIT, SVT, IC03, IC13, IC15, SVTP, CUTE datasets, the algorithm reproduction effect is as follows:
+
+|Model|Backbone|config|Acc|Download link|
+| --- | --- | --- | --- | --- | --- | --- |
+|SAR|ResNet31|[rec_r31_sar.yml](../../configs/rec/rec_r31_sar.yml)|87.20%|[train model](https://paddleocr.bj.bcebos.com/dygraph_v2.1/rec/rec_r31_sar_train.tar)|
+
+Note:In addition to using the two text recognition datasets MJSynth and SynthText, [SynthAdd](https://pan.baidu.com/share/init?surl=uV0LtoNmcxbO-0YA7Ch4dg) data (extraction code: 627x), and some real data are used in training, the specific data details can refer to the paper.
+
+<a name="2"></a>
+## 2. Environment
+Please refer to ["Environment Preparation"](./environment.md) to configure the PaddleOCR environment, and refer to ["Project Clone"](./clone.md) to clone the project code.
+
+
+<a name="3"></a>
+## 3. Model Training / Evaluation / Prediction
+
+Please refer to [Text Recognition Tutorial](./recognition.md). PaddleOCR modularizes the code, and training different recognition models only requires **changing the configuration file**.
+
+Training:
+
+Specifically, after the data preparation is completed, the training can be started. The training command is as follows:
+
+```
+#Single GPU training (long training period, not recommended)
+python3 tools/train.py -c configs/rec/rec_r31_sar.yml
+
+#Multi GPU training, specify the gpu number through the --gpus parameter
+python3 -m paddle.distributed.launch --gpus '0,1,2,3'  tools/train.py -c configs/rec/rec_r31_sar.yml
+```
+
+Evaluation:
+
+```
+# GPU evaluation
+python3 -m paddle.distributed.launch --gpus '0' tools/eval.py -c configs/rec/rec_r31_sar.yml -o Global.pretrained_model={path/to/weights}/best_accuracy
+```
+
+Prediction:
+
+```
+# The configuration file used for prediction must match the training
+python3 tools/infer_rec.py -c configs/rec/rec_r31_sar.yml -o Global.pretrained_model={path/to/weights}/best_accuracy Global.infer_img=doc/imgs_words/en/word_1.png
+```
+
+<a name="4"></a>
+## 4. Inference and Deployment
+
+<a name="4-1"></a>
+### 4.1 Python Inference
+First, the model saved during the SAR text recognition training process is converted into an inference model. ( [Model download link](https://paddleocr.bj.bcebos.com/dygraph_v2.1/rec/rec_r31_sar_train.tar) ), you can use the following command to convert:
+
+```
+python3 tools/export_model.py -c configs/rec/rec_r31_sar.yml -o Global.pretrained_model=./rec_r31_sar_train/best_accuracy  Global.save_inference_dir=./inference/rec_sar
+```
+
+For SAR text recognition model inference, the following commands can be executed:
+
+```
+python3 tools/infer/predict_rec.py --image_dir="./doc/imgs_words/en/word_1.png" --rec_model_dir="./inference/rec_sar/" --rec_image_shape="3, 48, 48, 160" --rec_char_type="ch" --rec_algorithm="SAR" --rec_char_dict_path="ppocr/utils/dict90.txt" --max_text_length=30 --use_space_char=False
+```
+
+<a name="4-2"></a>
+### 4.2 C++ Inference
+
+Not supported
+
+<a name="4-3"></a>
+### 4.3 Serving
+
+Not supported
+
+<a name="4-4"></a>
+### 4.4 More
+
+Not supported
+
+<a name="5"></a>
+## 5. FAQ
+
+
+## Citation
+
+```bibtex
+@article{Li2019ShowAA,
+  title={Show, Attend and Read: A Simple and Strong Baseline for Irregular Text Recognition},
+  author={Hui Li and Peng Wang and Chunhua Shen and Guyu Zhang},
+  journal={ArXiv},
+  year={2019},
+  volume={abs/1811.00751}
+}
+```

