Merge branch 'PaddlePaddle:release/2.4' into polish_serving_docs

README.md

-README_ch.md
\ No newline at end of file
+README_en.md
\ No newline at end of file

README_ch.md

@@ -20,7 +20,7 @@
 
 
 ## 近期更新
-- 📢将于**<u>6月15-6月17日晚20:30</u>**进行为期三天的课程直播，详细介绍超轻量图像分类方案，对各场景模型优化原理及使用方式进行拆解，之后还有产业案例全流程实操，对各类痛难点解决方案进行手把手教学，加上现场互动答疑，抓紧扫码上车吧！
+- 📢将于**6月15-6月17日晚20:30** 进行为期三天的课程直播，详细介绍超轻量图像分类方案，对各场景模型优化原理及使用方式进行拆解，之后还有产业案例全流程实操，对各类痛难点解决方案进行手把手教学，加上现场互动答疑，抓紧扫码上车吧！
 
 <div align="center">
 <img src="https://user-images.githubusercontent.com/45199522/173483779-2332f990-4941-4f8d-baee-69b62035fc31.png" width = "200" height = "200"/>
@@ -47,11 +47,11 @@ PaddleClas发布了[PP-HGNet](docs/zh_CN/models/PP-HGNet.md)、[PP-LCNetv2](docs
 
 ## 欢迎加入技术交流群
 
-* 您可以扫描下面的QQ/微信二维码（添加小助手微信并回复“C”），加入PaddleClas微信交流群，获得更高效的问题答疑，与各行各业开发者充分交流，期待您的加入。
+* 您可以扫描下面的微信/QQ二维码（添加小助手微信并回复“C”），加入PaddleClas微信交流群，获得更高效的问题答疑，与各行各业开发者充分交流，期待您的加入。
 
 <div align="center">
-<img src="https://user-images.githubusercontent.com/80816848/164383225-e375eb86-716e-41b4-a9e0-4b8a3976c1aa.jpg" width="200"/>
 <img src="https://user-images.githubusercontent.com/48054808/160531099-9811bbe6-cfbb-47d5-8bdb-c2b40684d7dd.png" width="200"/>
+<img src="https://user-images.githubusercontent.com/80816848/164383225-e375eb86-716e-41b4-a9e0-4b8a3976c1aa.jpg" width="200"/>
 </div>
 
 ## 快速体验

README_en.md

@@ -7,7 +7,7 @@
 PaddleClas is an image classification and image recognition toolset for industry and academia, helping users train better computer vision models and apply them in real scenarios.
 
 <div align="center">
-<img src="./docs/images/class_simple.gif"  width = "600" />
+<img src="./docs/images/class_simple_en.gif"  width = "600" />
 
 PULC demo images
 </div>
@@ -21,24 +21,24 @@ PP-ShiTu demo images
 </div>
 
 **Recent updates**
-- 2022.6.15 Release [**P**ractical **U**ltra **L**ight-weight image **C**lassification solutions](./docs/en/PULC/PULC_quickstart_en.md). PULC models inference within 3ms on CPU devices, with accuracy comparable with SwinTransformer. We also release 9 practical models covering pedestrian, vehicle and OCR.
+- 2022.6.15 Release [**P**ractical **U**ltra **L**ight-weight image **C**lassification solutions](./docs/en/PULC/PULC_quickstart_en.md). PULC models inference within 3ms on CPU devices, with accuracy on par with SwinTransformer. We also release 9 practical classification models covering pedestrian, vehicle and OCR scenario.
 - 2022.4.21 Added the related [code](https://github.com/PaddlePaddle/PaddleClas/pull/1820/files) of the CVPR2022 oral paper [MixFormer](https://arxiv.org/pdf/2204.02557.pdf).
 
 - 2021.09.17 Add PP-LCNet series model developed by PaddleClas, these models show strong competitiveness on Intel CPUs.
-For the introduction of PP-LCNet, please refer to [paper](https://arxiv.org/pdf/2109.15099.pdf) or [PP-LCNet model introduction](docs/en/models/PP-LCNet_en.md). The metrics and pretrained model are available [here](docs/en/ImageNet_models_en.md).
+For the introduction of PP-LCNet, please refer to [paper](https://arxiv.org/pdf/2109.15099.pdf) or [PP-LCNet model introduction](docs/en/models/PP-LCNet_en.md). The metrics and pretrained model are available [here](docs/en/algorithm_introduction/ImageNet_models_en.md).
 
-- 2021.06.29 Add Swin-transformer series model，Highest top1 acc on ImageNet1k dataset reaches 87.2%, training, evaluation and inference are all supported. Pretrained models can be downloaded [here](docs/en/models/models_intro_en.md).
+- 2021.06.29 Add [Swin-transformer](docs/en/models/SwinTransformer_en.md)) series model，Highest top1 acc on ImageNet1k dataset reaches 87.2%, training, evaluation and inference are all supported. Pretrained models can be downloaded [here](docs/en/algorithm_introduction/ImageNet_models_en.md#16).
 - 2021.06.16 PaddleClas release/2.2. Add metric learning and vector search modules. Add product recognition, animation character recognition, vehicle recognition and logo recognition. Added 30 pretrained models of LeViT, Twins, TNT, DLA, HarDNet, and RedNet, and the accuracy is roughly the same as that of the paper.
-- [more](./docs/en/update_history_en.md)
+- [more](./docs/en/others/update_history_en.md)
 
 ## Features
 
-PaddleClas release PP-HGNet、PP-LCNetv2、 PP-LCNet and **S**imple **S**emi-supervised **L**abel **D**istillation algorithms, and support plenty of 
-image classification and image recognition algorithms. 
+PaddleClas release PP-HGNet、PP-LCNetv2、 PP-LCNet and **S**imple **S**emi-supervised **L**abel **D**istillation algorithms, and support plenty of
+image classification and image recognition algorithms.
 Based on th algorithms above, PaddleClas release PP-ShiTu image recognition system and [**P**ractical **U**ltra **L**ight-weight image **C**lassification solutions](docs/en/PULC/PULC_quickstart_en.md).
 
 
-![](https://user-images.githubusercontent.com/19523330/173347904-f2998e00-7b86-4adf-b546-23c684fc67b9.png)
+![](https://user-images.githubusercontent.com/19523330/173539361-68cf7ab1-7e3b-4e5e-b00f-1500719bd2a2.png)
 
 ## Welcome to Join the Technical Exchange Group
 
@@ -50,40 +50,46 @@ Based on th algorithms above, PaddleClas release PP-ShiTu image recognition syst
 </div>
 
 ## Quick Start
-Quick experience of PP-ShiTu image recognition system：[Link](./docs/en/tutorials/quick_start_recognition_en.md)
-Quick experience of **P**ractical **U**ltra **L**ight-weight image **C**lassification models：[Link](docs/en/PULC/PULC_quickstart.md)
+Quick experience of PP-ShiTu image recognition system：[Link](./docs/en/quick_start/quick_start_recognition_en.md)
+
+Quick experience of **P**ractical **U**ltra **L**ight-weight image **C**lassification models：[Link](docs/en/PULC/PULC_quickstart_en.md)
 
 ## Tutorials
 
-- [Quick Installation](./docs/en/tutorials/install_en.md)
-- [Practical Ultra Light-weight image Classification solutions](./docs/en/PULC/PULC_quickstart_en.md)
-- [Quick Start of Recognition](./docs/en/tutorials/quick_start_recognition_en.md)
+- [Install Paddle](./docs/en/installation/install_paddle_en.md)
+- [Install PaddleClas Environment](./docs/en/installation/install_paddleclas_en.md)
+- [Practical Ultra Light-weight image Classification solutions](./docs/en/PULC/PULC_train_en.md)
+  - [PULC Quick Start](docs/en/PULC/PULC_quickstart_en.md)
+  - [PULC Model Zoo](docs/en/PULC/PULC_model_list_en.md)
+    - [PULC Classification Model of Someone or Nobody](docs/en/PULC/PULC_person_exists_en.md)
+    - [PULC Recognition Model of Person Attribute](docs/en/PULC/PULC_person_attribute_en.md)
+    - [PULC Classification Model of Wearing or Unwearing Safety Helmet](docs/en/PULC/PULC_safety_helmet_en.md)
+    - [PULC Classification Model of Traffic Sign](docs/en/PULC/PULC_traffic_sign_en.md)
+    - [PULC Recognition Model of Vehicle Attribute](docs/en/PULC/PULC_vehicle_attribute_en.md)
+    - [PULC Classification Model of Containing or Uncontaining Car](docs/en/PULC/PULC_car_exists_en.md)
+    - [PULC Classification Model of Text Image Orientation](docs/en/PULC/PULC_text_image_orientation_en.md)
+    - [PULC Classification Model of Textline Orientation](docs/en/PULC/PULC_textline_orientation_en.md)
+    - [PULC Classification Model of Language](docs/en/PULC/PULC_language_classification_en.md)
+- [Quick Start of Recognition](./docs/en/quick_start/quick_start_recognition_en.md)
 - [Introduction to Image Recognition Systems](#Introduction_to_Image_Recognition_Systems)
-- [Demo images](#Demo_images)
+- [Image Recognition Demo images](#Rec_Demo_images)
+- [PULC demo images](#Clas_Demo_images)
 - Algorithms Introduction
-    - [Backbone Network and Pre-trained Model Library](./docs/en/ImageNet_models_en.md)
-    - [Mainbody Detection](./docs/en/application/mainbody_detection_en.md)
-    - [Image Classification](./docs/en/tutorials/image_classification_en.md)
-    - [Feature Learning](./docs/en/application/feature_learning_en.md)
-        - [Product Recognition](./docs/en/application/product_recognition_en.md)
-        - [Vehicle Recognition](./docs/en/application/vehicle_recognition_en.md)
-        - [Logo Recognition](./docs/en/application/logo_recognition_en.md)
-        - [Animation Character Recognition](./docs/en/application/cartoon_character_recognition_en.md)
+    - [Backbone Network and Pre-trained Model Library](./docs/en/algorithm_introduction/ImageNet_models_en.md)
+    - [Mainbody Detection](./docs/en/image_recognition_pipeline/mainbody_detection_en.md)
+    - [Feature Learning](./docs/en/image_recognition_pipeline/feature_extraction_en.md)
     - [Vector Search](./deploy/vector_search/README.md)
-- Models Training/Evaluation
-    - [Image Classification](./docs/en/tutorials/getting_started_en.md)
-    - [Feature Learning](./docs/en/tutorials/getting_started_retrieval_en.md)
 - Inference Model Prediction
-    - [Python Inference](./docs/en/inference.md)
+    - [Python Inference](./docs/en/inference_deployment/python_deploy_en.md)
     - [C++ Classfication Inference](./deploy/cpp/readme_en.md)， [C++ PP-ShiTu Inference](deploy/cpp_shitu/readme_en.md)
 - Model Deploy (only support classification for now, recognition coming soon)
     - [Hub Serving Deployment](./deploy/hubserving/readme_en.md)
     - [Mobile Deployment](./deploy/lite/readme_en.md)
-    - [Inference Using whl](./docs/en/whl_en.md)
+    - [Inference Using whl](./docs/en/inference_deployment/whl_deploy_en.md)
 - Advanced Tutorial
     - [Knowledge Distillation](./docs/en/advanced_tutorials/distillation/distillation_en.md)
-    - [Model Quantization](./docs/en/extension/paddle_quantization_en.md)
-    - [Data Augmentation](./docs/en/advanced_tutorials/image_augmentation/ImageAugment_en.md)
+    - [Model Quantization](./docs/en/algorithm_introduction/model_prune_quantization_en.md)
+    - [Data Augmentation](./docs/en/advanced_tutorials/DataAugmentation_en.md)
 - [License](#License)
 - [Contribution](#Contribution)
 
@@ -109,9 +115,10 @@ Image recognition can be divided into three steps:
 
 For a new unknown category, there is no need to retrain the model, just prepare images of new category, extract features and update retrieval database and the category can be recognised.
 
+<a name="Clas_Demo_images"></a>
 ## PULC demo images
 <div align="center">
-<img src="docs/images/classification.gif">
+<img src="docs/images/classification_en.gif">
 </div>
 
 <a name="Rec_Demo_images"></a>

deploy/paddle2onnx/readme.md

 # paddle2onnx 模型转化与预测
 
-本章节介绍 ResNet50_vd 模型如何转化为 ONNX 模型，并基于 ONNX 引擎预测。
+## 目录
+
+- [paddle2onnx 模型转化与预测](#paddle2onnx-模型转化与预测)
+  - [1. 环境准备](#1-环境准备)
+  - [2. 模型转换](#2-模型转换)
+  - [3. onnx 预测](#3-onnx-预测)
 
 ## 1. 环境准备
 
 需要准备 Paddle2ONNX 模型转化环境，和 ONNX 模型预测环境。
 
-Paddle2ONNX 支持将 PaddlePaddle 模型格式转化到 ONNX 模型格式，算子目前稳定支持导出 ONNX Opset 9~11，部分Paddle算子支持更低的ONNX Opset转换。
-更多细节可参考 [Paddle2ONNX](https://github.com/PaddlePaddle/Paddle2ONNX/blob/develop/README_zh.md)
+Paddle2ONNX 支持将 PaddlePaddle inference 模型格式转化到 ONNX 模型格式，算子目前稳定支持导出 ONNX Opset 9~11。
+更多细节可参考 [Paddle2ONNX](https://github.com/PaddlePaddle/Paddle2ONNX#paddle2onnx)
 
 - 安装 Paddle2ONNX
-```
-python3.7 -m pip install paddle2onnx
-```
+    ```shell
+    python3.7 -m pip install paddle2onnx
+    ```
 
-- 安装 ONNX 运行时
-```
-python3.7 -m pip install onnxruntime
-```
+- 安装 ONNX 推理引擎
+    ```shell
+    python3.7 -m pip install onnxruntime
+    ```
+下面以 ResNet50_vd 为例，介绍如何将 PaddlePaddle inference 模型转换为 ONNX 模型，并基于 ONNX 引擎预测。
 
 ## 2. 模型转换
 
 - ResNet50_vd inference模型下载
 
-```
-cd deploy
-mkdir models && cd models
-wget -nc https://paddle-imagenet-models-name.bj.bcebos.com/dygraph/inference/ResNet50_vd_infer.tar && tar xf ResNet50_vd_infer.tar
-cd ..
-```
+    ```shell
+    cd deploy
+    mkdir models && cd models
+    wget -nc https://paddle-imagenet-models-name.bj.bcebos.com/dygraph/inference/ResNet50_vd_infer.tar && tar xf ResNet50_vd_infer.tar
+    cd ..
+    ```
 
 - 模型转换
 
-使用 Paddle2ONNX 将 Paddle 静态图模型转换为 ONNX 模型格式：
-```
-paddle2onnx --model_dir=./models/ResNet50_vd_infer/ \
---model_filename=inference.pdmodel \
---params_filename=inference.pdiparams \
---save_file=./models/ResNet50_vd_infer/inference.onnx \
---opset_version=10 \
---enable_onnx_checker=True
-```
+    使用 Paddle2ONNX 将 Paddle 静态图模型转换为 ONNX 模型格式：
+    ```shell
+    paddle2onnx --model_dir=./models/ResNet50_vd_infer/ \
+    --model_filename=inference.pdmodel \
+    --params_filename=inference.pdiparams \
+    --save_file=./models/ResNet50_vd_infer/inference.onnx \
+    --opset_version=10 \
+    --enable_onnx_checker=True
+    ```
 
-执行完毕后，ONNX 模型 `inference.onnx` 会被保存在 `./models/ResNet50_vd_infer/` 路径下
+转换完毕后，生成的ONNX 模型 `inference.onnx` 会被保存在 `./models/ResNet50_vd_infer/` 路径下
 
 ## 3. onnx 预测
 
 执行如下命令：
-```
+```shell
 python3.7 python/predict_cls.py \
 -c configs/inference_cls.yaml \
 -o Global.use_onnx=True \

deploy/paddle2onnx/readme_en.md

+# Paddle2ONNX: Converting To ONNX and Deployment
+
+This section introduce that how to convert the Paddle Inference Model ResNet50_vd to ONNX model and deployment based on ONNX engine.
+
+## 1. Installation
+
+First, you need to install Paddle2ONNX and onnxruntime. Paddle2ONNX is a toolkit to convert Paddle Inference Model to ONNX model. Please refer to [Paddle2ONNX](https://github.com/PaddlePaddle/Paddle2ONNX/blob/develop/README_en.md) for more information.
+
+- Paddle2ONNX Installation
+```
+python3.7 -m pip install paddle2onnx
+```
+
+- ONNX Installation
+```
+python3.7 -m pip install onnxruntime
+```
+
+## 2. Converting to ONNX
+
+Download the Paddle Inference Model ResNet50_vd:
+
+```
+cd deploy
+mkdir models && cd models
+wget -nc https://paddle-imagenet-models-name.bj.bcebos.com/dygraph/inference/ResNet50_vd_infer.tar && tar xf ResNet50_vd_infer.tar
+cd ..
+```
+
+Converting to ONNX model:
+
+```
+paddle2onnx --model_dir=./models/ResNet50_vd_infer/ \
+--model_filename=inference.pdmodel \
+--params_filename=inference.pdiparams \
+--save_file=./models/ResNet50_vd_infer/inference.onnx \
+--opset_version=10 \
+--enable_onnx_checker=True
+```
+
+After running the above command, the ONNX model file converted would be save in  `./models/ResNet50_vd_infer/`.
+
+## 3. Deployment
+
+Deployment with ONNX model, command is as shown below.
+
+```
+python3.7 python/predict_cls.py \
+-c configs/inference_cls.yaml \
+-o Global.use_onnx=True \
+-o Global.use_gpu=False \
+-o Global.inference_model_dir=./models/ResNet50_vd_infer
+```
+
+The prediction results:
+
+```
+ILSVRC2012_val_00000010.jpeg:   class id(s): [153, 204, 229, 332, 155], score(s): [0.69, 0.10, 0.02, 0.01, 0.01], label_name(s): ['Maltese dog, Maltese terrier, Maltese', 'Lhasa, Lhasa apso', 'Old English sheepdog, bobtail', 'Angora, Angora rabbit', 'Shih-Tzu']
+```

deploy/slim/quant_post_static.py

@@ -41,8 +41,11 @@ def main():
                      'inference.pdmodel')) and os.path.exists(
                          os.path.join(config["Global"]["save_inference_dir"],
                                       'inference.pdiparams'))
+    if "Query" in config["DataLoader"]["Eval"]:
+        config["DataLoader"]["Eval"] = config["DataLoader"]["Eval"]["Query"]
     config["DataLoader"]["Eval"]["sampler"]["batch_size"] = 1
     config["DataLoader"]["Eval"]["loader"]["num_workers"] = 0
+
     init_logger()
     device = paddle.set_device("cpu")
     train_dataloader = build_dataloader(config["DataLoader"], "Eval", device,
@@ -67,6 +70,7 @@ def main():
         quantize_model_path=os.path.join(
             config["Global"]["save_inference_dir"], "quant_post_static_model"),
         sample_generator=sample_generator(train_dataloader),
+        batch_size=config["DataLoader"]["Eval"]["sampler"]["batch_size"],
         batch_nums=10)
 
 