doc/doc_en/algorithm_rec_srn_en.md

+# SRN
+
+- [1. Introduction](#1)
+- [2. Environment](#2)
+- [3. Model Training / Evaluation / Prediction](#3)
+    - [3.1 Training](#3-1)
+    - [3.2 Evaluation](#3-2)
+    - [3.3 Prediction](#3-3)
+- [4. Inference and Deployment](#4)
+    - [4.1 Python Inference](#4-1)
+    - [4.2 C++ Inference](#4-2)
+    - [4.3 Serving](#4-3)
+    - [4.4 More](#4-4)
+- [5. FAQ](#5)
+
+<a name="1"></a>
+## 1. Introduction
+
+Paper:
+> [Towards Accurate Scene Text Recognition with Semantic Reasoning Networks](https://arxiv.org/abs/2003.12294#)
+> Deli Yu, Xuan Li, Chengquan Zhang, Junyu Han, Jingtuo Liu, Errui Ding
+> CVPR,2020
+
+Using MJSynth and SynthText two text recognition datasets for training, and evaluating on IIIT, SVT, IC03, IC13, IC15, SVTP, CUTE datasets, the algorithm reproduction effect is as follows:
+
+|Model|Backbone|config|Acc|Download link|
+| --- | --- | --- | --- | --- | --- | --- |
+|SRN|Resnet50_vd_fpn|[rec_r50_fpn_srn.yml](../../configs/rec/rec_r50_fpn_srn.yml)|86.31%|[train model](https://paddleocr.bj.bcebos.com/dygraph_v2.0/en/rec_r50_vd_srn_train.tar)|
+
+
+<a name="2"></a>
+## 2. Environment
+Please refer to ["Environment Preparation"](./environment.md) to configure the PaddleOCR environment, and refer to ["Project Clone"](./clone.md) to clone the project code.
+
+
+<a name="3"></a>
+## 3. Model Training / Evaluation / Prediction
+
+Please refer to [Text Recognition Tutorial](./recognition.md). PaddleOCR modularizes the code, and training different recognition models only requires **changing the configuration file**.
+
+Training:
+
+Specifically, after the data preparation is completed, the training can be started. The training command is as follows:
+
+```
+#Single GPU training (long training period, not recommended)
+python3 tools/train.py -c configs/rec/rec_r50_fpn_srn.yml
+
+#Multi GPU training, specify the gpu number through the --gpus parameter
+python3 -m paddle.distributed.launch --gpus '0,1,2,3'  tools/train.py -c configs/rec/rec_r50_fpn_srn.yml
+```
+
+Evaluation:
+
+```
+# GPU evaluation
+python3 -m paddle.distributed.launch --gpus '0' tools/eval.py -c configs/rec/rec_r50_fpn_srn.yml -o Global.pretrained_model={path/to/weights}/best_accuracy
+```
+
+Prediction:
+
+```
+# The configuration file used for prediction must match the training
+python3 tools/infer_rec.py -c configs/rec/rec_r50_fpn_srn.yml -o Global.pretrained_model={path/to/weights}/best_accuracy Global.infer_img=doc/imgs_words/en/word_1.png
+```
+
+<a name="4"></a>
+## 4. Inference and Deployment
+
+<a name="4-1"></a>
+### 4.1 Python Inference
+First, the model saved during the SRN text recognition training process is converted into an inference model. ( [Model download link](https://paddleocr.bj.bcebos.com/dygraph_v2.0/en/rec_r50_vd_srn_train.tar) ), you can use the following command to convert:
+
+```
+python3 tools/export_model.py -c configs/rec/rec_r50_fpn_srn.yml -o Global.pretrained_model=./rec_r50_vd_srn_train/best_accuracy  Global.save_inference_dir=./inference/rec_srn
+```
+
+For SRN text recognition model inference, the following commands can be executed:
+
+```
+python3 tools/infer/predict_rec.py --image_dir="./doc/imgs_words/en/word_1.png" --rec_model_dir="./inference/rec_srn/" --rec_image_shape="1,64,256" --rec_char_type="ch" --rec_algorithm="SRN" --rec_char_dict_path="ppocr/utils/ic15_dict.txt" --use_space_char=False
+```
+
+<a name="4-2"></a>
+### 4.2 C++ Inference
+
+Not supported
+
+<a name="4-3"></a>
+### 4.3 Serving
+
+Not supported
+
+<a name="4-4"></a>
+### 4.4 More
+
+Not supported
+
+<a name="5"></a>
+## 5. FAQ
+
+
+## Citation
+
+```bibtex
+@article{Yu2020TowardsAS,
+  title={Towards Accurate Scene Text Recognition With Semantic Reasoning Networks},
+  author={Deli Yu and Xuan Li and Chengquan Zhang and Junyu Han and Jingtuo Liu and Errui Ding},
+  journal={2020 IEEE/CVF Conference on Computer Vision and Pattern Recognition (CVPR)},
+  year={2020},
+  pages={12110-12119}
+}
+```