docs/en/PULC/PULC_language_classification_en.md

+# PULC Classification Model of Language
+
+------
+
+## Catalogue
+
+- [1. Introduction](#1)
+- [2. Quick Start](#2)
+    - [2.1 PaddlePaddle Installation](#2.1)
+    - [2.2 PaddleClas Installation](#2.2)
+    - [2.3 Prediction](#2.3)
+- [3. Training, Evaluation and Inference](#3)
+    - [3.1 Installation](#3.1)
+    - [3.2 Dataset](#3.2)
+      - [3.2.1 Dataset Introduction](#3.2.1)
+      - [3.2.2 Getting Dataset](#3.2.2)
+    - [3.3 Training](#3.3)
+    - [3.4 Evaluation](#3.4)
+    - [3.5 Inference](#3.5)
+- [4. Model Compression](#4)
+  - [4.1 SKL-UGI Knowledge Distillation](#4.1)
+    - [4.1.1 Teacher Model Training](#4.1.1)
+    - [4.1.2 Knowledge Distillation Training](#4.1.2)
+- [5. SHAS](#5)
+- [6. Inference Deployment](#6)
+  - [6.1 Getting Paddle Inference Model](#6.1)
+    - [6.1.1 Exporting Paddle Inference Model](#6.1.1)
+    - [6.1.2 Downloading Inference Model](#6.1.2)
+  - [6.2 Prediction with Python](#6.2)
+    - [6.2.1 Image Prediction](#6.2.1)
+    - [6.2.2 Images Prediction](#6.2.2)
+  - [6.3 Deployment with C++](#6.3)
+  - [6.4 Deployment as Service](#6.4)
+  - [6.5 Deployment on Mobile](#6.5)
+  - [6.6 Converting To ONNX and Deployment](#6.6)
+
+<a name="1"></a>
+
+## 1. Introduction
+
+This case provides a way for users to quickly build a lightweight, high-precision and practical classification model of language in the image using PaddleClas PULC (Practical Ultra Lightweight image Classification). The model can be widely used in various scenarios involving multilingual OCR processing, such as finance and government affairs.
+
+The following table lists the relevant indicators of the model. The first two lines means that using SwinTransformer_tiny and MobileNetV3_small_x0_35 as the backbone to training. The third to sixth lines means that the backbone is replaced by PPLCNet, additional use of EDA strategy and additional use of EDA strategy and SKL-UGI knowledge distillation strategy. When replacing the backbone with PPLCNet_x1_0, the input shape of model is changed to [192, 48], and the stride of the network is changed to [2, [2, 1], [2, 1], [2, 1]].
+
+| Backbone | Top1-Acc(%) | Latency(ms) | Size(M)| Training Strategy |
+| ----------------------- | --------- | -------- | ------- | ---------------------------------------------- |
+| SwinTranformer_tiny     | 98.12     | 89.09    | 111     | using ImageNet pretrained model |
+| MobileNetV3_small_x0_35 | 95.92     | 2.98     | 3.7      | using ImageNet pretrained model |
+| PPLCNet_x1_0            | 98.35     | 2.58     | 7.1     | using ImageNet pretrained model |
+| PPLCNet_x1_0            | 98.7      | 2.58     | 7.1     | using SSLD pretrained model |
+| PPLCNet_x1_0            | 99.12     | 2.58     | 7.1     | using SSLD pretrained model + EDA strategy  |
+| **PPLCNet_x1_0**        | **99.26** | **2.58** | **7.1** | using SSLD pretrained model + EDA strategy + SKL-UGI knowledge distillation strategy|
+
+It can be seen that high accuracy can be getted when backbone is SwinTranformer_tiny, but the speed is slow. Replacing backbone with the lightweight model MobileNetV3_small_x0_35, the speed can be greatly improved, but the accuracy will be greatly reduced. Replacing backbone with faster backbone PPLCNet_x1_0 and changing the input shape and stride of network, the accuracy is higher more 2.43 percentage points than MobileNetv3_small_x0_35. At the same time, the speed can be more than 20% faster. After additional using the SSLD pretrained model, the accuracy can be improved by about 0.35 percentage points without affecting the inference speed. Further, additional using the EDA strategy, the accuracy can be increased by 0.42 percentage points. Finally, after additional using the SKL-UGI knowledge distillation, the accuracy can be further improved by 0.14 percentage points. At this point, the accuracy is higher than that of SwinTranformer_tiny, but the speed is more faster. The training method and deployment instructions of PULC will be introduced in detail below.
+
+**Note**:
+
+* The Latency is tested on Intel(R) Xeon(R) Gold 6148 CPU @ 2.40GHz. The MKLDNN is enabled and the number of threads is 10.
+* About PP-LCNet, please refer to [PP-LCNet Introduction](../models/PP-LCNet_en.md) and [PP-LCNet Paper](https://arxiv.org/abs/2109.15099).
+
+<a name="2"></a>
+
+## 2. Quick Start
+
+<a name="2.1"></a>  
+
+### 2.1 PaddlePaddle Installation
+
+- Run the following command to install if CUDA9 or CUDA10 is available.
+
+```bash
+python3 -m pip install paddlepaddle-gpu -i https://mirror.baidu.com/pypi/simple
+```
+
+- Run the following command to install if GPU device is unavailable.
+
+```bash
+python3 -m pip install paddlepaddle -i https://mirror.baidu.com/pypi/simple
+```
+
+Please refer to [PaddlePaddle Installation](https://www.paddlepaddle.org.cn/install/quick?docurl=/documentation/docs/en/install/pip/linux-pip_en.html) for more information about installation, for examples other versions.
+
+<a name="2.2"></a>  
+
+### 2.2 PaddleClas wheel Installation
+
+The command of PaddleClas installation as bellow:
+
+```bash
+pip3 install paddleclas
+```
+
+<a name="2.3"></a>
+
+### 2.3 Prediction
+
+First, please click [here](https://paddleclas.bj.bcebos.com/data/PULC/pulc_demo_imgs.zip) to download and unzip to get the test demo images.
+
+* Prediction with CLI
+
+```bash
+paddleclas --model_name=language_classification --infer_imgs=pulc_demo_imgs/language_classification/word_35404.png
+```
+
+Results:
+
+```
+>>> result
+class_ids: [4, 6], scores: [0.88672, 0.01434], label_names: ['japan', 'korean'], filename: pulc_demo_imgs/language_classification/word_35404.png
+Predict complete!
+```
+
+**Note**: If you want to test other images, only need to specify the `--infer_imgs` argument, and the directory containing images is also supported.
+
+* Prediction in Python
+
+```python
+import paddleclas
+model = paddleclas.PaddleClas(model_name="language_classification")
+result = model.predict(input_data="pulc_demo_imgs/language_classification/word_35404.png")
+print(next(result))
+```
+
+**Note**: The `result` returned by `model.predict()` is a generator, so you need to use the `next()` function to call it or `for` loop to loop it. And it will predict with `batch_size` size batch and return the prediction results when called. The default `batch_size` is 1, and you also specify the `batch_size` when instantiating, such as `model = paddleclas.PaddleClas(model_name="language_classification",  batch_size=2)`. The result of demo above:
+
+```
+>>> result
+[{'class_ids': [4, 6], 'scores': [0.88672, 0.01434], 'label_names': ['japan', 'korean'], 'filename': 'pulc_demo_imgs/language_classification/word_35404.png'}]
+```
+
+<a name="3"></a>
+
+## 3. Training, Evaluation and Inference
+
+<a name="3.1"></a>  
+
+### 3.1 Installation
+
+Please refer to [Installation](../installation/install_paddleclas_en.md) to get the description about installation.
+
+<a name="3.2"></a>
+
+### 3.2 Dataset
+
+<a name="3.2.1"></a>
+
+#### 3.2.1 Dataset Introduction
+
+The models wo provided are trained with internal data, which is not open source yet. So it is suggested that constructing dataset based on open source dataset [Multi-lingual scene text detection and recognition](https://rrc.cvc.uab.es/?ch=15&com=downloads) to experience the this case.
+
+Some image of the processed dataset is as follows:
+
+![](../../images/PULC/docs/language_classification_original_data.png)
+
+<a name="3.2.2"></a>  
+
+#### 3.2.2 Getting Dataset
+
+The models provided support to classcify 10 languages, which as shown in the following list:
+
+`0` : means Arabic
+`1` : means chinese_cht
+`2` : means cyrillic
+`3` : means devanagari
+`4` : means Japanese
+`5` : means ka
+`6` : means Korean
+`7` : means ta
+`8` : means te
+`9` : means Latin
+
+In the `Multi-lingual scene text detection and recognition`, only Arabic, Japanese, Korean and Latin data are included. 1600 images from each of the four languages are taken as the training data of this case, 300 images as the evaluation data, and 400 images as the supplementary data is used for the `SKL-UGI Knowledge Distillation`.
+
+Therefore, for the demo dataset in this case, the language categories are shown in following list:
+`0` : means arabic
+`4` : means japan
+`6` : means korean
+`9` : means latin
+
+**Note**: The images used in this task should be cropped by text from original image. Only the text line part is used as the image data.
+
+If you want to create your own dataset, you can collect and sort out the data of the required languages in your task as required. And you can also download the data processed directly.
+
+```
+cd path_to_PaddleClas
+```
+
+Enter the `dataset/` directory, download and unzip the dataset.
+
+```shell
+cd dataset
+wget https://paddleclas.bj.bcebos.com/data/PULC/language_classification.tar
+tar -xf language_classification.tar
+cd ../
+```
+
+The datas under `language_classification` directory:
+
+```
+├── img
+│   ├── word_1.png
+│   ├── word_2.png
+...
+├── train_list.txt
+├── train_list_for_distill.txt
+├── test_list.txt
+└── label_list.txt
+```
+
+Where `img/` is the directory including 9200 images in 4 languages. The `train_list.txt` and `test_list.txt` are label files of training data and validation data respectively. `label_list.txt` is the mapping file corresponding to the four languages. `train_list_for_distill.txt` is the label list of images used for `SKL-UGI Knowledge Distillation`.
+
+**Note**:
+
+* About the contents format of `train_list.txt` and `val_list.txt`, please refer to [Description about Classification Dataset in PaddleClas](../data_preparation/classification_dataset_en.md).
+* About the `train_list_for_distill.txt`, please refer to [Knowledge Distillation Label](../advanced_tutorials/distillation/distillation_en.md).
+
+<a name="3.3"></a>
+
+### 3.3 Training
+
+The details of training config in `ppcls/configs/PULC/person_exists/PPLCNet_x1_0.yaml`. The command about training as follows:
+
+```shell
+export CUDA_VISIBLE_DEVICES=0,1,2,3
+python3 -m paddle.distributed.launch \
+    --gpus="0,1,2,3" \
+    tools/train.py \
+        -c ./ppcls/configs/PULC/language_classification/PPLCNet_x1_0.yaml \
+        -o Arch.class_num=4
+```
+
+**Note**: Because the class num of demo dataset is 4, the argument `-o Arch.class_num=4` should be specifed to change the prediction class num of model to 4.
+
+<a name="3.4"></a>
+
+### 3.4 Evaluation
+
+After training, you can use the following commands to evaluate the model.
+
+```bash
+python3 tools/eval.py \
+    -c ./ppcls/configs/PULC/language_classification/PPLCNet_x1_0.yaml \
+    -o Global.pretrained_model="output/PPLCNet_x1_0/best_model" \
+    -o Arch.class_num=4
+```
+
+Among the above command, the argument `-o Global.pretrained_model="output/PPLCNet_x1_0/best_model"` specify the path of the best model weight file. You can specify other path if needed.
+
+<a name="3.5"></a>
+
+### 3.5 Inference
+
+After training, you can use the model that trained to infer. Command is as follow:
+
+```bash
+python3 tools/infer.py \
+    -c ./ppcls/configs/PULC/language_classification/PPLCNet_x1_0.yaml \
+    -o Global.pretrained_model="output/PPLCNet_x1_0/best_model" \
+    -o Arch.class_num=4
+```
+
+The results:
+
+```
+[{'class_ids': [4, 9], 'scores': [0.96809, 0.01001], 'file_name': 'deploy/images/PULC/language_classification/word_35404.png', 'label_names': ['japan', 'latin']}]
+```
+
+**Note**:
+
+* Among the above command, argument `-o Global.pretrained_model="output/PPLCNet_x1_0/best_model"` specify the path of the best model weight file. You can specify other path if needed.
+* The default test image is `deploy/images/PULC/person_exists/objects365_02035329.jpg`. And you can test other image, only need to specify the argument `-o Infer.infer_imgs=path_to_test_image`.
+* Among the prediction results, `japan` means japanese and `korean` means korean.
+
+<a name="4"></a>
+
+## 4. Model Compression
+
+<a name="4.1"></a>
+
+### 4.1 SKL-UGI Knowledge Distillation
+
+SKL-UGI is a simple but effective knowledge distillation algrithem proposed by PaddleClas.
+
+<!-- todo -->
+<!-- Please refer to [SKL-UGI](../advanced_tutorials/distillation/distillation_en.md) for more details. -->
+
+<a name="4.1.1"></a>
+
+#### 4.1.1 Teacher Model Training
+
+Training the teacher model with hyperparameters specified in `ppcls/configs/PULC/language_classification/PPLCNet/PPLCNet_x1_0.yaml`. The command is as follow:
+
+```shell
+export CUDA_VISIBLE_DEVICES=0,1,2,3
+python3 -m paddle.distributed.launch \
+    --gpus="0,1,2,3" \
+    tools/train.py \
+        -c ./ppcls/configs/PULC/language_classification/PPLCNet_x1_0.yaml \
+        -o Arch.name=ResNet101_vd \
+        -o Arch.class_num=4
+```
+
+The best teacher model weight would be saved in file `output/ResNet101_vd/best_model.pdparams`.
+
+**Note**: Training the ResNet101_vd model requires more GPU memory. If the memory is not enough, you can reduce the learning rate and batch size in the same proportion.
+
+<a name="4.1.2"></a>
+
+#### 4.1.2 Knowledge Distillation Training
+
+The training strategy, specified in training config file `ppcls/configs/PULC/language_classification/PPLCNet_x1_0_distillation.yaml`, the teacher model is `ResNet101_vd`, the student model is `PPLCNet_x1_0`. The command is as follow:
+
+```shell
+export CUDA_VISIBLE_DEVICES=0,1,2,3
+python3 -m paddle.distributed.launch \
+    --gpus="0,1,2,3" \
+    tools/train.py \
+        -c ./ppcls/configs/PULC/language_classification/PPLCNet_x1_0_distillation.yaml \
+        -o Arch.models.0.Teacher.pretrained=output/ResNet101_vd/best_model \
+        -o Arch.class_num=4
+```
+
+The best student model weight would be saved in file `output/DistillationModel/best_model_student.pdparams`.
+
+<a name="5"></a>
+
+## 5. Hyperparameters Searching
+
+The hyperparameters used by [3.2 section](#3.2) and [4.1 section](#4.1) are according by `Hyperparameters Searching` in PaddleClas. If you want to get better results on your own dataset, you can refer to [Hyperparameters Searching](PULC_train_en.md#4) to get better hyperparameters.
+
+**Note**: This section is optional. Because the search process will take a long time, you can selectively run according to your specific. If not replace the dataset, you can ignore this section.
+
+<a name="6"></a>
+
+## 6. Inference Deployment
+
+<a name="6.1"></a>
+
+### 6.1 Getting Paddle Inference Model
+
+Paddle Inference is the original Inference Library of the PaddlePaddle, provides high-performance inference for server deployment. And compared with  directly based on the pretrained model, Paddle Inference can use tools to accelerate prediction, so as to achieve better inference performance. Please refer to [Paddle Inference](https://www.paddlepaddle.org.cn/documentation/docs/zh/guides/infer/inference/inference_cn.html) for more information.
+
+Paddle Inference need Paddle Inference Model to predict. Two process provided to get Paddle Inference Model. If want to use the provided by PaddleClas, you can download directly, click [Downloading Inference Model](#6.1.2).
+
+<a name="6.1.1"></a>
+
+### 6.1.1 Exporting Paddle Inference Model
+
+The command about exporting Paddle Inference Model is as follow:
+
+```bash
+python3 tools/export_model.py \
+    -c ./ppcls/configs/PULC/language_classification/PPLCNet_x1_0.yaml \
+    -o Global.pretrained_model=output/DistillationModel/best_model_student \
+    -o Global.save_inference_dir=deploy/models/PPLCNet_x1_0_language_classification_infer
+```
+
+After running above command, the inference model files would be saved in `deploy/models/PPLCNet_x1_0_language_classification_infer`, as shown below:
+
+```
+├── PPLCNet_x1_0_language_classification_infer
+│   ├── inference.pdiparams
+│   ├── inference.pdiparams.info
+│   └── inference.pdmodel
+```
+
+**Note**: The best model is from knowledge distillation training. If knowledge distillation training is not used, the best model would be saved in `output/PPLCNet_x1_0/best_model.pdparams`.
+
+<a name="6.1.2"></a>
+
+### 6.1.2 Downloading Inference Model
+
+You can also download directly.
+
+```
+cd deploy/models
+# download the inference model and decompression
+wget https://paddleclas.bj.bcebos.com/models/PULC/language_classification_infer.tar && tar -xf language_classification_infer.tar
+```
+
+After decompression, the directory `models` should be shown below.
+
+```
+├── language_classification_infer
+│   ├── inference.pdiparams
+│   ├── inference.pdiparams.info
+│   └── inference.pdmodel
+```
+
+<a name="6.2"></a>
+
+### 6.2 Prediction with Python
+
+<a name="6.2.1"></a>  
+
+#### 6.2.1 Image Prediction
+
+Return the directory `deploy`:
+
+```
+cd ../
+```
+
+Run the following command to classify language about the image `./images/PULC/language_classification/word_35404.png`.
+
+```shell
+# Use the following command to predict with GPU.
+python3.7 python/predict_cls.py -c configs/PULC/language_classification/inference_language_classification.yaml
+# Use the following command to predict with CPU.
+python3.7 python/predict_cls.py -c configs/PULC/language_classification/inference_language_classification.yaml -o Global.use_gpu=False
+```
+
+The prediction results:
+
+```
+word_35404.png:    class id(s): [4, 6], score(s): [0.89, 0.01], label_name(s): ['japan', 'korean']
+```
+
+**Note**: Among the prediction results, `japan` means japanese and `korean` means korean.
+
+<a name="6.2.2"></a>  
+
+#### 6.2.2 Images Prediction
+
+If you want to predict images in directory, please specify the argument `Global.infer_imgs` as directory path by `-o Global.infer_imgs`. The command is as follow.
+
+```shell
+# Use the following command to predict with GPU. If want to replace with CPU, you can add argument -o Global.use_gpu=False
+python3.7 python/predict_cls.py -c configs/PULC/language_classification/inference_language_classification.yaml -o Global.infer_imgs="./images/PULC/language_classification/"
+```
+
+All prediction results will be printed, as shown below.
+
+```
+word_17.png:    class id(s): [9, 4], score(s): [0.80, 0.09], label_name(s): ['latin', 'japan']
+word_20.png:    class id(s): [0, 4], score(s): [0.91, 0.02], label_name(s): ['arabic', 'japan']
+word_35404.png:    class id(s): [4, 6], score(s): [0.89, 0.01], label_name(s): ['japan', 'korean']
+```
+
+Among the prediction results above, `japan` means japanese, `latin` means latin, `arabic` means arabic and `korean` means korean.
+
+<a name="6.3"></a>
+
+### 6.3 Deployment with C++
+
+PaddleClas provides an example about how to deploy with C++. Please refer to [Deployment with C++](../inference_deployment/cpp_deploy_en.md).
+
+<a name="6.4"></a>
+
+### 6.4 Deployment as Service
+
+Paddle Serving is a flexible, high-performance carrier for machine learning models, and supports different protocol, such as RESTful, gRPC, bRPC and so on, which provides different deployment solutions for a variety of heterogeneous hardware and operating system environments. Please refer [Paddle Serving](https://github.com/PaddlePaddle/Serving) for more information.
+
+PaddleClas provides an example about how to deploy as service by Paddle Serving. Please refer to [Paddle Serving Deployment](../inference_deployment/paddle_serving_deploy_en.md).
+
+<a name="6.5"></a>
+
+### 6.5 Deployment on Mobile
+
+Paddle-Lite is an open source deep learning framework that designed to make easy to perform inference on mobile, embeded, and IoT devices. Please refer to [Paddle-Lite](https://github.com/PaddlePaddle/Paddle-Lite) for more information.
+
+PaddleClas provides an example of how to deploy on mobile by Paddle-Lite. Please refer to [Paddle-Lite deployment](../inference_deployment/paddle_lite_deploy_en.md).
+
+<a name="6.6"></a>
+
+### 6.6 Converting To ONNX and Deployment
+
+Paddle2ONNX support convert Paddle Inference model to ONNX model. And you can deploy with ONNX model on different inference engine, such as TensorRT, OpenVINO, MNN/TNN, NCNN and so on. About Paddle2ONNX details, please refer to [Paddle2ONNX](https://github.com/PaddlePaddle/Paddle2ONNX).
+
+PaddleClas provides an example of how to convert Paddle Inference model to ONNX model by paddle2onnx toolkit and predict by ONNX model. You can refer to [paddle2onnx](../../../deploy/paddle2onnx/readme_en.md) for deployment details.

docs/en/PULC/PULC_model_list_en.md

@@ -7,15 +7,15 @@ The PULC model zoo is provided here, mainly providing indicators, model storage
 
 |Model name| Model Description | Metrics |Storage Size| Latency| Download Address|
 | --- | --- | --- | --- | --- | --- |
-| person_exists |[Human Exists Classification](PULC_person_exists_en.md)| 95.60 |6.5M|2.58ms|[inference model](https://paddleclas.bj.bcebos.com/models/PULC/inference/person_exists_infer.tar) / [pretrained model](https://paddleclas.bj.bcebos.com/models/PULC/pretrained/person_exists_pretrained.pdparams)|
-| person_attribute |[Pedestrian Attribute Classification](PULC_person_attribute_en.md)| 78.59 |6.6M|2.01ms|[inference model](https://paddleclas.bj.bcebos.com/models/PULC/inference/person_attribute_infer.tar) / [pretrained model](https://paddleclas.bj.bcebos.com/models/PULC/pretrained/person_attribute_pretrained.pdparams)|
-| safety_helmet |[Classification of Wheather Wearing Safety Helmet](PULC_safety_helmet_en.md)| 99.38 |6.5M|2.03ms|[inference model](https://paddleclas.bj.bcebos.com/models/PULC/inference/safety_helmet_infer.tar) / [pretrained model](https://paddleclas.bj.bcebos.com/models/PULC/pretrained/safety_helmet_pretrained.pdparams)|
+| person_exists |[Human Exists Classification](PULC_person_exists_en.md)| 96.23 |7.0M|2.58ms|[inference model](https://paddleclas.bj.bcebos.com/models/PULC/inference/person_exists_infer.tar) / [pretrained model](https://paddleclas.bj.bcebos.com/models/PULC/pretrained/person_exists_pretrained.pdparams)|
+| person_attribute |[Pedestrian Attribute Classification](PULC_person_attribute_en.md)| 78.59 |7.2M|2.01ms|[inference model](https://paddleclas.bj.bcebos.com/models/PULC/inference/person_attribute_infer.tar) / [pretrained model](https://paddleclas.bj.bcebos.com/models/PULC/pretrained/person_attribute_pretrained.pdparams)|
+| safety_helmet |[Classification of Wheather Wearing Safety Helmet](PULC_safety_helmet_en.md)| 99.38 |7.1M|2.03ms|[inference model](https://paddleclas.bj.bcebos.com/models/PULC/inference/safety_helmet_infer.tar) / [pretrained model](https://paddleclas.bj.bcebos.com/models/PULC/pretrained/safety_helmet_pretrained.pdparams)|
 | traffic_sign |[Traffic Sign Classification](PULC_traffic_sign_en.md)| 98.35 |8.2M|2.10ms|[inference model](https://paddleclas.bj.bcebos.com/models/PULC/inference/traffic_sign_infer.tar) / [pretrained model](https://paddleclas.bj.bcebos.com/models/PULC/pretrained/traffic_sign_pretrained.pdparams)|
 | vehicle_attribute |[Vehicle Attribute Classification](PULC_vehicle_attribute_en.md)| 90.81 |7.2M|2.36ms|[inference model](https://paddleclas.bj.bcebos.com/models/PULC/inference/vehicle_attribute_infer.tar) / [pretrained model](https://paddleclas.bj.bcebos.com/models/PULC/pretrained/vehicle_attribute_pretrained.pdparams)|
-| car_exists |[Car Exists Classification](PULC_car_exists_en.md) | 95.92 | 6.6M | 2.38ms |[inference model](https://paddleclas.bj.bcebos.com/models/PULC/inference/car_exists_infer.tar) / [pretrained model](https://paddleclas.bj.bcebos.com/models/PULC/pretrained/car_exists_pretrained.pdparams)|
-| text_image_orientation |[Text Image Orientation Classification](PULC_text_image_orientation_en.md)| 99.06 | 6.5M | 2.16ms |[inference model](https://paddleclas.bj.bcebos.com/models/PULC/inference/text_image_orientation_infer.tar) / [pretrained model](https://paddleclas.bj.bcebos.com/models/PULC/pretrained/text_image_orientation_pretrained.pdparams)|
-| textline_orientation |[Text-line Orientation Classification](PULC_textline_orientation_en.md)| 96.01 |6.5M|2.72ms|[inference model](https://paddleclas.bj.bcebos.com/models/PULC/inference/textline_orientation_infer.tar) / [pretrained model](https://paddleclas.bj.bcebos.com/models/PULC/pretrained/textline_orientation_pretrained.pdparams)|
-| language_classification |[Language Classification](PULC_language_classification_en.md)| 99.26 |6.5M|2.58ms|[inference model](https://paddleclas.bj.bcebos.com/models/PULC/inference/language_classification_infer.tar) / [pretrained model](https://paddleclas.bj.bcebos.com/models/PULC/pretrained/language_classification_pretrained.pdparams)|
+| car_exists |[Car Exists Classification](PULC_car_exists_en.md) | 95.92 | 7.1M | 2.38ms |[inference model](https://paddleclas.bj.bcebos.com/models/PULC/inference/car_exists_infer.tar) / [pretrained model](https://paddleclas.bj.bcebos.com/models/PULC/pretrained/car_exists_pretrained.pdparams)|
+| text_image_orientation |[Text Image Orientation Classification](PULC_text_image_orientation_en.md)| 99.06 | 7.1M | 2.16ms |[inference model](https://paddleclas.bj.bcebos.com/models/PULC/inference/text_image_orientation_infer.tar) / [pretrained model](https://paddleclas.bj.bcebos.com/models/PULC/pretrained/text_image_orientation_pretrained.pdparams)|
+| textline_orientation |[Text-line Orientation Classification](PULC_textline_orientation_en.md)| 96.01 |7.0M|2.72ms|[inference model](https://paddleclas.bj.bcebos.com/models/PULC/inference/textline_orientation_infer.tar) / [pretrained model](https://paddleclas.bj.bcebos.com/models/PULC/pretrained/textline_orientation_pretrained.pdparams)|
+| language_classification |[Language Classification](PULC_language_classification_en.md)| 99.26 |7.1M|2.58ms|[inference model](https://paddleclas.bj.bcebos.com/models/PULC/inference/language_classification_infer.tar) / [pretrained model](https://paddleclas.bj.bcebos.com/models/PULC/pretrained/language_classification_pretrained.pdparams)|
 
 
 **Note：**

docs/en/PULC/PULC_person_attribute_en.md

+# PULC Recognition Model of Person Attribute
+
+------
+
+## Catalogue
+
+- [1. Introduction](#1)
+- [2. Quick Start](#2)
+    - [2.1 PaddlePaddle Installation](#2.1)
+    - [2.2 PaddleClas Installation](#2.2)
+    - [2.3 Prediction](#2.3)
+- [3. Training, Evaluation and Inference](#3)
+    - [3.1 Installation](#3.1)
+    - [3.2 Dataset](#3.2)
+      - [3.2.1 Dataset Introduction](#3.2.1)
+      - [3.2.2 Getting Dataset](#3.2.2)
+    - [3.3 Training](#3.3)
+    - [3.4 Evaluation](#3.4)
+    - [3.5 Inference](#3.5)
+- [4. Model Compression](#4)
+  - [4.1 SKL-UGI Knowledge Distillation](#4.1)
+    - [4.1.1 Teacher Model Training](#4.1.1)
+    - [4.1.2 Knowledge Distillation Training](#4.1.2)
+- [5. SHAS](#5)
+- [6. Inference Deployment](#6)
+  - [6.1 Getting Paddle Inference Model](#6.1)
+    - [6.1.1 Exporting Paddle Inference Model](#6.1.1)
+    - [6.1.2 Downloading Inference Model](#6.1.2)
+  - [6.2 Prediction with Python](#6.2)
+    - [6.2.1 Image Prediction](#6.2.1)
+    - [6.2.2 Images Prediction](#6.2.2)
+  - [6.3 Deployment with C++](#6.3)
+  - [6.4 Deployment as Service](#6.4)
+  - [6.5 Deployment on Mobile](#6.5)
+  - [6.6 Converting To ONNX and Deployment](#6.6)
+
+<a name="1"></a>
+
+## 1. Introduction
+
+This case provides a way for users to quickly build a lightweight, high-precision and practical classification model of person attribute using PaddleClas PULC (Practical Ultra Lightweight image Classification). The model can be widely used in
+Pedestrian analysis scenarios, pedestrian tracking scenarios, etc.
+
+The following table lists the relevant indicators of the model. The first three lines means that using Res2Net200_vd_26w_4s, SwinTransformer_tiny and MobileNetV3_small_x0_35 as the backbone to training. The fourth to seventh lines means that the backbone is replaced by PPLCNet, additional use of EDA strategy and additional use of EDA strategy and SKL-UGI knowledge distillation strategy.
+
+
+| Backbone | ma（%） | Latency(ms) | Size(M) | Training Strategy |
+|-------|-----------|----------|---------------|---------------|
+| Res2Net200_vd_26w_4s  | 81.25 | 77.51  | 293 | using ImageNet pretrained |
+| SwinTransformer_tiny  | 80.17 | 89.51  | 111 | using ImageNet pretrained |
+| MobileNetV3_small_x0_35  | 70.79 | 2.90  | 1.7 | using ImageNet pretrained |
+| PPLCNet_x1_0  | 76.31 | 2.01  | 7.1 | using ImageNet pretrained |
+| PPLCNet_x1_0  | 77.31 | 2.01  | 7.1 | using SSLD pretrained |
+| PPLCNet_x1_0  | 77.71 | 2.01  | 7.1 | using SSLD pretrained + EDA strategy|
+| <b>PPLCNet_x1_0<b>  | <b>78.59<b> | <b>2.01<b>  | <b>7.1<b> | using SSLD pretrained + EDA strategy + SKL-UGI knowledge distillation strategy|
+
+It can be seen that high ma metric can be getted when backbone are Res2Net200_vd_26w_4s and SwinTranformer_tiny, but the speed is slow. Replacing backbone with the lightweight model MobileNetV3_small_x0_35, the speed can be greatly improved, but the ma metric will be greatly reduced. Replacing backbone with faster backbone PPLCNet_x1_0, the ma metric is higher more 5.5 percentage points higher than MobileNetv3_small_x0_35. At the same time, the speed can be more than 20% faster. After additional using the SSLD pretrained model, the ma metric can be improved by about 1 percentage points without affecting the inference speed. Further, additional using the EDA strategy, the ma metric can be increased by 0.4 percentage points. Finally, after additional using the SKL-UGI knowledge distillation, the ma matric can be further improved by 0.88 percentage points. At this time, the ma metric of PPLCNet_x1_0 is only 1.58% different from SwinTransformer_tiny, but the speed is more than 44 times faster. The training method and deployment instructions of PULC will be introduced in detail below.
+
+**Note**:
+
+* The Latency is tested on Intel(R) Xeon(R) Gold 6148 CPU @ 2.40GHz. The MKLDNN is enabled and the number of threads is 10.
+* About PP-LCNet, please refer to [PP-LCNet Introduction](../models/PP-LCNet_en.md) and [PP-LCNet Paper](https://arxiv.org/abs/2109.15099).
+
+<a name="2"></a>
+
+## 2. Quick Start
+
+<a name="2.1"></a>  
+
+### 2.1 PaddlePaddle Installation
+
+- Run the following command to install if CUDA9 or CUDA10 is available.
+
+```bash
+python3 -m pip install paddlepaddle-gpu -i https://mirror.baidu.com/pypi/simple
+```
+
+- Run the following command to install if GPU device is unavailable.
+
+```bash
+python3 -m pip install paddlepaddle -i https://mirror.baidu.com/pypi/simple
+```
+
+Please refer to [PaddlePaddle Installation](https://www.paddlepaddle.org.cn/install/quick?docurl=/documentation/docs/en/install/pip/linux-pip_en.html) for more information about installation, for examples other versions.
+
+<a name="2.2"></a>  
+
+### 2.2 PaddleClas wheel Installation
+
+The command of PaddleClas installation as bellow:
+
+```bash
+pip3 install paddleclas
+```
+
+<a name="2.3"></a>
+
+### 2.3 Prediction
+
+First, please click [here](https://paddleclas.bj.bcebos.com/data/PULC/pulc_demo_imgs.zip) to download and unzip to get the test demo images.
+
+
+* Prediction with CLI
+
+```bash
+paddleclas --model_name=person_attribute --infer_imgs=pulc_demo_imgs/person_attribute/090004.jpg
+```
+
+Results:
+```
+>>> result
+attributes: ['Male', 'Age18-60', 'Back', 'Glasses: False', 'Hat: False', 'HoldObjectsInFront: False', 'Backpack', 'Upper: LongSleeve UpperPlaid', 'Lower:  Trousers', 'No boots'], output: [0, 0, 0, 1, 0, 0, 1, 0, 0, 0, 0, 1, 0, 0, 0, 0, 0, 1, 0, 0, 1, 0, 0, 0, 0, 1], filename: pulc_demo_imgs/person_attribute/090004.jpg
+Predict complete!
+```
+
+**Note**: If you want to test other images, only need to specify the `--infer_imgs` argument, and the directory containing images is also supported.
+
+* Prediction in Python
+
+```python
+import paddleclas
+model = paddleclas.PaddleClas(model_name="person_attribute")
+result = model.predict(input_data="pulc_demo_imgs/person_attribute/090004.jpg")
+print(next(result))
+```
+
+**Note**: The `result` returned by `model.predict()` is a generator, so you need to use the `next()` function to call it or `for` loop to loop it. And it will predict with `batch_size` size batch and return the prediction results when called. The default `batch_size` is 1, and you also specify the `batch_size` when instantiating, such as `model = paddleclas.PaddleClas(model_name="person_attribute",  batch_size=2)`. The result of demo above:
+
+```
+>>> result
+[{'attributes': ['Male', 'Age18-60', 'Back', 'Glasses: False', 'Hat: False', 'HoldObjectsInFront: False', 'Backpack', 'Upper: LongSleeve UpperPlaid', 'Lower:  Trousers', 'No boots'], 'output': [0, 0, 0, 1, 0, 0, 1, 0, 0, 0, 0, 1, 0, 0, 0, 0, 0, 1, 0, 0, 1, 0, 0, 0, 0, 1], 'filename': 'pulc_demo_imgs/person_attribute/090004.jpg'}]
+```
+
+<a name="3"></a>
+
+## 3. Training, Evaluation and Inference
+
+<a name="3.1"></a>  
+
+### 3.1 Installation
+
+Please refer to [Installation](../installation/install_paddleclas_en.md) to get the description about installation.
+
+<a name="3.2"></a>
+
+### 3.2 Dataset
+
+<a name="3.2.1"></a>
+
+#### 3.2.1 Dataset Introduction
+
+The data used in this case is the [pa100k dataset](https://www.v7labs.com/open-datasets/pa-100k).
+
+<a name="3.2.2"></a>  
+
+#### 3.2.2 Getting Dataset
+
+Some image of the processed dataset is as follows:
+
+![](../../images/PULC/docs/person_attribute_data_demo.png)
+
+
+We converted the data into a PaddleClas multi-label readable data format that can be downloaded directly.
+
+```
+cd path_to_PaddleClas
+```
+
+Enter the `dataset/` directory, download and unzip the dataset.
+
+```shell
+cd dataset
+wget https://paddleclas.bj.bcebos.com/data/PULC/pa100k.tar
+tar -xf pa100k.tar
+cd ../
+```
+
+The datas under `pa100k` directory:
+
+```
+pa100k
+├── train
+│   ├── 000001.jpg
+│   ├── 000002.jpg
+...
+├── val
+│   ├── 080001.jpg
+│   ├── 080002.jpg
+...  
+├── test
+│   ├── 090001.jpg
+│   ├── 090002.jpg
+...
+...
+├── train_list.txt
+├── train_val_list.txt
+├── val_list.txt
+├── test_list.txt
+```
+
+Where `train/`, `val/`, `test/` are training set, validation set and test set respectively. `train_list.txt`, `val_list.txt`, `test_list.txt` are the label files of the training set, validation set, and test set, respectively. In this example, `test_list.txt` is not used for now.
+
+
+<a name="3.3"></a>
+
+### 3.3 Training
+
+The details of training config in ./ppcls/configs/PULC/person_attribute/PPLCNet_x1_0.yaml`. The command about training as follows:
+
+```shell
+export CUDA_VISIBLE_DEVICES=0,1,2,3
+python3 -m paddle.distributed.launch \
+    --gpus="0,1,2,3" \
+    tools/train.py \
+        -c ./ppcls/configs/PULC/person_attribute/PPLCNet_x1_0.yaml
+```
+
+The best metric for the validation set is around `77.71%` (the dataset is small and generally fluctuates around 0.3%).
+
+
+<a name="3.4"></a>
+
+### 3.4 Evaluation
+
+After training, you can use the following commands to evaluate the model.
+
+```bash
+python3 tools/eval.py \
+    -c ./ppcls/configs/PULC/person_attribute/PPLCNet_x1_0.yaml \
+    -o Global.pretrained_model="output/PPLCNet_x1_0/best_model"
+```
+
+Among the above command, the argument `-o Global.pretrained_model="output/PPLCNet_x1_0/best_model"` specify the path of the best model weight file. You can specify other path if needed.
+
+<a name="3.5"></a>
+
+### 3.5 Inference
+
+After training, you can use the model that trained to infer. Command is as follow:
+
+```python
+python3 tools/infer.py \
+    -c ./ppcls/configs/PULC/person_attribute/PPLCNet_x1_0.yaml \
+    -o Global.pretrained_model=output/PPLCNet_x1_0/best_model
+```
+
+The results:
+
+```
+[{'attributes': ['Male', 'Age18-60', 'Back', 'Glasses: False', 'Hat: False', 'HoldObjectsInFront: False', 'Backpack', 'Upper: LongSleeve UpperPlaid', 'Lower:  Trousers', 'No boots'], 'output': [0, 0, 0, 1, 0, 0, 1, 0, 0, 0, 0, 1, 0, 0, 0, 0, 0, 1, 0, 0, 1, 0, 0, 0, 0, 1]}]
+```
+
+**Note**:
+
+* Among the above command, argument `-o Global.pretrained_model="output/PPLCNet_x1_0/best_model"` specify the path of the best model weight file. You can specify other path if needed.
+* The default test image is `deploy/images/PULC/person_attribute/090004.jpg`. And you can test other image, only need to specify the argument `-o Infer.infer_imgs=path_to_test_image`.
+
+<a name="4"></a>
+
+## 4. Model Compression
+
+<a name="4.1"></a>
+
+### 4.1 SKL-UGI Knowledge Distillation
+
+SKL-UGI is a simple but effective knowledge distillation algrithem proposed by PaddleClas.
+
+<!-- todo -->
+<!-- Please refer to [SKL-UGI](../advanced_tutorials/distillation/distillation_en.md) for more details. -->
+
+<a name="4.1.1"></a>
+
+#### 4.1.1 Teacher Model Training
+
+Training the teacher model with hyperparameters specified in `ppcls/configs/PULC/person_attribute/PPLCNet_x1_0.yaml`. The command is as follow:
+
+```shell
+export CUDA_VISIBLE_DEVICES=0,1,2,3
+python3 -m paddle.distributed.launch \
+    --gpus="0,1,2,3" \
+    tools/train.py \
+        -c ./ppcls/configs/PULC/person_attribute/PPLCNet_x1_0.yaml \
+        -o Arch.name=ResNet101_vd
+```
+
+The best metric for the validation set is around `80.10%`. The best teacher model weight would be saved in file `output/ResNet101_vd/best_model.pdparams`.
+
+<a name="4.1.2"></a>
+
+#### 4.1.2 Knowledge Distillation Training
+
+The training strategy, specified in training config file `ppcls/configs/PULC/person_attribute/PPLCNet_x1_0_Distillation.yaml`, the teacher model is `ResNet101_vd`, the student model is `PPLCNet_x1_0`. The command is as follow:
+
+```shell
+export CUDA_VISIBLE_DEVICES=0,1,2,3
+python3 -m paddle.distributed.launch \
+    --gpus="0,1,2,3" \
+    tools/train.py \
+        -c ./ppcls/configs/PULC/person_attribute/PPLCNet_x1_0_Distillation.yaml \
+        -o Arch.models.0.Teacher.pretrained=output/ResNet101_vd/best_model
+```
+
+The best metric for the validation set is around `78.5%`. The best student model weight would be saved in file `output/DistillationModel/best_model_student.pdparams`.
+
+<a name="5"></a>
+
+## 5. Hyperparameters Searching
+
+The hyperparameters used by [3.2 section](#3.2) and [4.1 section](#4.1) are according by `Hyperparameters Searching` in PaddleClas. If you want to get better results on your own dataset, you can refer to [Hyperparameters Searching](PULC_train_en.md#4) to get better hyperparameters.
+
+**Note**: This section is optional. Because the search process will take a long time, you can selectively run according to your specific. If not replace the dataset, you can ignore this section.
+
+<a name="6"></a>
+
+## 6. Inference Deployment
+
+<a name="6.1"></a>
+
+### 6.1 Getting Paddle Inference Model
+
+Paddle Inference is the original Inference Library of the PaddlePaddle, provides high-performance inference for server deployment. And compared with  directly based on the pretrained model, Paddle Inference can use tools to accelerate prediction, so as to achieve better inference performance. Please refer to [Paddle Inference](https://www.paddlepaddle.org.cn/documentation/docs/zh/guides/infer/inference/inference_cn.html) for more information.
+
+Paddle Inference need Paddle Inference Model to predict. Two process provided to get Paddle Inference Model. If want to use the provided by PaddleClas, you can download directly, click [Downloading Inference Model](#6.1.2).
+
+<a name="6.1.1"></a>
+
+### 6.1.1 Exporting Paddle Inference Model
+
+The command about exporting Paddle Inference Model is as follow:
+
+```bash
+python3 tools/export_model.py \
+    -c ./ppcls/configs/PULC/person_attribute/PPLCNet_x1_0.yaml \
+    -o Global.pretrained_model=output/DistillationModel/best_model_student \
+    -o Global.save_inference_dir=deploy/models/PPLCNet_x1_0_person_attribute_infer
+```
+
+After running above command, the inference model files would be saved in `PPLCNet_x1_0_person_attribute_infer`, as shown below:
+
+```
+├── PPLCNet_x1_0_person_attribute_infer
+│   ├── inference.pdiparams
+│   ├── inference.pdiparams.info
+│   └── inference.pdmodel
+```
+
+**Note**: The best model is from knowledge distillation training. If knowledge distillation training is not used, the best model would be saved in `output/PPLCNet_x1_0/best_model.pdparams`.
+
+<a name="6.1.2"></a>
+
+### 6.1.2 Downloading Inference Model
+
+You can also download directly.
+
+```
+cd deploy/models
+# download the inference model and decompression
+wget https://paddleclas.bj.bcebos.com/models/PULC/person_attribute_infer.tar && tar -xf person_attribute_infer.tar
+```
+
+After decompression, the directory `models` should be shown below.
+
+```
+├── person_attribute_infer
+│   ├── inference.pdiparams
+│   ├── inference.pdiparams.info
+│   └── inference.pdmodel
+```
+
+<a name="6.2"></a>
+
+### 6.2 Prediction with Python
+
+<a name="6.2.1"></a>  
+
+#### 6.2.1 Image Prediction
+
+Return the directory `deploy`:
+
+```
+cd ../
+```
+
+Run the following command to classify whether there are human in the image `./images/PULC/person_attribute/090004.jpg`.
+
+```shell
+# Use the following command to predict with GPU.
+python3.7 python/predict_cls.py -c configs/PULC/person_attribute/inference_person_attribute.yaml -o Global.use_gpu=True
+# Use the following command to predict with CPU.
+python3.7 python/predict_cls.py -c configs/PULC/person_attribute/inference_person_attribute.yaml -o Global.use_gpu=False
+```
+
+The prediction results:
+
+```
+090004.jpg:     {'attributes': ['Male', 'Age18-60', 'Back', 'Glasses: False', 'Hat: False', 'HoldObjectsInFront: False', 'Backpack', 'Upper: LongSleeve UpperPlaid', 'Lower:  Trousers', 'No boots'], 'output': [0, 0, 0, 1, 0, 0, 1, 0, 0, 0, 0, 1, 0, 0, 0, 0, 0, 1, 0, 0, 1, 0, 0, 0, 0, 1]}
+```
+
+
+<a name="6.2.2"></a>  
+
+#### 6.2.2 Images Prediction
+
+If you want to predict images in directory, please specify the argument `Global.infer_imgs` as directory path by `-o Global.infer_imgs`. The command is as follow.
+
+```shell
+# Use the following command to predict with GPU. If want to replace with CPU, you can add argument -o Global.use_gpu=False
+python3.7 python/predict_cls.py -c configs/PULC/person_attribute/inference_person_attribute.yaml -o Global.infer_imgs="./images/PULC/person_attribute/"
+```
+
+All prediction results will be printed, as shown below.
+
+```
+090004.jpg:     {'attributes': ['Male', 'Age18-60', 'Back', 'Glasses: False', 'Hat: False', 'HoldObjectsInFront: False', 'Backpack', 'Upper: LongSleeve UpperPlaid', 'Lower:  Trousers', 'No boots'], 'output': [0, 0, 0, 1, 0, 0, 1, 0, 0, 0, 0, 1, 0, 0, 0, 0, 0, 1, 0, 0, 1, 0, 0, 0, 0, 1]}
+090007.jpg:     {'attributes': ['Female', 'Age18-60', 'Side', 'Glasses: False', 'Hat: False', 'HoldObjectsInFront: False', 'No bag', 'Upper: ShortSleeve', 'Lower:  Skirt&Dress', 'No boots'], 'output': [0, 0, 1, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 1, 0, 0, 0, 0, 0, 0, 1, 0, 1, 0, 1, 0]}
+```
+
+Among the prediction results above, `someone` means that there is a human in the image, `nobody` means that there is no human in the image.
+
+<a name="6.3"></a>
+
+### 6.3 Deployment with C++
+
+PaddleClas provides an example about how to deploy with C++. Please refer to [Deployment with C++](../inference_deployment/cpp_deploy_en.md).
+
+<a name="6.4"></a>
+
+### 6.4 Deployment as Service
+
+Paddle Serving is a flexible, high-performance carrier for machine learning models, and supports different protocol, such as RESTful, gRPC, bRPC and so on, which provides different deployment solutions for a variety of heterogeneous hardware and operating system environments. Please refer [Paddle Serving](https://github.com/PaddlePaddle/Serving) for more information.
+
+PaddleClas provides an example about how to deploy as service by Paddle Serving. Please refer to [Paddle Serving Deployment](../inference_deployment/paddle_serving_deploy_en.md).
+
+<a name="6.5"></a>
+
+### 6.5 Deployment on Mobile
+
+Paddle-Lite is an open source deep learning framework that designed to make easy to perform inference on mobile, embeded, and IoT devices. Please refer to [Paddle-Lite](https://github.com/PaddlePaddle/Paddle-Lite) for more information.
+
+PaddleClas provides an example of how to deploy on mobile by Paddle-Lite. Please refer to [Paddle-Lite deployment](../inference_deployment/paddle_lite_deploy_en.md).
+
+<a name="6.6"></a>
+
+### 6.6 Converting To ONNX and Deployment
+
+Paddle2ONNX support convert Paddle Inference model to ONNX model. And you can deploy with ONNX model on different inference engine, such as TensorRT, OpenVINO, MNN/TNN, NCNN and so on. About Paddle2ONNX details, please refer to [Paddle2ONNX](https://github.com/PaddlePaddle/Paddle2ONNX).
+
+PaddleClas provides an example of how to convert Paddle Inference model to ONNX model by paddle2onnx toolkit and predict by ONNX model. You can refer to [paddle2onnx](../../../deploy/paddle2onnx/readme_en.md) for deployment details.