doc/doc_en/detection_en.md

@@ -6,10 +6,13 @@ This section uses the icdar2015 dataset as an example to introduce the training,
   - [1.1 Data Preparation](#11-data-preparation)
   - [1.2 Download Pre-trained Model](#12-download-pre-trained-model)
 - [2. Training](#2-training)
-  - [2.1 Start Training](#21-start-training)
-  - [2.2 Load Trained Model and Continue Training](#22-load-trained-model-and-continue-training)
-  - [2.3 Training with New Backbone](#23-training-with-new-backbone)
-  - [2.4 Training with knowledge distillation](#24-training-with-knowledge-distillation)
+  * [2.1 Start Training](#21-start-training)
+  * [2.2 Load Trained Model and Continue Training](#22-load-trained-model-and-continue-training)
+  * [2.3 Training with New Backbone](#23-training-with-new-backbone)
+  * [2.4 Mixed Precision Training](#24-amp-training)
+  * [2.5 Distributed Training](#25-distributed-training)
+  * [2.6 Training with knowledge distillation](#26)
+  * [2.7 Training on other platform(Windows/macOS/Linux DCU)](#27)
 - [3. Evaluation and Test](#3-evaluation-and-test)
   - [3.1 Evaluation](#31-evaluation)
   - [3.2 Test](#32-test)
@@ -137,11 +140,44 @@ After adding the four-part modules of the network, you only need to configure th
 