docs/en/PULC/PULC_safety_helmet_en.md

+# PULC Classification Model of Wheather Wearing Safety Helmet or Not
+
+-----
+
+## Catalogue
+
+- [1. Introduction](#1)
+- [2. Quick Start](#2)
+    - [2.1 PaddlePaddle Installation](#2.1)
+    - [2.2 PaddleClas Installation](#2.2)
+    - [2.3 Prediction](#2.3)
+- [3. Training, Evaluation and Inference](#3)
+    - [3.1 Installation](#3.1)
+    - [3.2 Dataset](#3.2)
+      - [3.2.1 Dataset Introduction](#3.2.1)
+      - [3.2.2 Getting Dataset](#3.2.2)
+    - [3.3 Training](#3.3)
+    - [3.4 Evaluation](#3.4)
+    - [3.5 Inference](#3.5)
+- [4. Model Compression](#4)
+  - [4.1 SKL-UGI Knowledge Distillation](#4.1)
+    - [4.1.1 Teacher Model Training](#4.1.1)
+    - [4.1.2 Knowledge Distillation Training](#4.1.2)
+- [5. SHAS](#5)
+- [6. Inference Deployment](#6)
+  - [6.1 Getting Paddle Inference Model](#6.1)
+    - [6.1.1 Exporting Paddle Inference Model](#6.1.1)
+    - [6.1.2 Downloading Inference Model](#6.1.2)
+  - [6.2 Prediction with Python](#6.2)
+    - [6.2.1 Image Prediction](#6.2.1)
+    - [6.2.2 Images Prediction](#6.2.2)
+  - [6.3 Deployment with C++](#6.3)
+  - [6.4 Deployment as Service](#6.4)
+  - [6.5 Deployment on Mobile](#6.5)
+  - [6.6 Converting To ONNX and Deployment](#6.6)
+
+<a name="1"></a>
+
+## 1. Introduction
+
+This case provides a way for users to quickly build a lightweight, high-precision and practical classification model of wheather wearing safety helmet using PaddleClas PULC (Practical Ultra Lightweight image Classification). The model can be widely used in construction scenes, factory workshop scenes, traffic scenes and so on.
+
+The following table lists the relevant indicators of the model. The first three lines means that using SwinTransformer_tiny, Res2Net200_vd_26w_4s and MobileNetV3_small_x0_35 as the backbone to training. The fourth to seventh lines means that the backbone is replaced by PPLCNet, additional use of EDA strategy and additional use of EDA strategy and SKL-UGI knowledge distillation strategy.
+
+| Backbone | Tpr(%) | Latency(ms) | Size(M)| Training Strategy |
+|-------|-----------|----------|---------------|---------------|
+| SwinTranformer_tiny  | 93.57 | 91.32  | 111 | using ImageNet pretrained model |
+| Res2Net200_vd_26w_4s  | 98.92 | 80.99 | 284 | using ImageNet pretrained model |
+| MobileNetV3_small_x0_35  | 84.83 | 2.85 | 2.6 | using ImageNet pretrained model |
+| PPLCNet_x1_0  | 93.27 | 2.03  | 7.1 | using ImageNet pretrained model |
+| PPLCNet_x1_0  | 98.16 | 2.03  | 7.1 | using SSLD pretrained model |
+| PPLCNet_x1_0  | 99.30 | 2.03  | 7.1 | using SSLD pretrained model + EDA strategy  |
+| <b>PPLCNet_x1_0<b>  | <b>99.38<b> | <b>2.03<b>  | <b>7.1<b> | using SSLD pretrained model + EDA strategy + SKL-UGI knowledge distillation strategy|
+
+It can be seen that high Tpr can be getted when backbone is Res2Net200_vd_26w_4s, but the speed is slow. Replacing backbone with the lightweight model MobileNetV3_small_x0_35, the speed can be greatly improved, but the Tpr will be greatly reduced. Replacing backbone with faster backbone PPLCNet_x1_0, the Tpr is higher more 8.5 percentage points than MobileNetv3_small_x0_35. At the same time, the speed can be more than 20% faster. After additional using the SSLD pretrained model, the Tpr can be improved by about 4.9 percentage points without affecting the inference speed. Further, additional using the EDA strategy, the Tpr can be increased by 1.1 percentage points. Finally, after additional using the UDML knowledge distillation, the Tpr can be further improved by 2.2 percentage points. At this point, the Tpr is higher than that of Res2Net200_vd_26w_4s, but the speed is more than 70 times faster. The training method and deployment instructions of PULC will be introduced in detail below.
+
+**Note**:
+
+* About `Tpr` metric, please refer to [3.2 section](#3.2) for more information .
+* The Latency is tested on Intel(R) Xeon(R) Gold 6148 CPU @ 2.40GHz. The MKLDNN is enabled and the number of threads is 10.
+* About PP-LCNet, please refer to [PP-LCNet Introduction](../models/PP-LCNet_en.md) and [PP-LCNet Paper](https://arxiv.org/abs/2109.15099).
+
+<a name="2"></a>
+
+## 2. Quick Start
+
+<a name="2.1"></a>  
+
+### 2.1 PaddlePaddle Installation
+
+- Run the following command to install if CUDA9 or CUDA10 is available.
+
+```bash
+python3 -m pip install paddlepaddle-gpu -i https://mirror.baidu.com/pypi/simple
+```
+
+- Run the following command to install if GPU device is unavailable.
+
+```bash
+python3 -m pip install paddlepaddle -i https://mirror.baidu.com/pypi/simple
+```
+
+Please refer to [PaddlePaddle Installation](https://www.paddlepaddle.org.cn/install/quick?docurl=/documentation/docs/en/install/pip/linux-pip_en.html) for more information about installation, for examples other versions.
+
+<a name="2.2"></a>  
+
+### 2.2 PaddleClas wheel Installation
+
+The command of PaddleClas installation as bellow:
+
+```bash
+pip3 install paddleclas
+```
+
+<a name="2.3"></a>
+
+### 2.3 Prediction
+
+First, please click [here](https://paddleclas.bj.bcebos.com/data/PULC/pulc_demo_imgs.zip) to download and unzip to get the test demo images.
+
+* Prediction with CLI
+
+```bash
+paddleclas --model_name=safety_helmet --infer_imgs=pulc_demo_imgs/safety_helmet/safety_helmet_test_1.png
+```
+
+Results:
+
+```
+>>> result
+class_ids: [1], scores: [0.9986255], label_names: ['unwearing_helmet'], filename: pulc_demo_imgs/safety_helmet/safety_helmet_test_1.png
+Predict complete!
+```
+**Note**: If you want to test other images, only need to specify the `--infer_imgs` argument, and the directory containing images is also supported.
+
+* Prediction in Python
+
+```python
+import paddleclas
+model = paddleclas.PaddleClas(model_name="safety_helmet")
+result = model.predict(input_data="pulc_demo_imgs/safety_helmet/safety_helmet_test_1.png")
+print(next(result))
+```
+
+**Note**: The `result` returned by `model.predict()` is a generator, so you need to use the `next()` function to call it or `for` loop to loop it. And it will predict with `batch_size` size batch and return the prediction results when called. The default `batch_size` is 1, and you also specify the `batch_size` when instantiating, such as `model = paddleclas.PaddleClas(model_name="safety_helmet",  batch_size=2)`. The result of demo above:
+
+```
+>>> result
+[{'class_ids': [1], 'scores': [0.9986255], 'label_names': ['unwearing_helmet'], 'filename': 'pulc_demo_imgs/safety_helmet/safety_helmet_test_1.png'}]
+```
+
+<a name="3"></a>
+
+## 3. Training, Evaluation and Inference
+
+<a name="3.1"></a>  
+
+### 3.1 Installation
+
+Please refer to [Installation](../installation/install_paddleclas_en.md) to get the description about installation.
+
+<a name="3.2"></a>
+
+### 3.2 Dataset
+
+<a name="3.2.1"></a>
+
+#### 3.2.1 Dataset Introduction
+
+All datasets used in this case are open source data. Train data is the subset of [Safety-Helmet-Wearing-Dataset](https://github.com/njvisionpower/Safety-Helmet-Wearing-Dataset), [hard-hat-detection](https://www.kaggle.com/datasets/andrewmvd/hard-hat-detection) and [Large-scale CelebFaces Attributes (CelebA) Dataset](https://mmlab.ie.cuhk.edu.hk/projects/CelebA.html).
+
+<a name="3.2.2"></a>  
+
+#### 3.2.2 Getting Dataset
+
+The data used in this case can be getted by processing the open source data. The detailed processes are as follows:
+
+* `Safety-Helmet-Wearing-Dataset`: according to the bbox label data, the image is cropped by enlarging width and height of bbox by 3 times. The label is 0 if wearing safety helmet in the image, and the label is 1 if not;
+* `hard-hat-detection`: Only use the image that labeled "hat" and crop it with bbox. The label is 0;
+* `CelebA`: Only use the image labeled "wearing_hat" and crop it with bbox. The label is 0;
+
+After processing, the dataset totals about 150000 images, of which the number of images with and without wearing safety helmet is about 28000 and 121000 respectively. Then 5600 images are randomly selected in the two labels as the valuation data, a total of about 11200 images, and about 138000 other images as the training data.
+
+Some image of the processed dataset is as follows:
+
+![](../../images/PULC/docs/safety_helmet_data_demo.jpg)
+
+And you can also download the data processed directly.
+
+```
+cd path_to_PaddleClas
+```
+
+Enter the `dataset/` directory, download and unzip the dataset.
+
+```shell
+cd dataset
+wget https://paddleclas.bj.bcebos.com/data/PULC/safety_helmet.tar
+tar -xf safety_helmet.tar
+cd ../
+```
+
+The datas under `safety_helmet` directory:
+
+```
+├── images
+│   ├── VOC2028_part2_001209_1.jpg
+│   ├── HHD_hard_hat_workers23_1.jpg
+│   ├── CelebA_077809.jpg
+│   ├── ...
+│   └── ...
+├── train_list.txt
+└── val_list.txt
+```
+
+The `train_list.txt` and `val_list.txt` are label files of training data and validation data respectively. All images in `images/` directory.
+
+**Note**:
+
+* About the contents format of `train_list.txt` and `val_list.txt`, please refer to [Description about Classification Dataset in PaddleClas](../data_preparation/classification_dataset_en.md).
+
+<a name="3.3"></a>
+
+### 3.3 Training
+
+The details of training config in `ppcls/configs/PULC/person_exists/PPLCNet_x1_0.yaml`. The command about training as follows:
+
+```shell
+export CUDA_VISIBLE_DEVICES=0,1,2,3
+python3 -m paddle.distributed.launch \
+    --gpus="0,1,2,3" \
+    tools/train.py \
+    -c ./ppcls/configs/PULC/safety_helmet/PPLCNet_x1_0.yaml
+```
+
+The best metric of validation data is between `0.985` and `0.993`. There would be fluctuations because the data size is small.
+
+**Note**:
+
+* The metric Tpr, that describe the True Positive Rate when False Positive Rate is less than a certain threshold(1/10000 used in this case), is one of the commonly used metric for binary classification. About the details of Fpr and Tpr, please refer [here](https://en.wikipedia.org/wiki/Receiver_operating_characteristic).
+* When evaluation, the best metric TprAtFpr will be printed that include `Fpr`, `Tpr` and the current `threshold`. The `Tpr` means the Recall rate under the current `Fpr`. The `Tpr` higher, the model better. The `threshold` would be used in deployment, which means the classification threshold under best `Fpr` metric.
+
+<a name="3.4"></a>
+
+### 3.4 Evaluation
+
+After training, you can use the following commands to evaluate the model.
+
+
+```bash
+python3 tools/eval.py \
+    -c ./ppcls/configs/PULC/safety_helmet/PPLCNet_x1_0.yaml \
+    -o Global.pretrained_model=output/PPLCNet_x1_0/best_model
+```
+
+Among the above command, the argument `-o Global.pretrained_model="output/PPLCNet_x1_0/best_model"` specify the path of the best model weight file. You can specify other path if needed.
+
+<a name="3.5"></a>
+
+### 3.5 Inference
+
+After training, you can use the model that trained to infer. Command is as follow:
+
+```python
+python3 tools/infer.py \
+    -c ./ppcls/configs/PULC/safety_helmet/PPLCNet_x1_0.yaml \
+    -o Global.pretrained_model=output/PPLCNet_x1_0/best_model
+```
+
+The results:
+
+```
+[{'class_ids': [1], 'scores': [0.9524797], 'label_names': ['unwearing_helmet'], 'file_name': 'deploy/images/PULC/safety_helmet/safety_helmet_test_1.png'}]
+```
+
+**备注：**
+
+* Among the above command, argument `-o Global.pretrained_model="output/PPLCNet_x1_0/best_model"` specify the path of the best model weight file. You can specify other path if needed.
+* The default test image is `deploy/images/PULC/safety_helmet/safety_helmet_test_1.png`. And you can test other image, only need to specify the argument `-o Infer.infer_imgs=path_to_test_image`.
+* The default threshold is `0.5`. If needed, you can specify the argument `Infer.PostProcess.threshold`, such as: `-o Infer.PostProcess.threshold=0.9167`. And the argument `threshold` is needed to be specified according by specific case. The `0.9167` is the best threshold when `Fpr` is less than `1/10000` in this valuation dataset.
+
+<a name="4"></a>
+
+## 4. Model Compression
+
+<a name="4.1"></a>
+
+### 4.1 UDML  Knowledge Distillation
+
+UDML is a simple but effective knowledge distillation algrithem proposed by PaddleClas. Please refer to [UDML 知识蒸馏](../advanced_tutorials/knowledge_distillation_en.md#1.2.3) for more details.
+
+<a name="4.1.1"></a>
+
+#### 4.1.1  Knowledge Distillation Training
+
+Training with hyperparameters specified in `ppcls/configs/PULC/safety_helmet/PPLCNet_x1_0_distillation.yaml`. The command is as follow:
+
+```shell
+export CUDA_VISIBLE_DEVICES=0,1,2,3
+python3 -m paddle.distributed.launch \
+    --gpus="0,1,2,3" \
+    tools/train.py \
+    -c ./ppcls/configs/PULC/safety_helmet/PPLCNet_x1_0_distillation.yaml
+```
+
+The best metric is between `0.990` and `0.993`. The best student model weight would be saved in file `output/DistillationModel/best_model_student.pdparams`.
+
+<a name="5"></a>
+
+## 5. Hyperparameters Searching
+
+The hyperparameters used by [3.2 section](#3.2) and [4.1 section](#4.1) are according by `Hyperparameters Searching` in PaddleClas. If you want to get better results on your own dataset, you can refer to [Hyperparameters Searching](PULC_train_en.md#4) to get better hyperparameters.
+
+**Note**: This section is optional. Because the search process will take a long time, you can selectively run according to your specific. If not replace the dataset, you can ignore this section.
+
+<a name="6"></a>
+
+## 6. Inference Deployment
+
+<a name="6.1"></a>
+
+### 6.1 Getting Paddle Inference Model
+
+Paddle Inference is the original Inference Library of the PaddlePaddle, provides high-performance inference for server deployment. And compared with  directly based on the pretrained model, Paddle Inference can use tools to accelerate prediction, so as to achieve better inference performance. Please refer to [Paddle Inference](https://www.paddlepaddle.org.cn/documentation/docs/zh/guides/infer/inference/inference_cn.html) for more information.
+
+Paddle Inference need Paddle Inference Model to predict. Two process provided to get Paddle Inference Model. If want to use the provided by PaddleClas, you can download directly, click [Downloading Inference Model](#6.1.2).
+
+<a name="6.1.1"></a>
+
+### 6.1.1 Exporting Paddle Inference Model
+
+The command about exporting Paddle Inference Model is as follow:
+
+```bash
+python3 tools/export_model.py \
+    -c ./ppcls/configs/PULC/safety_helmet/PPLCNet_x1_0.yaml \
+    -o Global.pretrained_model=output/DistillationModel/best_model_student \
+    -o Global.save_inference_dir=deploy/models/PPLCNet_x1_0_safety_helmet_infer
+```
+
+After running above command, the inference model files would be saved in `deploy/models/PPLCNet_x1_0_safety_helmet_infer`, as shown below:
+
+```
+├── PPLCNet_x1_0_safety_helmet_infer
+│   ├── inference.pdiparams
+│   ├── inference.pdiparams.info
+│   └── inference.pdmodel
+```
+
+**Note**: The best model is from knowledge distillation training. If knowledge distillation training is not used, the best model would be saved in `output/PPLCNet_x1_0/best_model.pdparams`.
+
+<a name="6.1.2"></a>
+
+### 6.1.2 Downloading Inference Model
+
+You can also download directly.
+
+```
+cd deploy/models
+# download the inference model and decompression
+wget https://paddleclas.bj.bcebos.com/models/PULC/safety_helmet_infer.tar && tar -xf safety_helmet_infer.tar
+```
+
+After decompression, the directory `models` should be shown below.
+
+```
+├── safety_helmet_infer
+│   ├── inference.pdiparams
+│   ├── inference.pdiparams.info
+│   └── inference.pdmodel
+```
+
+<a name="6.2"></a>
+
+### 6.2 Prediction with Python
+
+<a name="6.2.1"></a>  
+
+#### 6.2.1 Image Prediction
+
+Return the directory `deploy`:
+
+```
+cd ../
+```
+
+Run the following command to classify whether wearing safety helmet about the image `./images/PULC/safety_helmet/safety_helmet_test_1.png`.
+
+```shell
+# Use the following command to predict with GPU.
+python3.7 python/predict_cls.py -c configs/PULC/safety_helmet/inference_safety_helmet.yaml
+# Use the following command to predict with CPU.
+python3.7 python/predict_cls.py -c configs/PULC/safety_helmet/inference_safety_helmet.yaml -o Global.use_gpu=False
+```
+
+The prediction results:
+
+```
+safety_helmet_test_1.png:       class id(s): [1], score(s): [1.00], label_name(s): ['unwearing_helmet']
+```
+
+**Note**: The default threshold is `0.5`. If needed, you can specify the argument `Infer.PostProcess.threshold`, such as: `-o Infer.PostProcess.threshold=0.9167`. And the argument `threshold` is needed to be specified according by specific case. The `0.9167` is the best threshold when `Fpr` is less than `1/10000` in this valuation dataset. Please refer to [3.3 section](#3.3) for details.
+
+<a name="6.2.2"></a>  
+
+#### 6.2.2 Images Prediction
+
+If you want to predict images in directory, please specify the argument `Global.infer_imgs` as directory path by `-o Global.infer_imgs`. The command is as follow.
+
+```shell
+# Use the following command to predict with GPU. If want to replace with CPU, you can add argument -o Global.use_gpu=False
+python3.7 python/predict_cls.py -c configs/PULC/safety_helmet/inference_safety_helmet.yaml -o Global.infer_imgs="./images/PULC/safety_helmet/"
+```
+
+All prediction results will be printed, as shown below.
+
+```
+safety_helmet_test_1.png:       class id(s): [1], score(s): [1.00], label_name(s): ['unwearing_helmet']
+safety_helmet_test_2.png:       class id(s): [0], score(s): [1.00], label_name(s): ['wearing_helmet']
+```
+
+Among the prediction results above, `wearing_helmet` means that wearing safety helmet about the image, `unwearing_helmet` means not.
+
+<a name="6.3"></a>
+
+### 6.3 Deployment with C++
+
+PaddleClas provides an example about how to deploy with C++. Please refer to [Deployment with C++](../inference_deployment/cpp_deploy_en.md).
+
+<a name="6.4"></a>
+
+### 6.4 Deployment as Service
+
+Paddle Serving is a flexible, high-performance carrier for machine learning models, and supports different protocol, such as RESTful, gRPC, bRPC and so on, which provides different deployment solutions for a variety of heterogeneous hardware and operating system environments. Please refer [Paddle Serving](https://github.com/PaddlePaddle/Serving) for more information.
+
+PaddleClas provides an example about how to deploy as service by Paddle Serving. Please refer to [Paddle Serving Deployment](../inference_deployment/paddle_serving_deploy_en.md).
+
+<a name="6.5"></a>
+
+### 6.5 Deployment on Mobile
+
+Paddle-Lite is an open source deep learning framework that designed to make easy to perform inference on mobile, embeded, and IoT devices. Please refer to [Paddle-Lite](https://github.com/PaddlePaddle/Paddle-Lite) for more information.
+
+PaddleClas provides an example of how to deploy on mobile by Paddle-Lite. Please refer to [Paddle-Lite deployment](../inference_deployment/paddle_lite_deploy_en.md).
+
+<a name="6.6"></a>
+
+### 6.6 Converting To ONNX and Deployment
+
+Paddle2ONNX support convert Paddle Inference model to ONNX model. And you can deploy with ONNX model on different inference engine, such as TensorRT, OpenVINO, MNN/TNN, NCNN and so on. About Paddle2ONNX details, please refer to [Paddle2ONNX](https://github.com/PaddlePaddle/Paddle2ONNX).
+
+PaddleClas provides an example of how to convert Paddle Inference model to ONNX model by paddle2onnx toolkit and predict by ONNX model. You can refer to [paddle2onnx](../../../deploy/paddle2onnx/readme_en.md) for deployment details.

docs/en/PULC/PULC_textline_orientation_en.md

+# PULC Classification Model of Textline Orientation
+
+------
+
+## Catalogue
+
+- [1. Introduction](#1)
+- [2. Quick Start](#2)
+    - [2.1 PaddlePaddle Installation](#2.1)
+    - [2.2 PaddleClas Installation](#2.2)
+    - [2.3 Prediction](#2.3)
+- [3. Training, Evaluation and Inference](#3)
+    - [3.1 Installation](#3.1)
+    - [3.2 Dataset](#3.2)
+      - [3.2.1 Dataset Introduction](#3.2.1)
+      - [3.2.2 Getting Dataset](#3.2.2)
+    - [3.3 Training](#3.3)
+    - [3.4 Evaluation](#3.4)
+    - [3.5 Inference](#3.5)
+- [4. Model Compression](#4)
+  - [4.1 SKL-UGI Knowledge Distillation](#4.1)
+    - [4.1.1 Teacher Model Training](#4.1.1)
+    - [4.1.2 Knowledge Distillation Training](#4.1.2)
+- [5. SHAS](#5)
+- [6. Inference Deployment](#6)
+  - [6.1 Getting Paddle Inference Model](#6.1)
+    - [6.1.1 Exporting Paddle Inference Model](#6.1.1)
+    - [6.1.2 Downloading Inference Model](#6.1.2)
+  - [6.2 Prediction with Python](#6.2)
+    - [6.2.1 Image Prediction](#6.2.1)
+    - [6.2.2 Images Prediction](#6.2.2)
+  - [6.3 Deployment with C++](#6.3)
+  - [6.4 Deployment as Service](#6.4)
+  - [6.5 Deployment on Mobile](#6.5)
+  - [6.6 Converting To ONNX and Deployment](#6.6)
+
+<a name="1"></a>
+
+## 1. Introduction
+
+This case provides a way for users to quickly build a lightweight, high-precision and practical classification model of textline orientation using PaddleClas PULC (Practical Ultra Lightweight image Classification). The model can be widely used in character correction, character recognition, etc.
+
+The following table lists the relevant indicators of the model. The first two lines means that using SwinTransformer_tiny and MobileNetV3_small_x0_35 as the backbone to training. The third to seventh lines means that the backbone is replaced by PPLCNet, additional use of EDA strategy and additional use of EDA strategy and SKL-UGI knowledge distillation strategy.
+
+| Backbone | Top-1 Acc(%) | Latency(ms) | Size(M)| Training Strategy |
+|-------|-----------|----------|---------------|---------------|
+| SwinTranformer_tiny  | 93.61 | 89.64  | 111 |  using ImageNet pretrained model |
+| MobileNetV3_small_x0_35  | 81.40 | 2.96  | 2.6 |  using ImageNet pretrained model |
+| PPLCNet_x1_0  | 89.99 | 2.11  | 7.0 |  using ImageNet pretrained model |
+| PPLCNet_x1_0*  | 94.06 | 2.68  | 7.0 | using ImageNet pretrained model |
+| PPLCNet_x1_0*  | 94.11 | 2.68  | 7.0 | using SSLD pretrained model |
+| <b>PPLCNet_x1_0**<b>  | <b>96.01<b> | <b>2.72<b>  | <b>7.0<b> | using SSLD pretrained model + EDA strategy  |
+| PPLCNet_x1_0**  | 95.86 | 2.72  | 7.0 | using SSLD pretrained model + EDA strategy + SKL-UGI knowledge distillation strategy|
+
+It can be seen that high accuracy can be getted when backbone is SwinTranformer_tiny, but the speed is slow. Replacing backbone with the lightweight model MobileNetV3_small_x0_35, the speed can be greatly improved, but the accuracy will be greatly reduced. Replacing backbone with faster backbone PPLCNet_x1_0, the accuracy is higher more 8.6 percentage points than MobileNetv3_small_x0_35. At the same time, the speed can be more than 10% faster. On this basis, by changing the resolution and stripe (refer to [PaddleOCR](https://github.com/PaddlePaddle/PaddleOCR)), the speed becomes 27% slower, but the accuracy can be improved by 4.5 percentage points. After additional using the SSLD pretrained model, the accuracy can be improved by about 0.05 percentage points without affecting the inference speed. Finally, additional using the EDA strategy, the accuracy can be increased by 1.9 percentage points. The training method and deployment instructions of PULC will be introduced in detail below.
+
+**Note**:
+* Backbone name without \* means the resolution is 224x224, and with \* means the resolution is 48x192 (h\*w). The stride of the network is changed to `[2, [2, 1], [2, 1], [2, 1]`. Please refer to [PaddleOCR]（ https://github.com/PaddlePaddle/PaddleOCR）for more details.
+* Backbone name with \*\* means that the resolution is 80x160 (h\*w), and the stride of the network is changed to `[2, [2, 1], [2, 1], [2, 1]]`. This resolution is searched by [Hyperparameter Searching](pulc_train_en.md#4).
+* The Latency is tested on Intel(R) Xeon(R) Gold 6148 CPU @ 2.40GHz. The MKLDNN is enabled and the number of threads is 10.
+* About PP-LCNet, please refer to [PP-LCNet Introduction](../models/PP-LCNet_en.md) and [PP-LCNet Paper](https://arxiv.org/abs/2109.15099).
+
+<a name="2"></a>
+
+## 2. Quick Start
+
+<a name="2.1"></a>  
+
+### 2.1 PaddlePaddle Installation
+
+- Run the following command to install if CUDA9 or CUDA10 is available.
+
+```bash
+python3 -m pip install paddlepaddle-gpu -i https://mirror.baidu.com/pypi/simple
+```
+
+- Run the following command to install if GPU device is unavailable.
+
+```bash
+python3 -m pip install paddlepaddle -i https://mirror.baidu.com/pypi/simple
+```
+
+Please refer to [PaddlePaddle Installation](https://www.paddlepaddle.org.cn/install/quick?docurl=/documentation/docs/en/install/pip/linux-pip_en.html) for more information about installation, for examples other versions.
+
+<a name="2.2"></a>  
+
+### 2.2 PaddleClas wheel Installation
+
+The command of PaddleClas installation as bellow:
+
+```bash
+pip3 install paddleclas
+```
+
+<a name="2.3"></a>
+
+### 2.3 Prediction
+
+First, please click [here](https://paddleclas.bj.bcebos.com/data/PULC/pulc_demo_imgs.zip) to download and unzip to get the test demo images.
+
+* Prediction with CLI
+
+```bash
+paddleclas --model_name=textline_orientation --infer_imgs=pulc_demo_imgs/textline_orientation/textline_orientation_test_0_0.png
+```
+
+Results:
+
+```
+>>> result
+class_ids: [0], scores: [1.0], label_names: ['0_degree'], filename: pulc_demo_imgs/textline_orientation/textline_orientation_test_0_0.png
+Predict complete!
+```
+
+**Note**: If you want to test other images, only need to specify the `--infer_imgs` argument, and the directory containing images is also supported.
+
+* Prediction in Python
+
+```python
+import paddleclas
+model = paddleclas.PaddleClas(model_name="textline_orientation")
+result = model.predict(input_data="pulc_demo_imgs/textline_orientation/textline_orientation_test_0_0.png")
+print(next(result))
+```
+
+**Note**: The `result` returned by `model.predict()` is a generator, so you need to use the `next()` function to call it or `for` loop to loop it. And it will predict with `batch_size` size batch and return the prediction results when called. The default `batch_size` is 1, and you also specify the `batch_size` when instantiating, such as `model = paddleclas.PaddleClas(model_name="textline_orientation",  batch_size=2)`. The result of demo above:
+
+```
+>>> result
+[{'class_ids': [0], 'scores': [1.0], 'label_names': ['0_degree'], 'filename': 'pulc_demo_imgs/textline_orientation/textline_orientation_test_0_0.png'}]
+```
+
+<a name="3"></a>
+
+## 3. Training, Evaluation and Inference
+
+<a name="3.1"></a>  
+
+### 3.1 Installation
+
+Please refer to [Installation](../installation/install_paddleclas_en.md) to get the description about installation.
+
+<a name="3.2"></a>
+
+### 3.2 Dataset
+
+<a name="3.2.1"></a>
+
+#### 3.2.1 Dataset Introduction
+
+The data used in this case come from internal data. If you want to experience the training process, you can use open source data, such as [ICDAR2019-LSVT](https://aistudio.baidu.com/aistudio/datasetdetail/8429).
+
+<a name="3.2.2"></a>  
+
+#### 3.2.2 Getting Dataset
+
+Take ICDAR2019-LSVT for example, images with ID numbers from 0 to 1999 would be processed and used. After rotation, it is divided into class 0 or class 1. Class 0 means that the textline rotation angle is 0 degrees, and class 1 means 180 degrees.
+
+- Training data: The images with ID number from 0 to 1799 are used as the training set. 3600 images in total.
+- Evaluation data: The images with ID number from 1800 to 1999 are used as the evaluation set. 400 images in total.
+
+Some image of the processed dataset is as follows:
+
+![](../../images/PULC/docs/textline_orientation_data_demo.png)
+
+And you can also download the data processed directly.
+
+```
+cd path_to_PaddleClas
+```
+
+Enter the `dataset/` directory, download and unzip the dataset.
+
+```shell
+cd dataset
+wget https://paddleclas.bj.bcebos.com/data/PULC/textline_orientation.tar
+tar -xf textline_orientation.tar
+cd ../
+```
+
+The datas under `textline_orientation` directory:
+
+```
+├── 0
+│   ├── img_0.jpg
+│   ├── img_1.jpg
+...
+├── 1
+│   ├── img_0.jpg
+│   ├── img_1.jpg
+...
+├── train_list.txt
+└── val_list.txt
+```
+
+其中 `0/` 和 `1/` 分别存放 0 类和 1 类的数据。`train_list.txt` 和 `val_list.txt` 分别为训练集和验证集的标签文件。
+
+Where `0/` and `1/` are class 0 and class 1 data respectively. The `train_list.txt` and `val_list.txt` are label files of training data and validation data respectively.
+
+**Note**:
+
+* About the contents format of `train_list.txt` and `val_list.txt`, please refer to [Description about Classification Dataset in PaddleClas](../data_preparation/classification_dataset_en.md).
+
+<a name="3.3"></a>
+
+### 3.3 Training
+
+The details of training config in `ppcls/configs/PULC/textline_orientation/PPLCNet_x1_0.yaml`. The command about training as follows:
+
+```shell
+export CUDA_VISIBLE_DEVICES=0,1,2,3
+python3 -m paddle.distributed.launch \
+    --gpus="0,1,2,3" \
+    tools/train.py \
+        -c ./ppcls/configs/PULC/textline_orientation/PPLCNet_x1_0.yaml
+```
+
+**Note**:
+
+* Because the ICDAR2019-LSVT data set is different from the dataset used in the provided pretrained model. If you want to get higher accuracy, you can process [ICDAR2019-LSVT](https://aistudio.baidu.com/aistudio/datasetdetail/8429).
+
+<a name="3.4"></a>
+
+### 3.4 Evaluation
+
+After training, you can use the following commands to evaluate the model.
+
+```bash
+python3 tools/eval.py \
+    -c ./ppcls/configs/PULC/textline_orientation/PPLCNet_x1_0.yaml \
+    -o Global.pretrained_model="output/PPLCNet_x1_0/best_model"
+```
+
+Among the above command, the argument `-o Global.pretrained_model="output/PPLCNet_x1_0/best_model"` specify the path of the best model weight file. You can specify other path if needed.
+
+<a name="3.5"></a>
+
+### 3.5 Inference
+
+After training, you can use the model that trained to infer. Command is as follow:
+
+```python
+python3 tools/infer.py \
+    -c ./ppcls/configs/PULC/textline_orientation/PPLCNet_x1_0.yaml \
+    -o Global.pretrained_model=output/PPLCNet_x1_0/best_model
+```
+
+The results:
+
+```
+[{'class_ids': [0], 'scores': [1.0], 'file_name': 'deploy/images/PULC/textline_orientation/textline_orientation_test_0_0.png', 'label_names': ['0_degree']}]
+```
+
+**Note**:
+
+* Among the above command, argument `-o Global.pretrained_model="output/PPLCNet_x1_0/best_model"` specify the path of the best model weight file. You can specify other path if needed.
+* The default test image is `deploy/images/PULC/textline_orientation/textline_orientation_test_0_0.png`. And you can test other image, only need to specify the argument `-o Infer.infer_imgs=path_to_test_image`.
+
+<a name="4"></a>
+
+## 4. Model Compression
+
+<a name="4.1"></a>
+
+### 4.1 SKL-UGI Knowledge Distillation
+
+SKL-UGI is a simple but effective knowledge distillation algrithem proposed by PaddleClas.
+
+<!-- todo -->
+<!-- Please refer to [SKL-UGI](../advanced_tutorials/distillation/distillation_en.md) for more details. -->
+
+<a name="4.1.1"></a>
+
+#### 4.1.1 Teacher Model Training
+
+Training the teacher model with hyperparameters specified in `ppcls/configs/PULC/textline_orientation/PPLCNet/PPLCNet_x1_0.yaml`. The command is as follow:
+
+```shell
+export CUDA_VISIBLE_DEVICES=0,1,2,3
+python3 -m paddle.distributed.launch \
+    --gpus="0,1,2,3" \
+    tools/train.py \
+        -c ./ppcls/configs/PULC/textline_orientation/PPLCNet_x1_0.yaml \
+        -o Arch.name=ResNet101_vd
+```
+
+The best metric of validation data is between `0.96` and `0.98`. The best teacher model weight would be saved in file `output/ResNet101_vd/best_model.pdparams`.
+
+<a name="4.1.2"></a>
+
+#### 4.1.2 Knowledge Distillation Training
+
+The training strategy, specified in training config file `ppcls/configs/PULC/textline_orientation/PPLCNet_x1_0_distillation.yaml`, the teacher model is `ResNet101_vd` and the student model is `PPLCNet_x1_0`. The command is as follow:
+
+```shell
+export CUDA_VISIBLE_DEVICES=0,1,2,3
+python3 -m paddle.distributed.launch \
+    --gpus="0,1,2,3" \
+    tools/train.py \
+        -c ./ppcls/configs/PULC/textline_orientation/PPLCNet_x1_0_distillation.yaml \
+        -o Arch.models.0.Teacher.pretrained=output/ResNet101_vd/best_model
+```
+
+The best metric is between `0.95` and `0.97`. The best student model weight would be saved in file `output/DistillationModel/best_model_student.pdparams`.
+
+<a name="5"></a>
+
+## 5. Hyperparameters Searching
+
+The hyperparameters used by [3.2 section](#3.2) and [4.1 section](#4.1) are according by `Hyperparameters Searching` in PaddleClas. If you want to get better results on your own dataset, you can refer to [Hyperparameters Searching](PULC_train_en.md#4) to get better hyperparameters.
+
+**Note**: This section is optional. Because the search process will take a long time, you can selectively run according to your specific. If not replace the dataset, you can ignore this section.
+
+<a name="6"></a>
+
+## 6. Inference Deployment
+
+<a name="6.1"></a>
+
+### 6.1 Getting Paddle Inference Model
+
+Paddle Inference is the original Inference Library of the PaddlePaddle, provides high-performance inference for server deployment. And compared with  directly based on the pretrained model, Paddle Inference can use tools to accelerate prediction, so as to achieve better inference performance. Please refer to [Paddle Inference](https://www.paddlepaddle.org.cn/documentation/docs/zh/guides/infer/inference/inference_cn.html) for more information.
+
+Paddle Inference need Paddle Inference Model to predict. Two process provided to get Paddle Inference Model. If want to use the provided by PaddleClas, you can download directly, click [Downloading Inference Model](#6.1.2).
+
+<a name="6.1.1"></a>
+
+### 6.1.1 Exporting Paddle Inference Model
+
+The command about exporting Paddle Inference Model is as follow:
+
+```bash
+python3 tools/export_model.py \
+    -c ./ppcls/configs/PULC/textline_orientation/PPLCNet_x1_0.yaml \
+    -o Global.pretrained_model=output/PPLCNet_x1_0/best_model \
+    -o Global.save_inference_dir=deploy/models/PPLCNet_x1_0_textline_orientation_infer
+```
+
+After running above command, the inference model files would be saved in `deploy/models/PPLCNet_x1_0_textline_orientation_infer`, as shown below:
+
+```
+├── PPLCNet_x1_0_textline_orientation_infer
+│   ├── inference.pdiparams
+│   ├── inference.pdiparams.info
+│   └── inference.pdmodel
+```
+
+**Note**: The best model is from knowledge distillation training. If knowledge distillation training is not used, the best model would be saved in `output/PPLCNet_x1_0/best_model.pdparams`.
+
+<a name="6.1.2"></a>
+
+### 6.1.2 Downloading Inference Model
+
+You can also download directly.
+
+```
+cd deploy/models
+# 下载 inference 模型并解压
+wget https://paddleclas.bj.bcebos.com/models/PULC/textline_orientation_infer.tar && tar -xf textline_orientation_infer.tar
+```
+
+After decompression, the directory `models` should be shown below.
+
+```
+├── textline_orientation_infer
+│   ├── inference.pdiparams
+│   ├── inference.pdiparams.info
+│   └── inference.pdmodel
+```
+
+<a name="6.2"></a>
+
+### 6.2 Prediction with Python
+
+<a name="6.2.1"></a>  
+
+#### 6.2.1 Image Prediction
+
+Return the directory `deploy`:
+
+```
+cd ../
+```
+
+Run the following command to classify the rotation of image `./images/PULC/textline_orientation/objects365_02035329.jpg`.
+
+```shell
+# Use the following command to predict with GPU.
+python3.7 python/predict_cls.py -c configs/PULC/textline_orientation/inference_textline_orientation.yaml
+# Use the following command to predict with CPU.
+python3.7 python/predict_cls.py -c configs/PULC/textline_orientation/inference_textline_orientation.yaml  -o Global.use_gpu=False
+```
+
+The prediction results:
+
+```
+textline_orientation_test_0_0.png:    class id(s): [0], score(s): [1.00], label_name(s): ['0_degree']
+```
+
+<a name="6.2.2"></a>  
+
+#### 6.2.2 Images Prediction
+
+If you want to predict images in directory, please specify the argument `Global.infer_imgs` as directory path by `-o Global.infer_imgs`. The command is as follow.
+
+```shell
+# Use the following command to predict with GPU. If want to replace with CPU, you can add argument -o Global.use_gpu=False
+python3.7 python/predict_cls.py -c configs/PULC/textline_orientation/inference_textline_orientation.yaml -o Global.infer_imgs="./images/PULC/textline_orientation/"
+```
+
+All prediction results will be printed, as shown below.
+
+```
+textline_orientation_test_0_0.png:    class id(s): [0], score(s): [1.00], label_name(s): ['0_degree']
+textline_orientation_test_0_1.png:    class id(s): [0], score(s): [1.00], label_name(s): ['0_degree']
+textline_orientation_test_1_0.png:    class id(s): [1], score(s): [1.00], label_name(s): ['180_degree']
+textline_orientation_test_1_1.png:    class id(s): [1], score(s): [1.00], label_name(s): ['180_degree']
+```
+
+Among the prediction results above, `0_degree` means that the rotation angle of the textline image is 0, and `180_degree` means that 180.
+
+<a name="6.3"></a>
+
+### 6.3 Deployment with C++
+
+PaddleClas provides an example about how to deploy with C++. Please refer to [Deployment with C++](../inference_deployment/cpp_deploy_en.md).
+
+<a name="6.4"></a>
+
+### 6.4 Deployment as Service
+
+Paddle Serving is a flexible, high-performance carrier for machine learning models, and supports different protocol, such as RESTful, gRPC, bRPC and so on, which provides different deployment solutions for a variety of heterogeneous hardware and operating system environments. Please refer [Paddle Serving](https://github.com/PaddlePaddle/Serving) for more information.
+
+PaddleClas provides an example about how to deploy as service by Paddle Serving. Please refer to [Paddle Serving Deployment](../inference_deployment/paddle_serving_deploy_en.md).
+
+<a name="6.5"></a>
+
+### 6.5 Deployment on Mobile
+
+Paddle-Lite is an open source deep learning framework that designed to make easy to perform inference on mobile, embeded, and IoT devices. Please refer to [Paddle-Lite](https://github.com/PaddlePaddle/Paddle-Lite) for more information.
+
+PaddleClas provides an example of how to deploy on mobile by Paddle-Lite. Please refer to [Paddle-Lite deployment](../inference_deployment/paddle_lite_deploy_en.md).
+
+<a name="6.6"></a>
+
+### 6.6 Converting To ONNX and Deployment
+
+Paddle2ONNX support convert Paddle Inference model to ONNX model. And you can deploy with ONNX model on different inference engine, such as TensorRT, OpenVINO, MNN/TNN, NCNN and so on. About Paddle2ONNX details, please refer to [Paddle2ONNX](https://github.com/PaddlePaddle/Paddle2ONNX).
+
+PaddleClas provides an example of how to convert Paddle Inference model to ONNX model by paddle2onnx toolkit and predict by ONNX model. You can refer to [paddle2onnx](../../../deploy/paddle2onnx/readme_en.md) for deployment details.

docs/en/PULC/PULC_traffic_sign_en.md

+# PULC Classification Model of Traffic Sign
+
+------
+
+## Catalogue
+
+- [1. Introduction](#1)
+- [2. Quick Start](#2)
+    - [2.1 PaddlePaddle Installation](#2.1)
+    - [2.2 PaddleClas Installation](#2.2)
+    - [2.3 Prediction](#2.3)
+- [3. Training, Evaluation and Inference](#3)
+    - [3.1 Installation](#3.1)
+    - [3.2 Dataset](#3.2)
+      - [3.2.1 Dataset Introduction](#3.2.1)
+      - [3.2.2 Getting Dataset](#3.2.2)
+    - [3.3 Training](#3.3)
+    - [3.4 Evaluation](#3.4)
+    - [3.5 Inference](#3.5)
+- [4. Model Compression](#4)
+  - [4.1 SKL-UGI Knowledge Distillation](#4.1)
+    - [4.1.1 Teacher Model Training](#4.1.1)
+    - [4.1.2 Knowledge Distillation Training](#4.1.2)
+- [5. SHAS](#5)
+- [6. Inference Deployment](#6)
+  - [6.1 Getting Paddle Inference Model](#6.1)
+    - [6.1.1 Exporting Paddle Inference Model](#6.1.1)
+    - [6.1.2 Downloading Inference Model](#6.1.2)
+  - [6.2 Prediction with Python](#6.2)
+    - [6.2.1 Image Prediction](#6.2.1)
+    - [6.2.2 Images Prediction](#6.2.2)
+  - [6.3 Deployment with C++](#6.3)
+  - [6.4 Deployment as Service](#6.4)
+  - [6.5 Deployment on Mobile](#6.5)
+  - [6.6 Converting To ONNX and Deployment](#6.6)
+
+<a name="1"></a>
+
+## 1. Introduction
+
+This case provides a way for users to quickly build a lightweight, high-precision and practical classification model of traffic sign using PaddleClas PULC (Practical Ultra Lightweight image Classification). The model can be widely used in automatic driving, road monitoring, etc.
+
+The following table lists the relevant indicators of the model. The first two lines means that using SwinTransformer_tiny and MobileNetV3_small_x0_35 as the backbone to training. The third to sixth lines means that the backbone is replaced by PPLCNet, additional use of EDA strategy and additional use of EDA strategy and SKL-UGI knowledge distillation strategy.
+
+| Backbone | Top-1 Acc(%) | Latency(ms) | Size(M)| Training Strategy |
+|-------|-----------|----------|---------------|---------------|
+| SwinTranformer_tiny  | 98.11 | 89.45  | 111 | using ImageNet pretrained model |
+| MobileNetV3_small_x0_35  | 93.88 | 3.01  | 3.9 | using ImageNet pretrained model |
+| PPLCNet_x1_0  | 97.78 | 2.10  | 8.2 | using ImageNet pretrained model |
+| PPLCNet_x1_0  | 97.84 | 2.10  | 8.2 | using SSLD pretrained model |
+| PPLCNet_x1_0  | 98.14 | 2.10  | 8.2 | using SSLD pretrained model + EDA strategy  |
+| <b>PPLCNet_x1_0<b>  | <b>98.35<b> | <b>2.10<b>  | <b>8.2<b> | using SSLD pretrained model + EDA strategy + SKL-UGI knowledge distillation strategy|
+
+It can be seen that high accuracy can be getted when backbone is SwinTranformer_tiny, but the speed is slow. Replacing backbone with the lightweight model MobileNetV3_small_x0_35, the speed can be greatly improved, but the accuracy will be greatly reduced. Replacing backbone with faster backbone PPLCNet_x1_0, the accuracy is lower 3.9 percentage points than MobileNetv3_small_x0_35. At the same time, the speed can be more than 43% faster. After additional using the SSLD pretrained model, the accuracy can be improved by about 0.06 percentage points without affecting the inference speed. Further, additional using the EDA strategy, the accuracy can be increased by 0.3 percentage points. Finally, after additional using the SKL-UGI knowledge distillation, the accuracy can be further improved by 0.21 percentage points. At this point, the accuracy exceeds that of SwinTranformer_tiny, but the speed is more than 41 times faster. The training method and deployment instructions of PULC will be introduced in detail below.
+
+**Note**:
+
+* The Latency is tested on Intel(R) Xeon(R) Gold 6148 CPU @ 2.40GHz. The MKLDNN is enabled and the number of threads is 10.
+* About PP-LCNet, please refer to [PP-LCNet Introduction](../models/PP-LCNet_en.md) and [PP-LCNet Paper](https://arxiv.org/abs/2109.15099).
+
+<a name="2"></a>
+
+## 2. Quick Start
+
+<a name="2.1"></a>  
+
+### 2.1 PaddlePaddle Installation
+
+- Run the following command to install if CUDA9 or CUDA10 is available.
+
+```bash
+python3 -m pip install paddlepaddle-gpu -i https://mirror.baidu.com/pypi/simple
+```
+
+- Run the following command to install if GPU device is unavailable.
+
+```bash
+python3 -m pip install paddlepaddle -i https://mirror.baidu.com/pypi/simple
+```
+
+Please refer to [PaddlePaddle Installation](https://www.paddlepaddle.org.cn/install/quick?docurl=/documentation/docs/en/install/pip/linux-pip_en.html) for more information about installation, for examples other versions.
+
+<a name="2.2"></a>  
+
+### 2.2 PaddleClas wheel Installation
+
+The command of PaddleClas installation as bellow:
+
+```bash
+pip3 install paddleclas
+```
+
+<a name="2.3"></a>
+
+### 2.3 Prediction
+
+First, please click [here](https://paddleclas.bj.bcebos.com/data/PULC/pulc_demo_imgs.zip) to download and unzip to get the test demo images.
+
+* Prediction with CLI
+
+```bash
+paddleclas --model_name=traffic_sign --infer_imgs=pulc_demo_imgs/traffic_sign/100999_83928.jpg
+```
+
+Results:
+
+```
+>>> result
+class_ids: [182, 179, 162, 128, 24], scores: [0.98623, 0.01255, 0.00022, 0.00021, 0.00012], label_names: ['pl110', 'pl100', 'pl120', 'p26', 'pm10'], filename: pulc_demo_imgs/traffic_sign/100999_83928.jpg
+Predict complete!
+```
+
+**Note**: If you want to test other images, only need to specify the `--infer_imgs` argument, and the directory containing images is also supported.
+
+* Prediction in Python
+
+```python
+import paddleclas
+model = paddleclas.PaddleClas(model_name="traffic_sign")
+result = model.predict(input_data="pulc_demo_imgs/traffic_sign/100999_83928.jpg")
+print(next(result))
+```
+
+**Note**: The `result` returned by `model.predict()` is a generator, so you need to use the `next()` function to call it or `for` loop to loop it. And it will predict with `batch_size` size batch and return the prediction results when called. The default `batch_size` is 1, and you also specify the `batch_size` when instantiating, such as `model = paddleclas.PaddleClas(model_name="traffic_sign",  batch_size=2)`. The result of demo above:
+
+```
+>>> result
+[{'class_ids': [182, 179, 162, 128, 24], 'scores': [0.98623, 0.01255, 0.00022, 0.00021, 0.00012], 'label_names': ['pl110', 'pl100', 'pl120', 'p26', 'pm10'], 'filename': 'pulc_demo_imgs/traffic_sign/100999_83928.jpg'}]
+```
+
+<a name="3"></a>
+
+## 3. Training, Evaluation and Inference
+
+<a name="3.1"></a>  
+
+### 3.1 Installation
+
+Please refer to [Installation](../installation/install_paddleclas_en.md) to get the description about installation.
+
+<a name="3.2"></a>
+
+### 3.2 Dataset
+
+<a name="3.2.1"></a>
+
+#### 3.2.1 Dataset Introduction
+
+All datasets used in this case are open source data. Train data is the subset of [MS-COCO](https://cocodataset.org/#overview) training data. And the validation data is the subset of [Object365](https://www.objects365.org/overview.html) training data. ImageNet_val is [ImageNet-1k](https://www.image-net.org/) validation data.
+
+The dataset used in this case is based on the [Tsinghua-Tencent 100K dataset (CC-BY-NC license), TT100K](https://cg.cs.tsinghua.edu.cn/traffic-sign/) randomly expanded and cropped according to the bounding box.
+
+<a name="3.2.2"></a>  
+
+#### 3.2.2 Getting Dataset
+
+The processing to `TT00K` includes randomly expansion and cropping, details are shown below.
+
+```python
+def get_random_crop_box(xmin, ymin, xmax, ymax, img_height, img_width, ratio=1.0):
+    h = ymax - ymin
+    w = ymax - ymin
+
+    xmin_diff = random.random() * ratio * min(w, xmin/ratio)
+    ymin_diff = random.random() * ratio * min(h, ymin/ratio)
+    xmax_diff = random.random() * ratio * min(w, (img_width-xmin-1)/ratio)
+    ymax_diff = random.random() * ratio * min(h, (img_height-ymin-1)/ratio)
+
+    new_xmin = round(xmin - xmin_diff)
+    new_ymin = round(ymin - ymin_diff)
+    new_xmax = round(xmax + xmax_diff)
+    new_ymax = round(ymax + ymax_diff)
+
+    return new_xmin, new_ymin, new_xmax, new_ymax
+```
+
+Some image of the processed dataset is as follows:
+
+<div align="center">
+<img src="../../images/PULC/docs/traffic_sign_data_demo.png"  width = "500" />
+</div>
+
+You can also download the data processed directly. And the process script file `deal.py` is also included.
+
+```
+cd path_to_PaddleClas
+```
+
+Enter the `dataset/` directory, download and unzip the dataset.
+
+```shell
+cd dataset
+wget https://paddleclas.bj.bcebos.com/data/PULC/traffic_sign.tar
+tar -xf traffic_sign.tar
+cd ../
+```
+
+The datas under `traffic_sign` directory:
+
+```
+traffic_sign
+├── train
+│   ├── 0_62627.jpg
+│   ├── 100000_89031.jpg
+│   ├── 100001_89031.jpg
+...
+├── test
+│   ├── 100423_2315.jpg
+│   ├── 100424_2315.jpg
+│   ├── 100425_2315.jpg
+...
+├── other
+│   ├── 100603_3422.jpg
+│   ├── 100604_3422.jpg
+...
+├── label_list_train.txt
+├── label_list_test.txt
+├── label_list_other.txt
+├── label_list_train_for_distillation.txt
+├── label_list_train.txt.debug
+├── label_list_test.txt.debug
+├── label_name_id.txt
+├── deal.py
+```
+
+Where `train/` and `test/` are training set and validation set respectively. The `label_list_train.txt` and `label_list_test.txt` are label files of training data and validation data respectively. The file `label_list_train.txt.debug` and `label_list_test.txt.debug` are subset of `train_list.txt` and `val_list.txt` respectively. `other` would be used for SKL-UGI knowledge distillation, and its label file is `label_list_train_for_distillation.txt`.
+
+**Note**:
+
+* About the contents format of `label_list_train.txt` and `label_list_train.txt`, please refer to [Description about Classification Dataset in PaddleClas](../data_preparation/classification_dataset_en.md).
+* About the `label_list_train_for_distillation.txt`, please refer to [Knowledge Distillation Label](../advanced_tutorials/distillation/distillation_en.md).
+
+<a name="3.3"></a>
+
+### 3.3 Training
+
+The details of training config in `ppcls/configs/PULC/traffic_sign/PPLCNet_x1_0.yaml`. The command about training as follows:
+
+```shell
+export CUDA_VISIBLE_DEVICES=0,1,2,3
+python3 -m paddle.distributed.launch \
+    --gpus="0,1,2,3" \
+    tools/train.py \
+        -c ./ppcls/configs/PULC/traffic_sign/PPLCNet_x1_0.yaml
+```
+
+The best metric of validation data is between `98.0` and `98.2`. There would be fluctuations because the data size is small.
+
+<a name="3.4"></a>
+
+### 3.4 Evaluation
+
+After training, you can use the following commands to evaluate the model.
+
+```bash
+python3 tools/eval.py \
+    -c ./ppcls/configs/PULC/traffic_sign/PPLCNet_x1_0.yaml \
+    -o Global.pretrained_model="output/PPLCNet_x1_0/best_model"
+```
+
+Among the above command, the argument `-o Global.pretrained_model="output/PPLCNet_x1_0/best_model"` specify the path of the best model weight file. You can specify other path if needed.
+
+<a name="3.5"></a>
+
+### 3.5 Inference
+
+After training, you can use the model that trained to infer. Command is as follow:
+
+```bash
+python3 tools/infer.py \
+    -c ./ppcls/configs/PULC/traffic_sign/PPLCNet_x1_0.yaml \
+    -o Global.pretrained_model=output/DistillationModel/best_model
+```
+
+The results:
+
+```
+99603_17806.jpg:        class id(s): [216, 145, 49, 207, 169], score(s): [1.00, 0.00, 0.00, 0.00, 0.00], label_name(s): ['pm20', 'pm30', 'pm40', 'pl25', 'pm15']
+```
+
+**Note**:
+
+* Among the above command, argument `-o Global.pretrained_model="output/PPLCNet_x1_0/best_model"` specify the path of the best model weight file. You can specify other path if needed.
+* The default test image is `deploy/images/PULC/traffic_sign/99603_17806.jpg`. And you can test other image, only need to specify the argument `-o Infer.infer_imgs=path_to_test_image`.
+
+
+<a name="4"></a>
+
+## 4. Model Compression
+
+<a name="4.1"></a>
+
+### 4.1 SKL-UGI Knowledge Distillation
+
+SKL-UGI is a simple but effective knowledge distillation algrithem proposed by PaddleClas.
+
+<!-- todo -->
+<!-- Please refer to [SKL-UGI](../advanced_tutorials/distillation/distillation_en.md) for more details. -->
+
+<a name="4.1.1"></a>
+
+#### 4.1.1 Teacher Model Training
+
+Training the teacher model with hyperparameters specified in `ppcls/configs/PULC/traffic_sign/PPLCNet_x1_0.yaml`. The command is as follow:
+
+```shell
+export CUDA_VISIBLE_DEVICES=0,1,2,3
+python3 -m paddle.distributed.launch \
+    --gpus="0,1,2,3" \
+    tools/train.py \
+        -c ./ppcls/configs/PULC/traffic_sign/PPLCNet_x1_0.yaml \
+        -o Arch.name=ResNet101_vd
+```
+
+The best metric of validation data is about `98.59%`. The best teacher model weight would be saved in file `output/ResNet101_vd/best_model.pdparams`.
+
+<a name="4.1.2"></a>
+
+#### 4.1.2 Knowledge Distillation Training
+
+The training strategy, specified in training config file `ppcls/configs/PULC/traffic_sign/PPLCNet_x1_0_distillation.yaml`, the teacher model is `ResNet101_vd`, the student model is `PPLCNet_x1_0` and the additional unlabeled training data is validation data of ImageNet1k. The command is as follow:
+
+```shell
+export CUDA_VISIBLE_DEVICES=0,1,2,3
+python3 -m paddle.distributed.launch \
+    --gpus="0,1,2,3" \
+    tools/train.py \
+        -c ./ppcls/configs/PULC/traffic_sign/PPLCNet_x1_0_distillation.yaml \
+        -o Arch.models.0.Teacher.pretrained=output/ResNet101_vd/best_model
+```
+
+The best metric is about `98.35%`. The best student model weight would be saved in file `output/DistillationModel/best_model_student.pdparams`.
+
+<a name="5"></a>
+
+## 5. Hyperparameters Searching
+
+The hyperparameters used by [3.2 section](#3.2) and [4.1 section](#4.1) are according by `Hyperparameters Searching` in PaddleClas. If you want to get better results on your own dataset, you can refer to [Hyperparameters Searching](PULC_train_en.md#4) to get better hyperparameters.
+
+**Note**: This section is optional. Because the search process will take a long time, you can selectively run according to your specific. If not replace the dataset, you can ignore this section.
+
+<a name="6"></a>
+
+## 6. Inference Deployment
+
+<a name="6.1"></a>
+
+### 6.1 Getting Paddle Inference Model
+
+Paddle Inference is the original Inference Library of the PaddlePaddle, provides high-performance inference for server deployment. And compared with  directly based on the pretrained model, Paddle Inference can use tools to accelerate prediction, so as to achieve better inference performance. Please refer to [Paddle Inference](https://www.paddlepaddle.org.cn/documentation/docs/zh/guides/infer/inference/inference_cn.html) for more information.
+
+Paddle Inference need Paddle Inference Model to predict. Two process provided to get Paddle Inference Model. If want to use the provided by PaddleClas, you can download directly, click [Downloading Inference Model](#6.1.2).
+<a name="6.1.1"></a>
+
+### 6.1.1 Exporting Paddle Inference Model
+
+The command about exporting Paddle Inference Model is as follow:
+
+```bash
+python3 tools/export_model.py \
+    -c ./ppcls/configs/PULC/traffic_sign/PPLCNet_x1_0.yaml \
+    -o Global.pretrained_model=output/DistillationModel/best_model_student \
+    -o Global.save_inference_dir=deploy/models/PPLCNet_x1_0_traffic_sign_infer
+```
+
+After running above command, the inference model files would be saved in `deploy/models/PPLCNet_x1_0_traffic_sign_infer`, as shown below:
+
+```
+├── PPLCNet_x1_0_traffic_sign_infer
+│   ├── inference.pdiparams
+│   ├── inference.pdiparams.info
+│   └── inference.pdmodel
+```
+
+**Note**: The best model is from knowledge distillation training. If knowledge distillation training is not used, the best model would be saved in `output/PPLCNet_x1_0/best_model.pdparams`.
+
+<a name="6.1.2"></a>
+
+### 6.1.2 Downloading Inference Model
+
+You can also download directly.
+
+```
+cd deploy/models
+# download the inference model and decompression
+wget https://paddleclas.bj.bcebos.com/models/PULC/traffic_sign_infer.tar && tar -xf traffic_sign_infer.tar
+```
+
+After decompression, the directory `models` should be shown below.
+
+```
+├── traffic_sign_infer
+│   ├── inference.pdiparams
+│   ├── inference.pdiparams.info
+│   └── inference.pdmodel
+```
+
+<a name="6.2"></a>
+
+### 6.2 Prediction with Python
+
+<a name="6.2.1"></a>  
+
+#### 6.2.1 Image Prediction
+
+Return the directory `deploy`:
+
+```
+cd ../
+```
+
+Run the following command to classify traffic sign about the image `./images/PULC/traffic_sign/99603_17806.jpg`.
+
+```shell
+# Use the following command to predict with GPU.
+python3.7 python/predict_cls.py -c configs/PULC/traffic_sign/inference_traffic_sign.yaml
+# Use the following command to predict with CPU.
+python3.7 python/predict_cls.py -c configs/PULC/traffic_sign/inference_traffic_sign.yaml -o Global.use_gpu=False
+```
+
+The prediction results:
+
+```
+99603_17806.jpg:        class id(s): [216, 145, 49, 207, 169], score(s): [1.00, 0.00, 0.00, 0.00, 0.00], label_name(s): ['pm20', 'pm30', 'pm40', 'pl25', 'pm15']
+```
+
+<a name="6.2.2"></a>  
+
+#### 6.2.2 Images Prediction
+
+If you want to predict images in directory, please specify the argument `Global.infer_imgs` as directory path by `-o Global.infer_imgs`. The command is as follow.
+
+```shell
+# Use the following command to predict with GPU. If want to replace with CPU, you can add argument -o Global.use_gpu=False
+python3.7 python/predict_cls.py -c configs/PULC/traffic_sign/inference_traffic_sign.yaml -o Global.infer_imgs="./images/PULC/traffic_sign/"
+```
+
+All prediction results will be printed, as shown below.
+
+```
+100999_83928.jpg:    class id(s): [182, 179, 162, 128, 24], score(s): [0.99, 0.01, 0.00, 0.00, 0.00], label_name(s): ['pl110', 'pl100', 'pl120', 'p26', 'pm10']
+99603_17806.jpg:    class id(s): [216, 145, 49, 24, 169], score(s): [1.00, 0.00, 0.00, 0.00, 0.00], label_name(s): ['pm20', 'pm30', 'pm40', 'pm10', 'pm15']
+```
+
+About the `label_name` details, please refer to `dataset/traffic_sign/report.pdf`.
+
+<a name="6.3"></a>
+
+### 6.3 Deployment with C++
+
+PaddleClas provides an example about how to deploy with C++. Please refer to [Deployment with C++](../inference_deployment/cpp_deploy_en.md).
+
+<a name="6.4"></a>
+
+### 6.4 Deployment as Service
+
+Paddle Serving is a flexible, high-performance carrier for machine learning models, and supports different protocol, such as RESTful, gRPC, bRPC and so on, which provides different deployment solutions for a variety of heterogeneous hardware and operating system environments. Please refer [Paddle Serving](https://github.com/PaddlePaddle/Serving) for more information.
+
+PaddleClas provides an example about how to deploy as service by Paddle Serving. Please refer to [Paddle Serving Deployment](../inference_deployment/paddle_serving_deploy_en.md).
+
+<a name="6.5"></a>
+
+### 6.5 Deployment on Mobile
+
+Paddle-Lite is an open source deep learning framework that designed to make easy to perform inference on mobile, embeded, and IoT devices. Please refer to [Paddle-Lite](https://github.com/PaddlePaddle/Paddle-Lite) for more information.
+
+PaddleClas provides an example of how to deploy on mobile by Paddle-Lite. Please refer to [Paddle-Lite deployment](../inference_deployment/paddle_lite_deploy_en.md).
+
+<a name="6.6"></a>
+
+### 6.6 Converting To ONNX and Deployment
+
+Paddle2ONNX support convert Paddle Inference model to ONNX model. And you can deploy with ONNX model on different inference engine, such as TensorRT, OpenVINO, MNN/TNN, NCNN and so on. About Paddle2ONNX details, please refer to [Paddle2ONNX](https://github.com/PaddlePaddle/Paddle2ONNX).
+
+PaddleClas provides an example of how to convert Paddle Inference model to ONNX model by paddle2onnx toolkit and predict by ONNX model. You can refer to [paddle2onnx](../../../deploy/paddle2onnx/readme_en.md) for deployment details.