 **NOTE**: More details about replace Backbone and other mudule can be found in [doc](add_new_algorithm_en.md).
 
+### 2.4 Mixed Precision Training
 
-### 2.4 Training with knowledge distillation
+If you want to speed up your training further, you can use [Auto Mixed Precision Training](https://www.paddlepaddle.org.cn/documentation/docs/zh/guides/01_paddle2.0_introduction/basic_concept/amp_cn.html), taking a single machine and a single gpu as an example, the commands are as follows:
+
+```shell
+python3 tools/train.py -c configs/det/det_mv3_db.yml \
+     -o Global.pretrained_model=./pretrain_models/MobileNetV3_large_x0_5_pretrained \
+     Global.use_amp=True Global.scale_loss=1024.0 Global.use_dynamic_loss_scaling=True
+ ```
+
+### 2.5 Distributed Training
+
+During multi-machine multi-gpu training, use the `--ips` parameter to set the used machine IP address, and the `--gpus` parameter to set the used GPU ID:
+
+```bash
+python3 -m paddle.distributed.launch --ips="xx.xx.xx.xx,xx.xx.xx.xx" --gpus '0,1,2,3' tools/train.py -c configs/det/det_mv3_db.yml \
+     -o Global.pretrained_model=./pretrain_models/MobileNetV3_large_x0_5_pretrained
+```
+
+**Note:** When using multi-machine and multi-gpu training, you need to replace the ips value in the above command with the address of your machine, and the machines need to be able to ping each other. In addition, training needs to be launched separately on multiple machines. The command to view the ip address of the machine is `ifconfig`.
+
+### 2.6 Training with knowledge distillation
 
 Knowledge distillation is supported in PaddleOCR for text detection training process. For more details, please refer to [doc](./knowledge_distillation_en.md).
 
+### 2.7 Training on other platform(Windows/macOS/Linux DCU)
+
+- Windows GPU/CPU
+The Windows platform is slightly different from the Linux platform:
+Windows platform only supports `single gpu` training and inference, specify GPU for training `set CUDA_VISIBLE_DEVICES=0`
+On the Windows platform, DataLoader only supports single-process mode, so you need to set `num_workers` to 0;
+
+- macOS
+GPU mode is not supported, you need to set `use_gpu` to False in the configuration file, and the rest of the training evaluation prediction commands are exactly the same as Linux GPU.
+
+- Linux DCU
+Running on a DCU device requires setting the environment variable `export HIP_VISIBLE_DEVICES=0,1,2,3`, and the rest of the training and evaluation prediction commands are exactly the same as the Linux GPU.
+
 ## 3. Evaluation and Test
 
 ### 3.1 Evaluation

doc/doc_en/recognition_en.md

 # Text Recognition
 
-- [1. Data Preparation](#1-data-preparation)
-  - [1.1 DataSet Preparation](#11-dataset-preparation)
-  - [1.2 Dictionary](#12-dictionary)
-  - [1.4 Add Space Category](#14-add-space-category)
-- [2.Training](#2training)
-  - [2.1 Data Augmentation](#21-data-augmentation)
-  - [2.2 General Training](#22-general-training)
-  - [2.3 Multi-language Training](#23-multi-language-training)
-  - [2.4 Training with Knowledge Distillation](#24-training-with-knowledge-distillation)
-- [3. Evalution](#3-evalution)
-- [4. Prediction](#4-prediction)
-- [5. Convert to Inference Model](#5-convert-to-inference-model)
+- [1. Data Preparation](#DATA_PREPARATION)
+  * [1.1 Costom Dataset](#Costom_Dataset)
+  * [1.2 Dataset Download](#Dataset_download)
+  * [1.3 Dictionary](#Dictionary)  
+  * [1.4 Add Space Category](#Add_space_category)
+  * [1.5 Data Augmentation](#Data_Augmentation)
+- [2. Training](#TRAINING)
+  * [2.1 Start Training](#21-start-training)
+  * [2.2 Load Trained Model and Continue Training](#22-load-trained-model-and-continue-training)
+  * [2.3 Training with New Backbone](#23-training-with-new-backbone)
+  * [2.4 Mixed Precision Training](#24-amp-training)
+  * [2.5 Distributed Training](#25-distributed-training)
+  * [2.6 Training with knowledge distillation](#kd)
+  * [2.7 Multi-language Training](#Multi_language)
+  * [2.8 Training on other platform(Windows/macOS/Linux DCU)](#28)
+- [3. Evaluation and Test](#3-evaluation-and-test)
+  * [3.1 Evaluation](#31-evaluation)
+  * [3.2 Test](#32-test)
+- [4. Inference](#4-inference)
+- [5. FAQ](#5-faq)
 
 <a name="DATA_PREPARATION"></a>
 ## 1. Data Preparation
@@ -72,11 +80,8 @@ If you need to customize dic file, please add character_dict_path field in confi
 
 If you want to support the recognition of the `space` category, please set the `use_space_char` field in the yml file to `True`.
 
-<a name="TRAINING"></a>
-## 2.Training
-
 <a name="Data_Augmentation"></a>
-### 2.1 Data Augmentation
+### 1.5 Data Augmentation
 
 PaddleOCR provides a variety of data augmentation methods. All the augmentation methods are enabled by default.
 
@@ -84,11 +89,14 @@ The default perturbation methods are: cvtColor, blur, jitter, Gasuss noise, rand
 
 Each disturbance method is selected with a 40% probability during the training process. For specific code implementation, please refer to: [rec_img_aug.py](../../ppocr/data/imaug/rec_img_aug.py)
 
-<a name="Training"></a>
-### 2.2 General Training
+<a name="TRAINING"></a>
+## 2.Training
 
 PaddleOCR provides training scripts, evaluation scripts, and prediction scripts. In this section, the CRNN recognition model will be used as an example:
 
+<a name="21-start-training"></a>
+### 2.1 Start Training
+
 First download the pretrain model, you can download the trained model to finetune on the icdar2015 data:
 
 ```
@@ -204,8 +212,99 @@ Eval:
 ```
 **Note that the configuration file for prediction/evaluation must be consistent with the training.**
 
+<a name="22-load-trained-model-and-continue-training"></a>
+### 2.2 Load Trained Model and Continue Training
+
+If you expect to load trained model and continue the training again, you can specify the parameter `Global.checkpoints` as the model path to be loaded.
+
+For example:
+```shell
+python3 tools/train.py -c configs/rec/rec_icdar15_train.yml -o Global.checkpoints=./your/trained/model
+```
+
+**Note**: The priority of `Global.checkpoints` is higher than that of `Global.pretrained_model`, that is, when two parameters are specified at the same time, the model specified by `Global.checkpoints` will be loaded first. If the model path specified by `Global.checkpoints` is wrong, the one specified by `Global.pretrained_model` will be loaded.
+
+<a name="23-training-with-new-backbone"></a>
+### 2.3 Training with New Backbone
+
+The network part completes the construction of the network, and PaddleOCR divides the network into four parts, which are under [ppocr/modeling](../../ppocr/modeling). The data entering the network will pass through these four parts in sequence(transforms->backbones->
+necks->heads).
+
+```bash
+├── architectures # Code for building network
+├── transforms    # Image Transformation Module
+├── backbones     # Feature extraction module
+├── necks         # Feature enhancement module
+└── heads         # Output module
+```
+
+If the Backbone to be replaced has a corresponding implementation in PaddleOCR, you can directly modify the parameters in the `Backbone` part of the configuration yml file.
+
+However, if you want to use a new Backbone, an example of replacing the backbones is as follows:
+
+1. Create a new file under the [ppocr/modeling/backbones](../../ppocr/modeling/backbones) folder, such as my_backbone.py.
+2. Add code in the my_backbone.py file, the sample code is as follows:
+
+```python
+import paddle
+import paddle.nn as nn
+import paddle.nn.functional as F
+
+
+class MyBackbone(nn.Layer):
+    def __init__(self, *args, **kwargs):
+        super(MyBackbone, self).__init__()
+        # your init code
+        self.conv = nn.xxxx
+
+    def forward(self, inputs):
+        # your network forward
+        y = self.conv(inputs)
+        return y
+```
+
+3. Import the added module in the [ppocr/modeling/backbones/\__init\__.py](../../ppocr/modeling/backbones/__init__.py) file.
+
+After adding the four-part modules of the network, you only need to configure them in the configuration file to use, such as:
+
+```yaml
+  Backbone:
+    name: MyBackbone
+    args1: args1
+```
+
+**NOTE**: More details about replace Backbone and other mudule can be found in [doc](add_new_algorithm_en.md).
+
+<a name="24-amp-training"></a>
+### 2.4 Mixed Precision Training
+
+If you want to speed up your training further, you can use [Auto Mixed Precision Training](https://www.paddlepaddle.org.cn/documentation/docs/zh/guides/01_paddle2.0_introduction/basic_concept/amp_cn.html), taking a single machine and a single gpu as an example, the commands are as follows:
+
+```shell
+python3 tools/train.py -c configs/rec/rec_icdar15_train.yml \
+     -o Global.pretrained_model=./pretrain_models/rec_mv3_none_bilstm_ctc_v2.0_train \
+     Global.use_amp=True Global.scale_loss=1024.0 Global.use_dynamic_loss_scaling=True
+ ```
+
+<a name="25-distributed-training"></a>
+### 2.5 Distributed Training
+
+During multi-machine multi-gpu training, use the `--ips` parameter to set the used machine IP address, and the `--gpus` parameter to set the used GPU ID:
+
+```bash
+python3 -m paddle.distributed.launch --ips="xx.xx.xx.xx,xx.xx.xx.xx" --gpus '0,1,2,3' tools/train.py -c configs/rec/rec_icdar15_train.yml \
+     -o Global.pretrained_model=./pretrain_models/rec_mv3_none_bilstm_ctc_v2.0_train
+```
+
+**Note:** When using multi-machine and multi-gpu training, you need to replace the ips value in the above command with the address of your machine, and the machines need to be able to ping each other. In addition, training needs to be launched separately on multiple machines. The command to view the ip address of the machine is `ifconfig`.
+
+<a name="kd"></a>
+### 2.6 Training with Knowledge Distillation
+
+Knowledge distillation is supported in PaddleOCR for text recognition training process. For more details, please refer to [doc](./knowledge_distillation_en.md).
+
 <a name="Multi_language"></a>
-### 2.3 Multi-language Training
+### 2.7 Multi-language Training
 
 Currently, the multi-language algorithms supported by PaddleOCR are:
 