docs/en/PULC/PULC_train_en.md

@@ -14,7 +14,7 @@
     - [3.3 EDA strategy](#3.3)
     - [3.4 SKL-UGI knowledge distillation](#3.4)
     - [3.5 Summary](#3.5)
-- [4. Hyperparameter Search](#4)
+- [4. Hyperparameters Searching](#4)
     - [4.1 Search based on default configuration](#4.1)
     - [4.2 Custom search configuration](#4.2)
 
@@ -31,7 +31,7 @@ The PULC solution has been verified to be effective in many scenarios, such as h
 <img src="https://user-images.githubusercontent.com/19523330/173011854-b10fcd7a-b799-4dfd-a1cf-9504952a3c44.png"  width = "800" />
 </div>
 
-The  solution mainly includes 4 parts, namely: PP-LCNet lightweight backbone network, SSLD pre-trained model, Ensemble Data Augmentation (EDA) and SKL-UGI knowledge distillation algorithm. In addition, we also adopt the method of hyperparameter search to efficiently optimize the hyperparameters in training. Below, we take the person exists or not scene as an example to illustrate the solution.
+The  solution mainly includes 4 parts, namely: PP-LCNet lightweight backbone network, SSLD pre-trained model, Ensemble Data Augmentation (EDA) and SKL-UGI knowledge distillation algorithm. In addition, we also adopt the method of hyperparameters searching to efficiently optimize the hyperparameters in training. Below, we take the person exists or not scene as an example to illustrate the solution.
 
 **Note**：For some specific scenarios, we provide basic training documents for reference, such as [person exists or not classification model](PULC_person_exists_en.md), etc. You can find these documents [here](./PULC_model_list_en.md). If the methods in these documents do not meet your needs, or if you need a custom training task, you can refer to this document.
 
@@ -201,22 +201,22 @@ We also used the same optimization strategy in the other 8 scenarios and got the
 | Text Image Orientation Classification | SwinTransformer_tiny |99.12 | PPLCNet_x1_0 | 99.06 |
 | Text-line Orientation Classification | SwinTransformer_tiny | 93.61 | PPLCNet_x1_0 | 96.01 |
 | Language Classification | SwinTransformer_tiny | 98.12 | PPLCNet_x1_0 | 99.26 |
-    
+
 
 It can be seen from the results that the PULC scheme can improve the model accuracy in multiple application scenarios. Using the PULC scheme can greatly reduce the workload of model optimization and quickly obtain models with higher accuracy.
-    
+
 
 <a name="4"></a>
 
-### 4. Hyperparameter Search
+### 4. Hyperparameters Searching
 
-In the above training process, we adjusted parameters such as learning rate, data augmentation probability, and stage learning rate mult list. The optimal values of these parameters may not be the same in different scenarios. We provide a quick hyperparameter search script to automate the process of hyperparameter tuning. This script traverses the parameters in the search value list to replace the parameters in the default configuration, then trains in sequence, and finally selects the parameters corresponding to the model with the highest accuracy as the search result.
+In the above training process, we adjusted parameters such as learning rate, data augmentation probability, and stage learning rate mult list. The optimal values of these parameters may not be the same in different scenarios. We provide a quick hyperparameters searching script to automate the process of hyperparameter tuning. This script traverses the parameters in the search value list to replace the parameters in the default configuration, then trains in sequence, and finally selects the parameters corresponding to the model with the highest accuracy as the search result.
 
 <a name="4.1"></a>
 
 #### 4.1 Search based on default configuration
 
-The configuration file [search.yaml](../../../ppcls/configs/PULC/person_exists/search.yaml) defines the configuration of hyperparameter search in person exists or not scenarios. Use the following commands to complete hyperparameter search.
+The configuration file [search.yaml](../../../ppcls/configs/PULC/person_exists/search.yaml) defines the configuration of hyperparameters searching in person exists or not scenarios. Use the following commands to complete hyperparameters searching.
 
 ```bash
 python3 tools/search_strategy.py -c ppcls/configs/PULC/person_exists/search.yaml
@@ -228,8 +228,8 @@ python3 tools/search_strategy.py -c ppcls/configs/PULC/person_exists/search.yaml
 
 #### 4.2 Custom search configuration
 
-    
-You can also modify the configuration of hyperparameter search based on training results or your parameter tuning experience.
+
+You can also modify the configuration of hyperparameters searching based on training results or your parameter tuning experience.
 
 Modify the `search_values` field in `lrs` to modify the list of learning rate search values;
 

docs/en/PULC/PULC_vehicle_attribute_en.md

+# PULC Recognition Model of Vehicle Attribute
+
+------
+
+## Catalogue
+
+- [1. Introduction](#1)
+- [2. Quick Start](#2)
+    - [2.1 PaddlePaddle Installation](#2.1)
+    - [2.2 PaddleClas Installation](#2.2)
+    - [2.3 Prediction](#2.3)
+- [3. Training, Evaluation and Inference](#3)
+    - [3.1 Installation](#3.1)
+    - [3.2 Dataset](#3.2)
+      - [3.2.1 Dataset Introduction](#3.2.1)
+      - [3.2.2 Getting Dataset](#3.2.2)
+    - [3.3 Training](#3.3)
+    - [3.4 Evaluation](#3.4)
+    - [3.5 Inference](#3.5)
+- [4. Model Compression](#4)
+  - [4.1 SKL-UGI Knowledge Distillation](#4.1)
+    - [4.1.1 Teacher Model Training](#4.1.1)
+    - [4.1.2 Knowledge Distillation Training](#4.1.2)
+- [5. SHAS](#5)
+- [6. Inference Deployment](#6)
+  - [6.1 Getting Paddle Inference Model](#6.1)
+    - [6.1.1 Exporting Paddle Inference Model](#6.1.1)
+    - [6.1.2 Downloading Inference Model](#6.1.2)
+  - [6.2 Prediction with Python](#6.2)
+    - [6.2.1 Image Prediction](#6.2.1)
+    - [6.2.2 Images Prediction](#6.2.2)
+  - [6.3 Deployment with C++](#6.3)
+  - [6.4 Deployment as Service](#6.4)
+  - [6.5 Deployment on Mobile](#6.5)
+  - [6.6 Converting To ONNX and Deployment](#6.6)
+
+<a name="1"></a>
+
+## 1. Introduction
+
+This case provides a way for users to quickly build a lightweight, high-precision and practical classification model of vehicle attribute using PaddleClas PULC (Practical Ultra Lightweight image Classification). The model can be widely used in Vehicle identification, road monitoring and other scenarios.
+
+The following table lists the relevant indicators of the model. The first three lines means that using Res2Net200_vd_26w_4s, ResNet50 and MobileNetV3_small_x0_35 as the backbone to training. The fourth to seventh lines means that the backbone is replaced by PPLCNet, additional use of EDA strategy and additional use of EDA strategy and SKL-UGI knowledge distillation strategy.
+
+
+| Backbone | mA（%） | Latency(ms) | Size(M) | Training Strategy |
+|-------|-----------|----------|---------------|---------------|
+| Res2Net200_vd_26w_4s  | 91.36 | 79.46  | 293 | using ImageNet pretrained  |
+| ResNet50  | 89.98 | 12.83  | 92 | using ImageNet pretrained |
+| MobileNetV3_small_x0_35  | 87.41 | 2.91  | 2.8 | using ImageNet pretrained |
+| PPLCNet_x1_0  | 89.57 | 2.36  | 7.2 | using ImageNet pretrained |
+| PPLCNet_x1_0  | 90.07 | 2.36  | 7.2 | using SSLD pretrained  |
+| PPLCNet_x1_0  | 90.59 | 2.36  | 7.2 | using SSLD pretrained + EDA strategy|
+| <b>PPLCNet_x1_0<b>  | <b>90.81<b> | <b>2.36<b>  | <b>7.2<b> | using SSLD pretrained + EDA strategy + SKL-UGI knowledge distillation strategy|
+
+
+It can be seen from the table that the ma metric is higher when the backbone is Res2Net200_vd_26w_4s, but the inference speed is slower. After replacing the backbone with the lightweight model MobileNetV3_small_x0_35, the speed can be greatly improved, but the ma metric drops significantly. When the backbone is replaced by PPLCNet_x1_0, the ma metric is increased by 2 percentage points, and the speed is also increased by about 23%. On this basis, after using the SSLD pre-training model, the ma metric can be improved by about 0.5 percentage points without changing the inference speed. Further, when the EDA strategy is integrated, the ma metric can be improved by another 0.52 percentage points. Finally, using After SKL-UGI knowledge distillation, the ma metric can continue to improve by 0.23 percentage points. At this time, the ma metric of PPLCNet_x1_0 is only 0.55 percentage points away from Res2Net200_vd_26w_4s, but it is 32 times faster. The training method and deployment instructions of PULC will be introduced in detail below.
+
+
+**Note**:
+
+* The Latency is tested on Intel(R) Xeon(R) Gold 6148 CPU @ 2.40GHz. The MKLDNN is enabled and the number of threads is 10.
+* About PP-LCNet, please refer to [PP-LCNet Introduction](../models/PP-LCNet_en.md) and [PP-LCNet Paper](https://arxiv.org/abs/2109.15099).
+
+<a name="2"></a>
+
+## 2. Quick Start
+
+<a name="2.1"></a>  
+
+### 2.1 PaddlePaddle Installation
+
+- Run the following command to install if CUDA9 or CUDA10 is available.
+
+```bash
+python3 -m pip install paddlepaddle-gpu -i https://mirror.baidu.com/pypi/simple
+```
+
+- Run the following command to install if GPU device is unavailable.
+
+```bash
+python3 -m pip install paddlepaddle -i https://mirror.baidu.com/pypi/simple
+```
+
+Please refer to [PaddlePaddle Installation](https://www.paddlepaddle.org.cn/install/quick?docurl=/documentation/docs/en/install/pip/linux-pip_en.html) for more information about installation, for examples other versions.
+
+<a name="2.2"></a>  
+
+### 2.2 PaddleClas wheel Installation
+
+The command of PaddleClas installation as bellow:
+
+```bash
+pip3 install paddleclas
+```
+
+<a name="2.3"></a>
+
+### 2.3 Prediction
+
+First, please click [here](https://paddleclas.bj.bcebos.com/data/PULC/pulc_demo_imgs.zip) to download and unzip to get the test demo images.
+
+
+* Prediction with CLI
+
+```bash
+paddleclas --model_name=vehicle_attribute --infer_imgs=pulc_demo_imgs/vehicle_attribute/0002_c002_00030670_0.jpg
+```
+
+Results:
+```
+>>> result
+attributes: Color: (yellow, prob: 0.9893476963043213), Type: (hatchback, prob: 0.9734097719192505), output: [1, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 1, 0, 0, 0, 0, 0], filename: pulc_demo_imgs/vehicle_attribute/0002_c002_00030670_0.jpg
+Predict complete!
+```
+
+**Note**: If you want to test other images, only need to specify the `--infer_imgs` argument, and the directory containing images is also supported.
+
+* Prediction in Python
+
+```python
+import paddleclas
+model = paddleclas.PaddleClas(model_name="vehicle_attribute")
+result = model.predict(input_data="pulc_demo_imgs/vehicle_attribute/0002_c002_00030670_0.jpg")
+print(next(result))
+```
+
+**Note**: The `result` returned by `model.predict()` is a generator, so you need to use the `next()` function to call it or `for` loop to loop it. And it will predict with `batch_size` size batch and return the prediction results when called. The default `batch_size` is 1, and you also specify the `batch_size` when instantiating, such as `model = paddleclas.PaddleClas(model_name="vehicle_attribute",  batch_size=2)`. The result of demo above:
+
+```
+>>> result
+[{'attributes': 'Color: (yellow, prob: 0.9893476963043213), Type: (hatchback, prob: 0.9734097719192505)', 'output': [1, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 1, 0, 0, 0, 0, 0], 'filename': 'pulc_demo_imgs/vehicle_attribute/0002_c002_00030670_0.jpg'}]
+```
+
+<a name="3"></a>
+
+## 3. Training, Evaluation and Inference
+
+<a name="3.1"></a>  
+
+### 3.1 Installation
+
+Please refer to [Installation](../installation/install_paddleclas_en.md) to get the description about installation.
+
+<a name="3.2"></a>
+
+### 3.2 Dataset
+
+<a name="3.2.1"></a>
+
+#### 3.2.1 Dataset Introduction
+
+The data used in this case is the [pa100k dataset](https://www.v7labs.com/open-datasets/pa-100k).
+
+<a name="3.2.2"></a>  
+
+#### 3.2.2 Getting Dataset
+
+
+Part of the data visualization is shown below.
+
+<div align="center">
+<img src="../../images/PULC/docs/vehicle_attribute_data_demo.png"  width = "500" />
+</div>
+
+First, apply for and download data from [VeRi dataset official website](https://www.v7labs.com/open-datasets/veri-dataset), put it in the `dataset` directory of PaddleClas, the dataset directory name is `VeRi `, use the following command to enter the folder.
+
+
+```shell
+cd PaddleClas/dataset/VeRi/
+```
+
+Then use the following code to convert the label (you can execute the following command in the python terminal, or you can write it to a file and run the file using `python3 convert.py`).
+
+```python
+import os
+from xml.dom.minidom import parse
+
+vehicleids = []
+
+def convert_annotation(input_fp, output_fp):
+    in_file = open(input_fp)
+    list_file = open(output_fp, 'w')
+    tree = parse(in_file)
+
+    root = tree.documentElement
+
+    for item in root.getElementsByTagName("Item"):  
+        label = ['0', '0', '0', '0', '0', '0', '0', '0', '0', '0', '0', '0', '0', '0', '0', '0', '0', '0', '0']
+        if item.hasAttribute("imageName"):
+            name = item.getAttribute("imageName")
+        if item.hasAttribute("vehicleID"):
+            vehicleid = item.getAttribute("vehicleID")
+            if vehicleid not in vehicleids :
+                vehicleids.append(vehicleid)
+            vid = vehicleids.index(vehicleid)
+        if item.hasAttribute("colorID"):
+            colorid = int (item.getAttribute("colorID"))
+            label[colorid-1] = '1'
+        if item.hasAttribute("typeID"):
+            typeid = int (item.getAttribute("typeID"))
+            label[typeid+9] = '1'
+        label = ','.join(label)
+        list_file.write(os.path.join('image_train', name)  + "\t" + label + "\n")
+
+    list_file.close()
+
+convert_annotation('train_label.xml', 'train_list.txt')  #imagename vehiclenum colorid typeid
+convert_annotation('test_label.xml', 'test_list.txt')
+```
+
+
+After executing the above command, the `VeRi` directory has the following data:
+
+```
+VeRi
+├── image_train
+│   ├── 0001_c001_00016450_0.jpg
+│   ├── 0001_c001_00016460_0.jpg
+│   ├── 0001_c001_00016470_0.jpg
+...
+├── image_test
+│   ├── 0002_c002_00030600_0.jpg
+│   ├── 0002_c002_00030605_1.jpg
+│   ├── 0002_c002_00030615_1.jpg
+...
+...
+├── train_list.txt
+├── test_list.txt
+├── train_label.xml
+├── test_label.xml
+```
+
+where `train/` and `test/` are the training set and validation set, respectively. `train_list.txt` and `test_list.txt` are the converted label files for training and validation sets, respectively.
+
+
+<a name="3.3"></a>
+
+### 3.3 Training
+
+The details of training config in `./ppcls/configs/PULC/vehicle_attribute/PPLCNet_x1_0.yaml`. The command about training as follows:
+
+```shell
+export CUDA_VISIBLE_DEVICES=0,1,2,3
+python3 -m paddle.distributed.launch \
+    --gpus="0,1,2,3" \
+    tools/train.py \
+        -c ./ppcls/configs/PULC/vehicle_attribute/PPLCNet_x1_0.yaml
+```
+
+The best metric for the validation set is around `90.59%` (the dataset is small and generally fluctuates around 0.3%).
+
+
+<a name="3.4"></a>
+
+### 3.4 Evaluation
+
+After training, you can use the following commands to evaluate the model.
+
+```bash
+python3 tools/eval.py \
+    -c ./ppcls/configs/PULC/vehicle_attribute/PPLCNet_x1_0.yaml \
+    -o Global.pretrained_model="output/PPLCNet_x1_0/best_model"
+```
+
+Among the above command, the argument `-o Global.pretrained_model="output/PPLCNet_x1_0/best_model"` specify the path of the best model weight file. You can specify other path if needed.
+
+<a name="3.5"></a>
+
+### 3.5 Inference
+
+After training, you can use the model that trained to infer. Command is as follow:
+
+```python
+python3 tools/infer.py \
+    -c ./ppcls/configs/PULC/vehicle_attribute/PPLCNet_x1_0.yaml \
+    -o Global.pretrained_model=output/DistillationModel/best_model
+```
+
+The results:
+
+```
+[{'attr': 'Color: (yellow, prob: 0.9893478155136108), Type: (hatchback, prob: 0.9734100103378296)', 'pred': [1, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 1, 0, 0, 0, 0, 0], 'file_name': './deploy/images/PULC/vehicle_attribute/0002_c002_00030670_0.jpg'}]
+```
+
+**Note**:
+
+* Among the above command, argument `-o Global.pretrained_model="output/PPLCNet_x1_0/best_model"` specify the path of the best model weight file. You can specify other path if needed.
+* The default test image is `./deploy/images/PULC/vehicle_attribute/0002_c002_00030670_0.jpg`. And you can test other image, only need to specify the argument `-o Infer.infer_imgs=path_to_test_image`.
+
+<a name="4"></a>
+
+## 4. Model Compression
+
+<a name="4.1"></a>
+
+### 4.1 SKL-UGI Knowledge Distillation
+
+SKL-UGI is a simple but effective knowledge distillation algrithem proposed by PaddleClas.
+
+<!-- todo -->
+<!-- Please refer to [SKL-UGI](../advanced_tutorials/distillation/distillation_en.md) for more details. -->
+
+<a name="4.1.1"></a>
+
+#### 4.1.1 Teacher Model Training
+
+Training the teacher model with hyperparameters specified in `ppcls/configs/PULC/vehicle_attribute/PPLCNet_x1_0.yaml`. The command is as follow:
+
+```shell
+export CUDA_VISIBLE_DEVICES=0,1,2,3
+python3 -m paddle.distributed.launch \
+    --gpus="0,1,2,3" \
+    tools/train.py \
+        -c ./ppcls/configs/PULC/vehicle_attribute/PPLCNet_x1_0.yaml \
+        -o Arch.name=ResNet101_vd
+```
+
+The best metric for the validation set is around `91.60%`. The best teacher model weight would be saved in file `output/ResNet101_vd/best_model.pdparams`.
+
+<a name="4.1.2"></a>
+
+#### 4.1.2 Knowledge Distillation Training
+
+The training strategy, specified in training config file `ppcls/configs/PULC/vehicle_attribute/PPLCNet_x1_0_distillation.yaml`, the teacher model is `ResNet101_vd`, the student model is `PPLCNet_x1_0`. The command is as follow:
+
+```shell
+export CUDA_VISIBLE_DEVICES=0,1,2,3
+python3 -m paddle.distributed.launch \
+    --gpus="0,1,2,3" \
+    tools/train.py \
+        -c ./ppcls/configs/PULC/vehicle_attribute/PPLCNet_x1_0_distillation.yaml \
+        -o Arch.models.0.Teacher.pretrained=output/ResNet101_vd/best_model
+```
+
+The best metric for the validation set is around `90.81%`. The best student model weight would be saved in file `output/DistillationModel/best_model_student.pdparams`.
+
+<a name="5"></a>
+
+## 5. Hyperparameters Searching
+
+The hyperparameters used by [3.2 section](#3.2) and [4.1 section](#4.1) are according by `Hyperparameters Searching` in PaddleClas. If you want to get better results on your own dataset, you can refer to [Hyperparameters Searching](PULC_train_en.md#4) to get better hyperparameters.
+
+**Note**: This section is optional. Because the search process will take a long time, you can selectively run according to your specific. If not replace the dataset, you can ignore this section.
+
+<a name="6"></a>
+
+## 6. Inference Deployment
+
+<a name="6.1"></a>
+
+### 6.1 Getting Paddle Inference Model
+
+Paddle Inference is the original Inference Library of the PaddlePaddle, provides high-performance inference for server deployment. And compared with  directly based on the pretrained model, Paddle Inference can use tools to accelerate prediction, so as to achieve better inference performance. Please refer to [Paddle Inference](https://www.paddlepaddle.org.cn/documentation/docs/zh/guides/infer/inference/inference_cn.html) for more information.
+
+Paddle Inference need Paddle Inference Model to predict. Two process provided to get Paddle Inference Model. If want to use the provided by PaddleClas, you can download directly, click [Downloading Inference Model](#6.1.2).
+
+<a name="6.1.1"></a>
+
+### 6.1.1 Exporting Paddle Inference Model
+
+The command about exporting Paddle Inference Model is as follow:
+
+```bash
+python3 tools/export_model.py \
+    -c ./ppcls/configs/PULC/vehicle_attribute/PPLCNet_x1_0.yaml \
+    -o Global.pretrained_model=output/DistillationModel/best_model_student \
+    -o Global.save_inference_dir=deploy/models/PPLCNet_x1_0_vehicle_attribute_infer
+```
+
+After running above command, the inference model files would be saved in `PPLCNet_x1_0_vehicle_attribute_infer`, as shown below:
+
+```
+└── PPLCNet_x1_0_vehicle_attribute_infer
+    ├── inference.pdiparams
+    ├── inference.pdiparams.info
+    └── inference.pdmodel
+```
+
+**Note**: The best model is from knowledge distillation training. If knowledge distillation training is not used, the best model would be saved in `output/PPLCNet_x1_0/best_model.pdparams`.
+
+<a name="6.1.2"></a>
+
+### 6.1.2 Downloading Inference Model
+
+You can also download directly.
+
+```
+cd deploy/models
+# download the inference model and decompression
+wget https://paddleclas.bj.bcebos.com/models/PULC/vehicle_attribute_infer.tar && tar -xf vehicle_attribute_infer.tar
+```
+
+After decompression, the directory `models` should be shown below.
+
+```
+├── vehicle_attribute_infer
+│   ├── inference.pdiparams
+│   ├── inference.pdiparams.info
+│   └── inference.pdmodel
+```
+
+<a name="6.2"></a>
+
+### 6.2 Prediction with Python
+
+<a name="6.2.1"></a>  
+
+#### 6.2.1 Image Prediction
+
+Return the directory `deploy`:
+
+```
+cd ../
+```
+
+Run the following command to classify whether there are human in the image `../images/PULC/vehicle_attribute/0002_c002_00030670_0.jpg`.
+
+```shell
+# Use the following command to predict with GPU.
+python3.7 python/predict_cls.py -c configs/PULC/vehicle_attribute/inference_vehicle_attribute.yaml -o Global.use_gpu=True
+# Use the following command to predict with CPU.
+python3.7 python/predict_cls.py -c configs/PULC/vehicle_attribute/inference_vehicle_attribute.yaml -o Global.use_gpu=False
+```
+
+The prediction results:
+
+```
+0002_c002_00030670_0.jpg:     {'attributes': 'Color: (yellow, prob: 0.9893478155136108), Type: (hatchback, prob: 0.9734099507331848)', 'output': [1, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 1, 0, 0, 0, 0, 0]}
+```
+
+
+<a name="6.2.2"></a>  
+
+#### 6.2.2 Images Prediction
+
+If you want to predict images in directory, please specify the argument `Global.infer_imgs` as directory path by `-o Global.infer_imgs`. The command is as follow.
+
+```shell
+# Use the following command to predict with GPU. If want to replace with CPU, you can add argument -o Global.use_gpu=False
+python3.7 python/predict_cls.py -c configs/PULC/vehicle_attribute/inference_vehicle_attribute.yaml -o Global.infer_imgs="./images/PULC/vehicle_attribute/"
+```
+
+All prediction results will be printed, as shown below.
+
+```
+0002_c002_00030670_0.jpg:     {'attributes': 'Color: (yellow, prob: 0.9893476963043213), Type: (hatchback, prob: 0.9734097719192505)', 'output': [1, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 1, 0, 0, 0, 0, 0]}
+0014_c012_00040750_0.jpg:     {'attributes': 'Color: (red, prob: 0.999872088432312), Type: (sedan, prob: 0.999976634979248)', 'output': [0, 0, 0, 0, 1, 0, 0, 0, 0, 0, 1, 0, 0, 0, 0, 0, 0, 0, 0]}
+```
+
+Among the prediction results above, `someone` means that there is a human in the image, `nobody` means that there is no human in the image.
+
+<a name="6.3"></a>
+
+### 6.3 Deployment with C++
+
+PaddleClas provides an example about how to deploy with C++. Please refer to [Deployment with C++](../inference_deployment/cpp_deploy_en.md).
+
+<a name="6.4"></a>
+
+### 6.4 Deployment as Service
+
+Paddle Serving is a flexible, high-performance carrier for machine learning models, and supports different protocol, such as RESTful, gRPC, bRPC and so on, which provides different deployment solutions for a variety of heterogeneous hardware and operating system environments. Please refer [Paddle Serving](https://github.com/PaddlePaddle/Serving) for more information.
+
+PaddleClas provides an example about how to deploy as service by Paddle Serving. Please refer to [Paddle Serving Deployment](../inference_deployment/paddle_serving_deploy_en.md).
+
+<a name="6.5"></a>
+
+### 6.5 Deployment on Mobile
+
+Paddle-Lite is an open source deep learning framework that designed to make easy to perform inference on mobile, embeded, and IoT devices. Please refer to [Paddle-Lite](https://github.com/PaddlePaddle/Paddle-Lite) for more information.
+
+PaddleClas provides an example of how to deploy on mobile by Paddle-Lite. Please refer to [Paddle-Lite deployment](../inference_deployment/paddle_lite_deploy_en.md).
+
+<a name="6.6"></a>
+
+### 6.6 Converting To ONNX and Deployment
+
+Paddle2ONNX support convert Paddle Inference model to ONNX model. And you can deploy with ONNX model on different inference engine, such as TensorRT, OpenVINO, MNN/TNN, NCNN and so on. About Paddle2ONNX details, please refer to [Paddle2ONNX](https://github.com/PaddlePaddle/Paddle2ONNX).
+
+PaddleClas provides an example of how to convert Paddle Inference model to ONNX model by paddle2onnx toolkit and predict by ONNX model. You can refer to [paddle2onnx](../../../deploy/paddle2onnx/readme_en.md) for deployment details.

docs/en/algorithm_introduction/ImageNet_models_en.md

@@ -541,9 +541,9 @@ The accuracy and speed indicators of MobileViT series models are shown in the fo
 
 | Model       | Top-1 Acc | Top-5 Acc | time(ms)<br>bs=1 | time(ms)<br>bs=4 | time(ms)<br/>bs=8 | FLOPs(M) | Params(M) | Pretrained Model Download Address                                               | Inference Model Download Address                                      |
 | ---------- | --------- | --------- | ---------------- | ---------------- | -------- | --------- | ------------------------------------------------------------ | ------------------------------------------------------------ | ------------------------------------------------------------ |
-|  MobileViT_XXS    | 0.6867 | 0.8878 | - | - | - | 1849.35  |  5.59   | [Download link](https://paddle-imagenet-models-name.bj.bcebos.com/dygraph/MobileViT_XXS_pretrained.pdparams) | [Download link](https://paddle-imagenet-models-name.bj.bcebos.com/dygraph/inference/MobileViT_XXS_infer.tar) |
+|  MobileViT_XXS    | 0.6867 | 0.8878 | - | - | - | 337.24  |  1.28   | [Download link](https://paddle-imagenet-models-name.bj.bcebos.com/dygraph/MobileViT_XXS_pretrained.pdparams) | [Download link](https://paddle-imagenet-models-name.bj.bcebos.com/dygraph/inference/MobileViT_XXS_infer.tar) |
 |  MobileViT_XS    | 0.7454 | 0.9227 | - | - | - | 930.75  |  2.33   | [Download link](https://paddle-imagenet-models-name.bj.bcebos.com/dygraph/MobileViT_XS_pretrained.pdparams) | [Download link](https://paddle-imagenet-models-name.bj.bcebos.com/dygraph/inference/MobileViT_XS_infer.tar) |
-|  MobileViT_S    | 0.7814 | 0.9413 | - | - | - | 337.24  |   1.28   | [Download link](https://paddle-imagenet-models-name.bj.bcebos.com/dygraph/MobileViT_S_pretrained.pdparams) | [Download link](https://paddle-imagenet-models-name.bj.bcebos.com/dygraph/inference/MobileViT_S_infer.tar) |
+|  MobileViT_S    | 0.7814 | 0.9413 | - | - | - | 1849.35  |   5.59   | [Download link](https://paddle-imagenet-models-name.bj.bcebos.com/dygraph/MobileViT_S_pretrained.pdparams) | [Download link](https://paddle-imagenet-models-name.bj.bcebos.com/dygraph/inference/MobileViT_S_infer.tar) |
 
 <a name="26"></a>
 

docs/en/inference_deployment/whl_deploy_en.md

@@ -8,6 +8,8 @@ PaddleClas supports Python wheel package for prediction. At present, PaddleClas
 
 - [1. Installation](#1)
 - [2. Quick Start](#2)
+   - [2.1 ImageNet1k models](#2.1)
+   - [2.2 PULC models](#2.2)
 - [3. Definition of Parameters](#3)
 - [4. More usage](#4)
    - [4.1 View help information](#4.1)
@@ -75,7 +77,6 @@ filename: docs/images/inference_deployment/whl_demo.jpg, top-5, class_ids: [8, 7
 Predict complete!
 ```
 