@@ -261,25 +360,35 @@ Eval:
     ...
 ```
 
-<a name="kd"></a>
+<a name="28"></a>
+### 2.8 Training on other platform(Windows/macOS/Linux DCU)
 
-### 2.4 Training with Knowledge Distillation
+- Windows GPU/CPU
+The Windows platform is slightly different from the Linux platform:
+Windows platform only supports `single gpu` training and inference, specify GPU for training `set CUDA_VISIBLE_DEVICES=0`
+On the Windows platform, DataLoader only supports single-process mode, so you need to set `num_workers` to 0;
 
-Knowledge distillation is supported in PaddleOCR for text recognition training process. For more details, please refer to [doc](./knowledge_distillation_en.md).
+- macOS
+GPU mode is not supported, you need to set `use_gpu` to False in the configuration file, and the rest of the training evaluation prediction commands are exactly the same as Linux GPU.
 
-<a name="EVALUATION"></a>
+- Linux DCU
+Running on a DCU device requires setting the environment variable `export HIP_VISIBLE_DEVICES=0,1,2,3`, and the rest of the training and evaluation prediction commands are exactly the same as the Linux GPU.
 
-## 3. Evalution
+<a name="3-evaluation-and-test"></a>
+## 3. Evaluation and Test
 
-The evaluation dataset can be set by modifying the `Eval.dataset.label_file_list` field in the `configs/rec/rec_icdar15_train.yml` file.
+<a name="31-evaluation"></a>
+### 3.1 Evaluation
+
+The model parameters during training are saved in the `Global.save_model_dir` directory by default. When evaluating indicators, you need to set `Global.checkpoints` to point to the saved parameter file. The evaluation dataset can be set by modifying the `Eval.dataset.label_file_list` field in the `configs/rec/rec_icdar15_train.yml` file.
 
 ```
 # GPU evaluation, Global.checkpoints is the weight to be tested
 python3 -m paddle.distributed.launch --gpus '0' tools/eval.py -c configs/rec/rec_icdar15_train.yml -o Global.checkpoints={path/to/weights}/best_accuracy
 ```
 
-<a name="PREDICTION"></a>
-## 4. Prediction
+<a name="32-test"></a>
+### 3.2 Test
 
 
 Using the model trained by paddleocr, you can quickly get prediction through the following script.
@@ -341,9 +450,14 @@ infer_img: doc/imgs_words/ch/word_1.jpg
         result: ('韩国小馆', 0.997218)
 ```
 
-<a name="Inference"></a>
+<a name="4-inference"></a>
+## 4. Inference
+
+The inference model (the model saved by `paddle.jit.save`) is generally a solidified model saved after the model training is completed, and is mostly used to give prediction in deployment.
+
+The model saved during the training process is the checkpoints model, which saves the parameters of the model and is mostly used to resume training.
 
-## 5. Convert to Inference Model
+Compared with the checkpoints model, the inference model will additionally save the structural information of the model. Therefore, it is easier to deploy because the model structure and model parameters are already solidified in the inference model file, and is suitable for integration with actual systems.
 
 The recognition model is converted to the inference model in the same way as the detection, as follows:
 
@@ -361,7 +475,7 @@ If you have a model trained on your own dataset with a different dictionary file
 After the conversion is successful, there are three files in the model save directory:
 
 ```
-inference/det_db/
+inference/rec_crnn/
     ├── inference.pdiparams         # The parameter file of recognition inference model
     ├── inference.pdiparams.info    # The parameter information of recognition inference model, which can be ignored
     └── inference.pdmodel           # The program file of recognition model
@@ -374,3 +488,10 @@ inference/det_db/
   ```
   python3 tools/infer/predict_rec.py --image_dir="./doc/imgs_words_en/word_336.png" --rec_model_dir="./your inference model" --rec_image_shape="3, 32, 100" --rec_char_dict_path="your text dict path"
   ```
+
+<a name="5-faq"></a>
+## 5. FAQ
+
+Q1: After the training model is transferred to the inference model, the prediction effect is inconsistent?
+
+**A**: There are many such problems, and the problems are mostly caused by inconsistent preprocessing and postprocessing parameters when the trained model predicts and the preprocessing and postprocessing parameters when the inference model predicts. You can compare whether there are differences in preprocessing, postprocessing, and prediction in the configuration files used for training.