-
 <a name="2.2"></a>
 
 ### 2.2 PULC models

docs/en/models/MobileViT_en.md

@@ -18,6 +18,6 @@ MobileViT is a lightweight visual Transformer network that can be used as a gene
 
 | Models           | Top1 | Top5 | Reference<br>top1 | Reference<br>top5 | FLOPs<br>(M) | Params<br>(M) |
 |:--:|:--:|:--:|:--:|:--:|:--:|:--:|
-| MobileViT_XXS    | 0.6867 | 0.8878 | 0.690 | - | 1849.35  | 5.59   |
+| MobileViT_XXS    | 0.6867 | 0.8878 | 0.690 | - | 337.24  | 1.28   |
 | MobileViT_XS    | 0.7454 | 0.9227 | 0.747 | - | 930.75  | 2.33   |
-| MobileViT_S    | 0.7814 | 0.9413 | 0.783 | - | 337.24  | 1.28   |
+| MobileViT_S    | 0.7814 | 0.9413 | 0.783 | - | 1849.35  | 5.59   |

docs/zh_CN/PULC/PULC_car_exists.md

@@ -40,36 +40,36 @@
 
 ## 1. 模型和应用场景介绍
 
-该案例提供了用户使用 PaddleClas 的超轻量图像分类方案（PULC，Practical Ultra Lightweight Classification）快速构建轻量级、高精度、可落地的有车/无车的分类模型。该模型可以广泛应用于如监控场景、海量数据过滤场景等。
+该案例提供了用户使用 PaddleClas 的超轻量图像分类方案（PULC，Practical Ultra Lightweight image Classification）快速构建轻量级、高精度、可落地的有车/无车的分类模型。该模型可以广泛应用于如监控场景、海量数据过滤场景等。
 
 下表列出了判断图片中是否有车的二分类模型的相关指标，前两行展现了使用 SwinTranformer_tiny 和 MobileNetV3_small_x0_35 作为 backbone 训练得到的模型的相关指标，第三行至第六行依次展现了替换 backbone 为 PPLCNet_x1_0、使用 SSLD 预训练模型、使用 SSLD 预训练模型 + EDA 策略、使用 SSLD 预训练模型 + EDA 策略 + SKL-UGI 知识蒸馏策略训练得到的模型的相关指标。
 
 
 | 模型 | Tpr（%）@Fpr0.01 | 延时（ms） | 存储（M） | 策略 |
 |-------|----------------|----------|---------------|---------------|
-| SwinTranformer_tiny  | 97.71          | 95.30  | 107 | 使用 ImageNet 预训练模型 |
-| MobileNetV3_small_x0_35  | 81.23          | 2.85  | 1.6 | 使用 ImageNet 预训练模型 |
-| PPLCNet_x1_0  | 94.72          | 2.12  | 6.5 | 使用 ImageNet 预训练模型 |
-| PPLCNet_x1_0  | 95.48          | 2.12  | 6.5 | 使用 SSLD 预训练模型 |
-| PPLCNet_x1_0  | 95.48          | 2.12  | 6.5 | 使用 SSLD 预训练模型+EDA 策略|
-| <b>PPLCNet_x1_0<b>  | <b>95.92<b>    | <b>2.12<b>  | <b>6.5<b> | 使用 SSLD 预训练模型+EDA 策略+SKL-UGI 知识蒸馏策略|
+| SwinTranformer_tiny  | 97.71          | 95.30  | 111 | 使用 ImageNet 预训练模型 |
+| MobileNetV3_small_x0_35  | 81.23          | 2.85  | 2.7 | 使用 ImageNet 预训练模型 |
+| PPLCNet_x1_0  | 94.72          | 2.12  | 7.1 | 使用 ImageNet 预训练模型 |
+| PPLCNet_x1_0  | 95.48          | 2.12  | 7.1 | 使用 SSLD 预训练模型 |
+| PPLCNet_x1_0  | 95.48          | 2.12  | 7.1 | 使用 SSLD 预训练模型+EDA 策略|
+| <b>PPLCNet_x1_0<b>  | <b>95.92<b>    | <b>2.12<b>  | <b>7.1<b> | 使用 SSLD 预训练模型+EDA 策略+SKL-UGI 知识蒸馏策略|
 
-从表中可以看出，backbone 为 SwinTranformer_tiny 时精度较高，但是推理速度较慢。将 backboone 替换为轻量级模型 MobileNetV3_small_x0_35 后，速度可以大幅提升，但是会导致精度大幅下降。将 backbone 替换为速度更快的 PPLCNet_x1_0 时，精度较 MobileNetV3_small_x0_35 高 13 个百分点，与此同时速度依旧可以快 20% 以上。在此基础上，使用 SSLD 预训练模型后，在不改变推理速度的前提下，精度可以提升约 0.7 个百分点，进一步地，在使用 SKL-UGI 知识蒸馏后，精度可以继续提升 0.44 个百分点。此时，PPLCNet_x1_0 达到了接近 SwinTranformer_tiny 模型的精度，但是速度快 40 多倍。关于 PULC 的训练方法和推理部署方法将在下面详细介绍。
+从表中可以看出，backbone 为 SwinTranformer_tiny 时精度较高，但是推理速度较慢。将 backbone 替换为轻量级模型 MobileNetV3_small_x0_35 后，速度可以大幅提升，但是会导致精度大幅下降。将 backbone 替换为速度更快的 PPLCNet_x1_0 时，精度较 MobileNetV3_small_x0_35 高 13 个百分点，与此同时速度依旧可以快 20% 以上。在此基础上，使用 SSLD 预训练模型后，在不改变推理速度的前提下，精度可以提升约 0.7 个百分点，进一步地，在使用 SKL-UGI 知识蒸馏后，精度可以继续提升 0.44 个百分点。此时，PPLCNet_x1_0 达到了接近 SwinTranformer_tiny 模型的精度，但是速度快 40 多倍。关于 PULC 的训练方法和推理部署方法将在下面详细介绍。
 
 **备注：**
 
 * `Tpr`指标的介绍可以参考 [3.3节](#3.3)的备注部分，延时是基于 Intel(R) Xeon(R) Gold 6148 CPU @ 2.40GHz 测试得到，开启 MKLDNN 加速策略，线程数为10。
 * 关于PP-LCNet的介绍可以参考[PP-LCNet介绍](../models/PP-LCNet.md)，相关论文可以查阅[PP-LCNet paper](https://arxiv.org/abs/2109.15099)。
-    
-    
+
+
 <a name="2"></a>
 
 ## 2. 模型快速体验
 
 <a name="2.1"></a>  
-    
+
 ### 2.1 安装 paddlepaddle
-    
+
 - 您的机器安装的是 CUDA9 或 CUDA10，请运行以下命令安装
 
 ```bash
@@ -81,11 +81,11 @@ python3 -m pip install paddlepaddle-gpu -i https://mirror.baidu.com/pypi/simple
 ```bash
 python3 -m pip install paddlepaddle -i https://mirror.baidu.com/pypi/simple
 ```
-    
+
 更多的版本需求，请参照[飞桨官网安装文档](https://www.paddlepaddle.org.cn/install/quick)中的说明进行操作。
-    
+
 <a name="2.2"></a>
-    
+
 ### 2.2 安装 paddleclas
 
 使用如下命令快速安装 paddleclas
@@ -93,11 +93,11 @@ python3 -m pip install paddlepaddle -i https://mirror.baidu.com/pypi/simple
 ```  
 pip3 install paddleclas
 ```
-    
+
 <a name="2.3"></a>
 
 ### 2.3 预测
-    
+
 点击[这里](https://paddleclas.bj.bcebos.com/data/PULC/pulc_demo_imgs.zip)下载 demo 数据并解压，然后在终端中切换到相应目录。
 
 * 使用命令行快速预测
@@ -130,7 +130,7 @@ print(next(result))
 >>> result
 [{'class_ids': [1], 'scores': [0.9871138], 'label_names': ['contains_car'], 'filename': 'pulc_demo_imgs/car_exists/objects365_00001507.jpeg'}]
 ```
-    
+
 
 <a name="3"></a>
 
@@ -150,7 +150,7 @@ print(next(result))
 
 #### 3.2.1 数据集来源
 
-本案例中所使用的所有数据集均为开源数据，`train`和`val` 集合均为[Objects365 数据](https://www.objects365.org/overview.html)的训练集的子集，`ImageNet_val` 为[ImageNet-1k 数据](https://www.image-net.org/)的验证集。
+本案例中所使用的所有数据集均为开源数据，`train`和`val` 集合均为[Objects365 数据](https://www.objects365.org/overview.html)的子集，`ImageNet_val` 为[ImageNet-1k 数据](https://www.image-net.org/)的验证集。
 
 <a name="3.2.2"></a>  
 
@@ -326,7 +326,7 @@ python3 -m paddle.distributed.launch \
 
 ## 5. 超参搜索
 
-在 [3.3 节](#3.3)和 [4.1 节](#4.1)所使用的超参数是根据 PaddleClas 提供的 `SHAS 超参数搜索策略` 搜索得到的，如果希望在自己的数据集上得到更好的结果，可以参考[SHAS 超参数搜索策略](PULC_train.md#4-超参搜索)来获得更好的训练超参数。
+在 [3.3 节](#3.3)和 [4.1 节](#4.1)所使用的超参数是根据 PaddleClas 提供的 `超参数搜索策略` 搜索得到的，如果希望在自己的数据集上得到更好的结果，可以参考[超参数搜索策略](PULC_train.md#4-超参搜索)来获得更好的训练超参数。
 
 **备注：** 此部分内容是可选内容，搜索过程需要较长的时间，您可以根据自己的硬件情况来选择执行。如果没有更换数据集，可以忽略此节内容。
 

docs/zh_CN/PULC/PULC_language_classification.md

@@ -36,20 +36,20 @@
 
 ## 1. 模型和应用场景介绍
 
-该案例提供了用户使用 PaddleClas 的超轻量图像分类方案（PULC，Practical Ultra Lightweight Classification）快速构建轻量级、高精度、可落地的语种分类模型。使用该方法训练得到的模型可以快速判断图片中的文字语种，该模型可以广泛应用于金融、政务等各种涉及多语种OCR处理的场景中。
+该案例提供了用户使用 PaddleClas 的超轻量图像分类方案（PULC，Practical Ultra Lightweight image Classification）快速构建轻量级、高精度、可落地的语种分类模型。使用该方法训练得到的模型可以快速判断图片中的文字语种，该模型可以广泛应用于金融、政务等各种涉及多语种OCR处理的场景中。
 
 下表列出了语种分类模型的相关指标，前两行展现了使用 SwinTranformer_tiny 和 MobileNetV3_small_x0_35 作为 backbone 训练得到的模型的相关指标，第三行至第六行依次展现了替换 backbone 为 PPLCNet_x1_0、使用 SSLD 预训练模型、使用 SSLD 预训练模型 + EDA 策略、使用 SSLD 预训练模型 + EDA 策略 + SKL-UGI 知识蒸馏策略训练得到的模型的相关指标。其中替换 backbone 为 PPLCNet_x1_0时，将数据预处理时的输入尺寸变为[192,48]，且网络的下采样stride调整为[2, [2, 1], [2, 1], [2, 1], [2, 1]]。
 
 | 模型                    | 精度      | 延时     | 存储    | 策略                                           |
 | ----------------------- | --------- | -------- | ------- | ---------------------------------------------- |
-| SwinTranformer_tiny     | 98.12     | 89.09    | 107     | 使用ImageNet预训练模型                         |
-| MobileNetV3_small_x0_35 | 95.92     | 2.98     | 17      | 使用ImageNet预训练模型                         |
-| PPLCNet_x1_0            | 98.35     | 2.58     | 6.5     | 使用ImageNet预训练模型                         |
-| PPLCNet_x1_0            | 98.7      | 2.58     | 6.5     | 使用SSLD预训练模型                             |
-| PPLCNet_x1_0            | 99.12     | 2.58     | 6.5     | 使用SSLD预训练模型+EDA策略                     |
-| **PPLCNet_x1_0**        | **99.26** | **2.58** | **6.5** | 使用SSLD预训练模型+EDA策略+SKL-UGI知识蒸馏策略 |
+| SwinTranformer_tiny     | 98.12     | 89.09    | 111     | 使用ImageNet预训练模型                         |
+| MobileNetV3_small_x0_35 | 95.92     | 2.98     | 3.7      | 使用ImageNet预训练模型                         |
+| PPLCNet_x1_0            | 98.35     | 2.58     | 7.1     | 使用ImageNet预训练模型                         |
+| PPLCNet_x1_0            | 98.7      | 2.58     | 7.1     | 使用SSLD预训练模型                             |
+| PPLCNet_x1_0            | 99.12     | 2.58     | 7.1     | 使用SSLD预训练模型+EDA策略                     |
+| **PPLCNet_x1_0**        | **99.26** | **2.58** | **7.1** | 使用SSLD预训练模型+EDA策略+SKL-UGI知识蒸馏策略 |
 
-从表中可以看出，backbone 为 SwinTranformer_tiny 时精度比较高，但是推理速度较慢。将 backboone 替换为轻量级模型 MobileNetV3_small_x0_35 后，速度提升明显，但精度有了大幅下降。将 backbone 替换为 PPLCNet_x1_0 且调整预处理输入尺寸和网络的下采样stride时，速度略为提升，同时精度较 MobileNetV3_large_x1_0 高2.43个百分点。在此基础上，使用 SSLD 预训练模型后，在不改变推理速度的前提下，精度可以提升 0.35 个百分点，进一步地，当融合EDA策略后，精度可以再提升 0.42 个百分点，最后，在使用 SKL-UGI 知识蒸馏后，精度可以继续提升 0.14 个百分点。此时，PPLCNet_x1_0 超过了 SwinTranformer_tiny 模型的精度，并且速度有了明显提升。关于 PULC 的训练方法和推理部署方法将在下面详细介绍。
+从表中可以看出，backbone 为 SwinTranformer_tiny 时精度比较高，但是推理速度较慢。将 backbone 替换为轻量级模型 MobileNetV3_small_x0_35 后，速度提升明显，但精度有了大幅下降。将 backbone 替换为 PPLCNet_x1_0 且调整预处理输入尺寸和网络的下采样stride时，速度略为提升，同时精度较 MobileNetV3_large_x1_0 高2.43个百分点。在此基础上，使用 SSLD 预训练模型后，在不改变推理速度的前提下，精度可以提升 0.35 个百分点，进一步地，当融合EDA策略后，精度可以再提升 0.42 个百分点，最后，在使用 SKL-UGI 知识蒸馏后，精度可以继续提升 0.14 个百分点。此时，PPLCNet_x1_0 超过了 SwinTranformer_tiny 模型的精度，并且速度有了明显提升。关于 PULC 的训练方法和推理部署方法将在下面详细介绍。
 
 **备注：**
 
@@ -60,9 +60,9 @@
 ## 2. 模型快速体验
 
 <a name="2.1"></a>  
-    
+
 ### 2.1 安装 paddlepaddle
-    
+
 - 您的机器安装的是 CUDA9 或 CUDA10，请运行以下命令安装
 
 ```bash
@@ -74,23 +74,23 @@ python3 -m pip install paddlepaddle-gpu -i https://mirror.baidu.com/pypi/simple
 ```bash
 python3 -m pip install paddlepaddle -i https://mirror.baidu.com/pypi/simple
 ```
-    
+
 更多的版本需求，请参照[飞桨官网安装文档](https://www.paddlepaddle.org.cn/install/quick)中的说明进行操作。
-    
+
 <a name="2.2"></a>  
-    
+
 ### 2.2 安装 paddleclas
 
 使用如下命令快速安装 paddleclas
 
 ```  
 pip3 install paddleclas
-``` 
-    
+```
+
 <a name="2.3"></a>
 
 ### 2.3 预测
-    
+
 点击[这里](https://paddleclas.bj.bcebos.com/data/PULC/pulc_demo_imgs.zip)下载 demo 数据并解压，然后在终端中切换到相应目录。
 
 * 使用命令行快速预测
@@ -309,7 +309,7 @@ python3 -m paddle.distributed.launch \
 
 ## 5. 超参搜索
 
-在 [3.2 节](#3.2)和 [4.1 节](#4.1)所使用的超参数是根据 PaddleClas 提供的 `SHAS 超参数搜索策略` 搜索得到的，如果希望在自己的数据集上得到更好的结果，可以参考[SHAS 超参数搜索策略](PULC_train.md#4-超参搜索)来获得更好的训练超参数。
+在 [3.2 节](#3.2)和 [4.1 节](#4.1)所使用的超参数是根据 PaddleClas 提供的 `超参数搜索策略` 搜索得到的，如果希望在自己的数据集上得到更好的结果，可以参考[超参数搜索策略](PULC_train.md#4-超参搜索)来获得更好的训练超参数。
 
 **备注：** 此部分内容是可选内容，搜索过程需要较长的时间，您可以根据自己的硬件情况来选择执行。如果没有更换数据集，可以忽略此节内容。
 

docs/zh_CN/PULC/PULC_model_list.md

@@ -7,15 +7,15 @@
 
 |模型名称|模型简介|模型精度 |模型大小|推理耗时|下载地址|
 | --- | --- | --- | --- | --- | --- |
-| person_exists |[PULC有人/无人分类模型](PULC_person_exists.md)| 95.60 |6.5M|2.58ms|[推理模型](https://paddleclas.bj.bcebos.com/models/PULC/inference/person_exists_infer.tar) / [预训练模型](https://paddleclas.bj.bcebos.com/models/PULC/pretrained/person_exists_pretrained.pdparams)|
-| person_attribute |[PULC人体属性识别模型](PULC_person_attribute.md)| 78.59 |6.6M|2.01ms|[推理模型](https://paddleclas.bj.bcebos.com/models/PULC/inference/person_attribute_infer.tar) / [预训练模型](https://paddleclas.bj.bcebos.com/models/PULC/pretrained/person_attribute_pretrained.pdparams)|
-| safety_helmet |[PULC佩戴安全帽分类模型](PULC_safety_helmet.md)| 99.38 |6.5M|2.03ms|[推理模型](https://paddleclas.bj.bcebos.com/models/PULC/inference/safety_helmet_infer.tar) / [预训练模型](https://paddleclas.bj.bcebos.com/models/PULC/pretrained/safety_helmet_pretrained.pdparams)|
+| person_exists |[PULC有人/无人分类模型](PULC_person_exists.md)| 96.23 |7.0M|2.58ms|[推理模型](https://paddleclas.bj.bcebos.com/models/PULC/inference/person_exists_infer.tar) / [预训练模型](https://paddleclas.bj.bcebos.com/models/PULC/pretrained/person_exists_pretrained.pdparams)|
+| person_attribute |[PULC人体属性识别模型](PULC_person_attribute.md)| 78.59 |7.2M|2.01ms|[推理模型](https://paddleclas.bj.bcebos.com/models/PULC/inference/person_attribute_infer.tar) / [预训练模型](https://paddleclas.bj.bcebos.com/models/PULC/pretrained/person_attribute_pretrained.pdparams)|
+| safety_helmet |[PULC佩戴安全帽分类模型](PULC_safety_helmet.md)| 99.38 |7.1M|2.03ms|[推理模型](https://paddleclas.bj.bcebos.com/models/PULC/inference/safety_helmet_infer.tar) / [预训练模型](https://paddleclas.bj.bcebos.com/models/PULC/pretrained/safety_helmet_pretrained.pdparams)|
 | traffic_sign |[PULC交通标志分类模型](PULC_traffic_sign.md)| 98.35 |8.2M|2.10ms|[推理模型](https://paddleclas.bj.bcebos.com/models/PULC/inference/traffic_sign_infer.tar) / [预训练模型](https://paddleclas.bj.bcebos.com/models/PULC/pretrained/traffic_sign_pretrained.pdparams)|
 | vehicle_attribute |[PULC车辆属性识别模型](PULC_vehicle_attribute.md)| 90.81 |7.2M|2.36ms|[推理模型](https://paddleclas.bj.bcebos.com/models/PULC/inference/vehicle_attribute_infer.tar) / [预训练模型](https://paddleclas.bj.bcebos.com/models/PULC/pretrained/vehicle_attribute_pretrained.pdparams)|
-| car_exists |[PULC有车/无车分类模型](PULC_car_exists.md) | 95.92 | 6.6M | 2.38ms |[推理模型](https://paddleclas.bj.bcebos.com/models/PULC/inference/car_exists_infer.tar) / [预训练模型](https://paddleclas.bj.bcebos.com/models/PULC/pretrained/car_exists_pretrained.pdparams)|
-| text_image_orientation |[PULC含文字图像方向分类模型](PULC_text_image_orientation.md)| 99.06 | 6.5M | 2.16ms |[推理模型](https://paddleclas.bj.bcebos.com/models/PULC/inference/text_image_orientation_infer.tar) / [预训练模型](https://paddleclas.bj.bcebos.com/models/PULC/pretrained/text_image_orientation_pretrained.pdparams)|
-| textline_orientation |[PULC文本行方向分类模型](PULC_textline_orientation.md)| 96.01 |6.5M|2.72ms|[推理模型](https://paddleclas.bj.bcebos.com/models/PULC/inference/textline_orientation_infer.tar) / [预训练模型](https://paddleclas.bj.bcebos.com/models/PULC/pretrained/textline_orientation_pretrained.pdparams)|
-| language_classification |[PULC语种分类模型](PULC_language_classification.md)| 99.26 |6.5M|2.58ms|[推理模型](https://paddleclas.bj.bcebos.com/models/PULC/inference/language_classification_infer.tar) / [预训练模型](https://paddleclas.bj.bcebos.com/models/PULC/pretrained/language_classification_pretrained.pdparams)|
+| car_exists |[PULC有车/无车分类模型](PULC_car_exists.md) | 95.92 | 7.1M | 2.38ms |[推理模型](https://paddleclas.bj.bcebos.com/models/PULC/inference/car_exists_infer.tar) / [预训练模型](https://paddleclas.bj.bcebos.com/models/PULC/pretrained/car_exists_pretrained.pdparams)|
+| text_image_orientation |[PULC含文字图像方向分类模型](PULC_text_image_orientation.md)| 99.06 | 7.1M | 2.16ms |[推理模型](https://paddleclas.bj.bcebos.com/models/PULC/inference/text_image_orientation_infer.tar) / [预训练模型](https://paddleclas.bj.bcebos.com/models/PULC/pretrained/text_image_orientation_pretrained.pdparams)|
+| textline_orientation |[PULC文本行方向分类模型](PULC_textline_orientation.md)| 96.01 |7.0M|2.72ms|[推理模型](https://paddleclas.bj.bcebos.com/models/PULC/inference/textline_orientation_infer.tar) / [预训练模型](https://paddleclas.bj.bcebos.com/models/PULC/pretrained/textline_orientation_pretrained.pdparams)|
+| language_classification |[PULC语种分类模型](PULC_language_classification.md)| 99.26 |7.1M|2.58ms|[推理模型](https://paddleclas.bj.bcebos.com/models/PULC/inference/language_classification_infer.tar) / [预训练模型](https://paddleclas.bj.bcebos.com/models/PULC/pretrained/language_classification_pretrained.pdparams)|
 
 
 **备注：**

docs/zh_CN/PULC/PULC_person_attribute.md

@@ -40,25 +40,26 @@
 
 ## 1. 模型和应用场景介绍
 
-该案例提供了用户使用 PaddleClas 的超轻量图像分类方案（PULC，Practical Ultra Lightweight Classification）快速构建轻量级、高精度、可落地的人体属性识别模型。该模型可以广泛应用于行人分析、行人跟踪等场景。
+该案例提供了用户使用 PaddleClas 的超轻量图像分类方案（PULC，Practical Ultra Lightweight image Classification）快速构建轻量级、高精度、可落地的人体属性识别模型。该模型可以广泛应用于行人分析、行人跟踪等场景。
 
-下表列出了不同人体属性识别模型的相关指标，前两行展现了使用 SwinTransformer_tiny、Res2Net200_vd_26w_4s 和 MobileNetV3_small_x0_35 作为 backbone 训练得到的模型的相关指标，第三行至第六行依次展现了替换 backbone 为 PPLCNet_x1_0、使用 SSLD 预训练模型、使用 SSLD 预训练模型 + EDA 策略、使用 SSLD 预训练模型 + EDA 策略 + SKL-UGI 知识蒸馏策略训练得到的模型的相关指标。
+下表列出了不同人体属性识别模型的相关指标，前三行展现了使用 SwinTransformer_tiny、Res2Net200_vd_26w_4s 和 MobileNetV3_small_x0_35 作为 backbone 训练得到的模型的相关指标，第四行至第七行依次展现了替换 backbone 为 PPLCNet_x1_0、使用 SSLD 预训练模型、使用 SSLD 预训练模型 + EDA 策略、使用 SSLD 预训练模型 + EDA 策略 + SKL-UGI 知识蒸馏策略训练得到的模型的相关指标。
 
 
-| 模型 | ma（%） | 延时（ms） | 存储（M） | 策略 |
+| 模型 | mA（%） | 延时（ms） | 存储（M） | 策略 |
 |-------|-----------|----------|---------------|---------------|
 | Res2Net200_vd_26w_4s  | 81.25 | 77.51  | 293 | 使用ImageNet预训练模型 |
-| SwinTransformer_tiny  | 80.17 | 89.51  | 107 | 使用ImageNet预训练模型 |
+| SwinTransformer_tiny  | 80.17 | 89.51  | 111 | 使用ImageNet预训练模型 |
 | MobileNetV3_small_x0_35  | 70.79 | 2.90  | 1.7 | 使用ImageNet预训练模型 |
-| PPLCNet_x1_0  | 76.31 | 2.01  | 6.6 | 使用ImageNet预训练模型 |
-| PPLCNet_x1_0  | 77.31 | 2.01  | 6.6 | 使用SSLD预训练模型 |
-| PPLCNet_x1_0  | 77.71 | 2.01  | 6.6 | 使用SSLD预训练模型+EDA策略|
-| <b>PPLCNet_x1_0<b>  | <b>78.59<b> | <b>2.01<b>  | <b>6.6<b> | 使用SSLD预训练模型+EDA策略+SKL-UGI知识蒸馏策略|
+| PPLCNet_x1_0  | 76.31 | 2.01  | 7.1 | 使用ImageNet预训练模型 |
+| PPLCNet_x1_0  | 77.31 | 2.01  | 7.1 | 使用SSLD预训练模型 |
+| PPLCNet_x1_0  | 77.71 | 2.01  | 7.1 | 使用SSLD预训练模型+EDA策略|
+| <b>PPLCNet_x1_0<b>  | <b>78.59<b> | <b>2.01<b>  | <b>7.1<b> | 使用SSLD预训练模型+EDA策略+SKL-UGI知识蒸馏策略|
 
 从表中可以看出，backbone 为 Res2Net200_vd_26w_4s 和 SwinTransformer_tiny 时精度较高，但是推理速度较慢。将 backbone 替换为轻量级模型 MobileNetV3_small_x0_35 后，速度可以大幅提升，但是精度也大幅下降。将 backbone 替换为 PPLCNet_x1_0 时，精度较 MobileNetV3_small_x0_35 高 5.5%，于此同时，速度更快。在此基础上，使用 SSLD 预训练模型后，在不改变推理速度的前提下，精度可以提升 1%，进一步地，当融合EDA策略后，精度可以再提升 0.4%，最后，在使用 SKL-UGI 知识蒸馏后，精度可以继续提升 0.88%。此时，PPLCNet_x1_0 的精度与 SwinTransformer_tiny 仅相差1.58%，但是速度快 44 倍。关于 PULC 的训练方法和推理部署方法将在下面详细介绍。
 
 **备注：**
 
+* 延时是基于 Intel(R) Xeon(R) Gold 6148 CPU @ 2.40GHz 测试得到，开启 MKLDNN 加速策略，线程数为10。
 * 关于PP-LCNet的介绍可以参考[PP-LCNet介绍](../models/PP-LCNet.md)，相关论文可以查阅[PP-LCNet paper](https://arxiv.org/abs/2109.15099)。
 
 
@@ -67,9 +68,9 @@
 ## 2. 模型快速体验
 
 <a name="2.1"></a>  
-    
+
 ### 2.1 安装 paddlepaddle
-    
+
 - 您的机器安装的是 CUDA9 或 CUDA10，请运行以下命令安装
 
 ```bash
@@ -81,23 +82,23 @@ python3 -m pip install paddlepaddle-gpu -i https://mirror.baidu.com/pypi/simple
 ```bash
 python3 -m pip install paddlepaddle -i https://mirror.baidu.com/pypi/simple
 ```
-    
+
 更多的版本需求，请参照[飞桨官网安装文档](https://www.paddlepaddle.org.cn/install/quick)中的说明进行操作。
-    
+
 <a name="2.2"></a>  
-    
+
 ### 2.2 安装 paddleclas
 
 使用如下命令快速安装 paddleclas
 
 ```  
 pip3 install paddleclas
-``` 
-    
+```
+
 <a name="2.3"></a>
 
 ### 2.3 预测
-    
+
 点击[这里](https://paddleclas.bj.bcebos.com/data/PULC/pulc_demo_imgs.zip)下载 demo 数据并解压，然后在终端中切换到相应目录。
 
 * 使用命令行快速预测
@@ -223,7 +224,7 @@ python3 -m paddle.distributed.launch \
         -c ./ppcls/configs/PULC/person_attribute/PPLCNet_x1_0.yaml
 ```
 
-验证集的最佳指标在 `90.07%` 左右（数据集较小，一般有0.3%左右的波动）。
+验证集的最佳指标在 `77.71%` 左右（数据集较小，一般有0.3%左右的波动）。
 
 
 <a name="3.4"></a>
@@ -313,7 +314,7 @@ python3 -m paddle.distributed.launch \
 
 ## 5. 超参搜索
 
-在 [3.2 节](#3.2)和 [4.1 节](#4.1)所使用的超参数是根据 PaddleClas 提供的 `SHAS 超参数搜索策略` 搜索得到的，如果希望在自己的数据集上得到更好的结果，可以参考[SHAS 超参数搜索策略](PULC_train.md#4-超参搜索)来获得更好的训练超参数。
+在 [3.2 节](#3.2)和 [4.1 节](#4.1)所使用的超参数是根据 PaddleClas 提供的 `超参数搜索策略` 搜索得到的，如果希望在自己的数据集上得到更好的结果，可以参考[超参数搜索策略](PULC_train.md#4-超参搜索)来获得更好的训练超参数。
 
 **备注：** 此部分内容是可选内容，搜索过程需要较长的时间，您可以根据自己的硬件情况来选择执行。如果没有更换数据集，可以忽略此节内容。
 

docs/zh_CN/PULC/PULC_person_exists.md

@@ -40,21 +40,21 @@
 
 ## 1. 模型和应用场景介绍
 
-该案例提供了用户使用 PaddleClas 的超轻量图像分类方案（PULC，Practical Ultra Lightweight Classification）快速构建轻量级、高精度、可落地的有人/无人的分类模型。该模型可以广泛应用于如监控场景、人员进出管控场景、海量数据过滤场景等。
+该案例提供了用户使用 PaddleClas 的超轻量图像分类方案（PULC，Practical Ultra Lightweight image Classification）快速构建轻量级、高精度、可落地的有人/无人的分类模型。该模型可以广泛应用于如监控场景、人员进出管控场景、海量数据过滤场景等。
 
 下表列出了判断图片中是否有人的二分类模型的相关指标，前两行展现了使用 SwinTranformer_tiny 和 MobileNetV3_small_x0_35 作为 backbone 训练得到的模型的相关指标，第三行至第六行依次展现了替换 backbone 为 PPLCNet_x1_0、使用 SSLD 预训练模型、使用 SSLD 预训练模型 + EDA 策略、使用 SSLD 预训练模型 + EDA 策略 + SKL-UGI 知识蒸馏策略训练得到的模型的相关指标。
 
 
 | 模型 | Tpr（%） | 延时（ms） | 存储（M） | 策略 |
 |-------|-----------|----------|---------------|---------------|
-| SwinTranformer_tiny  | 95.69 | 95.30  | 107 | 使用 ImageNet 预训练模型 |
-| MobileNetV3_small_x0_35  | 68.25 | 2.85  | 1.6 | 使用 ImageNet 预训练模型 |
-| PPLCNet_x1_0  | 89.57 | 2.12  | 6.5 | 使用 ImageNet 预训练模型 |
-| PPLCNet_x1_0  | 92.10 | 2.12  | 6.5 | 使用 SSLD 预训练模型 |
-| PPLCNet_x1_0  | 93.43 | 2.12  | 6.5 | 使用 SSLD 预训练模型+EDA 策略|
-| <b>PPLCNet_x1_0<b>  | <b>95.60<b> | <b>2.12<b>  | <b>6.5<b> | 使用 SSLD 预训练模型+EDA 策略+SKL-UGI 知识蒸馏策略|
+| SwinTranformer_tiny  | 95.69 | 95.30  | 111 | 使用 ImageNet 预训练模型 |
+| MobileNetV3_small_x0_35  | 68.25 | 2.85  | 2.6 | 使用 ImageNet 预训练模型 |
+| PPLCNet_x1_0  | 89.57 | 2.12  | 7.0 | 使用 ImageNet 预训练模型 |
+| PPLCNet_x1_0  | 92.10 | 2.12  | 7.0 | 使用 SSLD 预训练模型 |
+| PPLCNet_x1_0  | 93.43 | 2.12  | 7.0 | 使用 SSLD 预训练模型+EDA 策略|
+| <b>PPLCNet_x1_0<b>  | <b>96.23<b> | <b>2.12<b>  | <b>7.0<b> | 使用 SSLD 预训练模型+EDA 策略+SKL-UGI 知识蒸馏策略|
 
-从表中可以看出，backbone 为 SwinTranformer_tiny 时精度较高，但是推理速度较慢。将 backboone 替换为轻量级模型 MobileNetV3_small_x0_35 后，速度可以大幅提升，但是会导致精度大幅下降。将 backbone 替换为速度更快的 PPLCNet_x1_0 时，精度较 MobileNetV3_small_x0_35 高 20 多个百分点，与此同时速度依旧可以快 20% 以上。在此基础上，使用 SSLD 预训练模型后，在不改变推理速度的前提下，精度可以提升约 2.6 个百分点，进一步地，当融合EDA策略后，精度可以再提升 1.3 个百分点，最后，在使用 SKL-UGI 知识蒸馏后，精度可以继续提升 2.2 个百分点。此时，PPLCNet_x1_0 达到了 SwinTranformer_tiny 模型的精度，但是速度快 40 多倍。关于 PULC 的训练方法和推理部署方法将在下面详细介绍。
+从表中可以看出，backbone 为 SwinTranformer_tiny 时精度较高，但是推理速度较慢。将 backbone 替换为轻量级模型 MobileNetV3_small_x0_35 后，速度可以大幅提升，但是会导致精度大幅下降。将 backbone 替换为速度更快的 PPLCNet_x1_0 时，精度较 MobileNetV3_small_x0_35 高 20 多个百分点，与此同时速度依旧可以快 20% 以上。在此基础上，使用 SSLD 预训练模型后，在不改变推理速度的前提下，精度可以提升约 2.6 个百分点，进一步地，当融合EDA策略后，精度可以再提升 1.3 个百分点，最后，在使用 SKL-UGI 知识蒸馏后，精度可以继续提升 2.8 个百分点。此时，PPLCNet_x1_0 达到了 SwinTranformer_tiny 模型的精度，但是速度快 40 多倍。关于 PULC 的训练方法和推理部署方法将在下面详细介绍。
 
 **备注：**
 
@@ -67,9 +67,9 @@
 ## 2. 模型快速体验
 
 <a name="2.1"></a>  
-    
+
 ### 2.1 安装 paddlepaddle
-    
+
 - 您的机器安装的是 CUDA9 或 CUDA10，请运行以下命令安装
 
 ```bash
@@ -81,23 +81,23 @@ python3 -m pip install paddlepaddle-gpu -i https://mirror.baidu.com/pypi/simple
 ```bash
 python3 -m pip install paddlepaddle -i https://mirror.baidu.com/pypi/simple
 ```
-    
+
 更多的版本需求，请参照[飞桨官网安装文档](https://www.paddlepaddle.org.cn/install/quick)中的说明进行操作。
-    
+
 <a name="2.2"></a>  
-    
+
 ### 2.2 安装 paddleclas
 
 使用如下命令快速安装 paddleclas
 
 ```  
 pip3 install paddleclas
-``` 
-    
+```
+
 <a name="2.3"></a>
 
 ### 2.3 预测
-    
+
 点击[这里](https://paddleclas.bj.bcebos.com/data/PULC/pulc_demo_imgs.zip)下载 demo 数据并解压，然后在终端中切换到相应目录。
 
 * 使用命令行快速预测
@@ -328,7 +328,7 @@ python3 -m paddle.distributed.launch \
 
 ## 5. 超参搜索
 
-在 [3.3 节](#3.3)和 [4.1 节](#4.1)所使用的超参数是根据 PaddleClas 提供的 `SHAS 超参数搜索策略` 搜索得到的，如果希望在自己的数据集上得到更好的结果，可以参考[SHAS 超参数搜索策略](PULC_train.md#4-超参搜索)来获得更好的训练超参数。
+在 [3.3 节](#3.3)和 [4.1 节](#4.1)所使用的超参数是根据 PaddleClas 提供的 `超参数搜索策略` 搜索得到的，如果希望在自己的数据集上得到更好的结果，可以参考[超参数搜索策略](PULC_train.md#4-超参搜索)来获得更好的训练超参数。
 
 **备注：** 此部分内容是可选内容，搜索过程需要较长的时间，您可以根据自己的硬件情况来选择执行。如果没有更换数据集，可以忽略此节内容。
 

docs/zh_CN/PULC/PULC_text_image_orientation.md

@@ -36,21 +36,21 @@
 
 ## 1. 模型和应用场景介绍
 
-在诸如文档扫描、证照拍摄等过程中，有时为了拍摄更清晰，会将拍摄设备进行旋转，导致得到的图片也是不同方向的。此时，标准的OCR流程无法很好地应对这些数据。利用图像分类技术，可以预先判断含文字图像的方向，并将其进行方向调整，从而提高OCR处理的准确性。该案例提供了用户使用 PaddleClas 的超轻量图像分类方案（PULC，Practical Ultra Lightweight Classification）快速构建轻量级、高精度、可落地的含文字图像方向的分类模型。该模型可以广泛应用于金融、政务等行业的旋转图片的OCR处理场景中。
+在诸如文档扫描、证照拍摄等过程中，有时为了拍摄更清晰，会将拍摄设备进行旋转，导致得到的图片也是不同方向的。此时，标准的OCR流程无法很好地应对这些数据。利用图像分类技术，可以预先判断含文字图像的方向，并将其进行方向调整，从而提高OCR处理的准确性。该案例提供了用户使用 PaddleClas 的超轻量图像分类方案（PULC，Practical Ultra Lightweight image Classification）快速构建轻量级、高精度、可落地的含文字图像方向的分类模型。该模型可以广泛应用于金融、政务等行业的旋转图片的OCR处理场景中。
 
-下表列出了判断含文字图像方向分类模型的相关指标，前两行展现了使用 SwinTranformer_tiny 和 MobileNetV3_small_x0_35 作为 backbone 训练得到的模型的相关指标，第三行至第五行依次展现了替换 backbone 为 PPLCNet_x1_0、使用 SSLD 预训练模型、使用 SHAS 超参数搜索策略训练得到的模型的相关指标。
+下表列出了判断含文字图像方向分类模型的相关指标，前两行展现了使用 SwinTranformer_tiny 和 MobileNetV3_small_x0_35 作为 backbone 训练得到的模型的相关指标，第三行至第五行依次展现了替换 backbone 为 PPLCNet_x1_0、使用 SSLD 预训练模型、使用EDA策略训练得到的模型的相关指标。
 
-| 模型                    | 精度（%） | 延时（ms） | 存储（M） | 策略                                  |
-| ----------------------- | --------- | ---------- | --------- | ------------------------------------- |
-| SwinTranformer_tiny     | 99.12     | 89.65      | 107       | 使用ImageNet预训练模型                |
-| MobileNetV3_small_x0_35 | 83.61     | 2.95       | 17        | 使用ImageNet预训练模型                |
-| PPLCNet_x1_0            | 97.85     | 2.16       | 6.5       | 使用ImageNet预训练模型                |
-| PPLCNet_x1_0            | 98.02     | 2.16       | 6.5       | 使用SSLD预训练模型                    |
-| **PPLCNet_x1_0**        | **99.06** | **2.16**   | **6.5**   | 使用SSLD预训练模型+SHAS超参数搜索策略 |
+| 模型                    | 精度（%） | 延时（ms） | 存储（M） | 策略                       |
+| ----------------------- | --------- | ---------- | --------- | -------------------------- |
+| SwinTranformer_tiny     | 99.12     | 89.65      | 111       | 使用ImageNet预训练模型     |
+| MobileNetV3_small_x0_35 | 83.61     | 2.95       | 2.6        | 使用ImageNet预训练模型     |
+| PPLCNet_x1_0            | 97.85     | 2.16       | 7.1       | 使用ImageNet预训练模型     |
+| PPLCNet_x1_0            | 99.02     | 2.16       | 7.1       | 使用SSLD预训练模型         |
+| **PPLCNet_x1_0**        | **99.06** | **2.16**   | **7.1**   | 使用SSLD预训练模型+EDA策略 |
 
-从表中可以看出，backbone 为 SwinTranformer_tiny 时精度比较高，但是推理速度较慢。将 backboone 替换为轻量级模型 MobileNetV3_small_x0_35 后，速度提升明显，但精度有了大幅下降。将 backbone 替换为 PPLCNet_x1_0 时，速度略为提升，同时精度较 MobileNetV3_small_x0_35 高了 14.24 个百分点。在此基础上，使用 SSLD 预训练模型后，在不改变推理速度的前提下，精度可以提升 0.17 个百分点，进一步地，当使用SHAS超参数搜索策略搜索最优超参数后，精度可以再提升 1.04 个百分点。此时，PPLCNet_x1_0 与 SwinTranformer_tiny 的精度差别不大，但是速度明显变快。关于 PULC 的训练方法和推理部署方法将在下面详细介绍。
+从表中可以看出，backbone 为 SwinTranformer_tiny 时精度比较高，但是推理速度较慢。将 backbone 替换为轻量级模型 MobileNetV3_small_x0_35 后，速度提升明显，但精度有了大幅下降。将 backbone 替换为 PPLCNet_x1_0 时，速度略为提升，同时精度较 MobileNetV3_small_x0_35 高了 14.24 个百分点。在此基础上，使用 SSLD 预训练模型后，在不改变推理速度的前提下，精度可以提升 1.17 个百分点，进一步地使用 EDA 策略后，精度可以再提升 0.04 个百分点。此时，PPLCNet_x1_0 与 SwinTranformer_tiny 的精度差别不大，但是速度明显变快。关于 PULC 的训练方法和推理部署方法将在下面详细介绍。
 
-**备注：** 
+**备注：**
 
 * 关于PP-LCNet的介绍可以参考[PP-LCNet介绍](../models/PP-LCNet.md)，相关论文可以查阅[PP-LCNet paper](https://arxiv.org/abs/2109.15099)。
 
@@ -59,9 +59,9 @@
 ## 2. 模型快速体验
 
 <a name="2.1"></a>  
-    
+
 ### 2.1 安装 paddlepaddle
-    
+
 - 您的机器安装的是 CUDA9 或 CUDA10，请运行以下命令安装
 
 ```bash
@@ -73,23 +73,23 @@ python3 -m pip install paddlepaddle-gpu -i https://mirror.baidu.com/pypi/simple
 ```bash
 python3 -m pip install paddlepaddle -i https://mirror.baidu.com/pypi/simple
 ```
-    
+
 更多的版本需求，请参照[飞桨官网安装文档](https://www.paddlepaddle.org.cn/install/quick)中的说明进行操作。
-    
+
 <a name="2.2"></a>  
-    
+
 ### 2.2 安装 paddleclas
 
 使用如下命令快速安装 paddleclas
 
 ```  
 pip3 install paddleclas
-``` 
-    
+```
+
 <a name="2.3"></a>
 
 ### 2.3 预测
-    
+
 点击[这里](https://paddleclas.bj.bcebos.com/data/PULC/pulc_demo_imgs.zip)下载 demo 数据并解压，然后在终端中切换到相应目录。
 
 * 使用命令行快速预测
@@ -319,7 +319,7 @@ python3 -m paddle.distributed.launch \
 
 ## 5. 超参搜索
 
-在 [3.2 节](#3.2)和 [4.1 节](#4.1)所使用的超参数是根据 PaddleClas 提供的 `SHAS 超参数搜索策略` 搜索得到的，如果希望在自己的数据集上得到更好的结果，可以参考[SHAS 超参数搜索策略](PULC_train.md#4-超参搜索)来获得更好的训练超参数。
+在 [3.2 节](#3.2)和 [4.1 节](#4.1)所使用的超参数是根据 PaddleClas 提供的 `超参数搜索策略` 搜索得到的，如果希望在自己的数据集上得到更好的结果，可以参考[超参数搜索策略](PULC_train.md#4-超参搜索)来获得更好的训练超参数。
 
 **备注：** 此部分内容是可选内容，搜索过程需要较长的时间，您可以根据自己的硬件情况来选择执行。如果没有更换数据集，可以忽略此节内容。