yujunjun/PaddleOCR
1
0
Fork 0
mirror of https://github.com/PaddlePaddle/PaddleOCR.git synced 2025-09-25 16:15:12 +00:00
Code Issues Packages Projects Releases Wiki Activity

docs: update paddle2onnx documentations (#14144)

Browse Source
This commit is contained in:
Wang Xin 2024-11-01 23:17:59 +08:00 committed by GitHub
parent d1bc41661f
commit d3d7e85883
No known key found for this signature in database
GPG Key ID: B5690EEEBB952194
4 changed files with 226 additions and 115 deletions
Show Stats Download Patch File Download Diff File Expand all files Collapse all files

BIN
docs/ppocr/infer_deploy/images/img_12.jpg Normal file
View File

Binary file not shown.
Side by Side

After

Width:  |  Height:  |  Size: 722 KiB

BIN
docs/ppocr/infer_deploy/images/img_12_result.jpg Normal file
View File

Binary file not shown.
Side by Side

After

Width:  |  Height:  |  Size: 561 KiB

207
docs/ppocr/infer_deploy/paddle2onnx.en.md
View File

@ -1,139 +1,196 @@
---
typora-copy-images-to: images
comments: true
---
# Paddle2ONNX model transformation and prediction
# Paddle2ONNX Model Conversion and Prediction
This chapter describes how the PaddleOCR model is converted into an ONNX model and predicted based on the ONNXRuntime engine.
This chapter introduces how to convert PaddleOCR models to ONNX models and perform predictions based on the ONNXRuntime engine.
## 1. Environment preparation
## 1. Environment Setup
Need to prepare PaddleOCR, Paddle2ONNX model conversion environment, and ONNXRuntime prediction environment
You need to prepare the environments for PaddleOCR, Paddle2ONNX model conversion, and ONNXRuntime prediction.
### PaddleOCR
Clone the PaddleOCR repository, use the release/2.6 branch, and install it.
Clone the PaddleOCR repository, use the main branch, and install it. Since the PaddleOCR repository is relatively large and cloning via `git clone` can be slow, this tutorial has already downloaded it.
```bash linenums="1"
git clone -b release/2.6 https://github.com/PaddlePaddle/PaddleOCR.git
cd PaddleOCR && python3.7 setup.py install
git clone -b main https://github.com/PaddlePaddle/PaddleOCR.git
cd PaddleOCR && python3 -m pip install -e .
```
### Paddle2ONNX
Paddle2ONNX supports converting the PaddlePaddle model format to the ONNX model format. The operator currently supports exporting ONNX Opset 9~11 stably, and some Paddle operators support lower ONNX Opset conversion.
For more details, please refer to [Paddle2ONNX](https://github.com/PaddlePaddle/Paddle2ONNX/blob/develop/README_en.md)
Paddle2ONNX supports converting models in the PaddlePaddle format to the ONNX format. Operators currently stably support exporting ONNX Opset versions 9~18, and some Paddle operators support conversion to lower ONNX Opsets. For more details, please refer to [Paddle2ONNX](https://github.com/PaddlePaddle/Paddle2ONNX/blob/develop/README_en.md).
- install Paddle2ONNX
- Install Paddle2ONNX
```bash linenums="1"
python3.7 -m pip install paddle2onnx
python3 -m pip install paddle2onnx
```
- install ONNXRuntime
- Install ONNXRuntime
```bash linenums="1"
# It is recommended to install version 1.9.0, and the version number can be changed according to the environment
python3.7 -m pip install onnxruntime==1.9.0
python3 -m pip install onnxruntime
```
## 2. Model conversion
## 2. Model Conversion
### Paddle model download
### Download Paddle Models
There are two ways to obtain the Paddle model: Download the prediction model provided by PaddleOCR in [model_list](../model_list.en.md);
There are two ways to obtain Paddle static graph models: download the prediction models provided by PaddleOCR in the [model list](../model_list.en.md); or refer to the [Model Export Instructions](https://paddlepaddle.github.io/PaddleOCR/latest/ppocr/infer_deploy/python_infer.html#inference) to convert trained weights into inference models.
Take the PP-OCRv3 detection, recognition, and classification model as an example:
Using the PP-OCR series English detection, recognition, and classification models as examples:
```bash linenums="1"
wget -nc -P ./inference https://paddleocr.bj.bcebos.com/PP-OCRv3/english/en_PP-OCRv3_det_infer.tar
cd ./inference && tar xf en_PP-OCRv3_det_infer.tar && cd ..
=== "PP-OCRv3"
wget -nc -P ./inference https://paddleocr.bj.bcebos.com/PP-OCRv3/english/en_PP-OCRv3_rec_infer.tar
cd ./inference && tar xf en_PP-OCRv3_rec_infer.tar && cd ..
```bash linenums="1"
wget -nc -P ./inference https://paddleocr.bj.bcebos.com/PP-OCRv3/english/en_PP-OCRv3_det_infer.tar
cd ./inference && tar xf en_PP-OCRv3_det_infer.tar && cd ..
wget -nc -P ./inference https://paddleocr.bj.bcebos.com/dygraph_v2.0/ch/ch_ppocr_mobile_v2.0_cls_infer.tar
cd ./inference && tar xf ch_ppocr_mobile_v2.0_cls_infer.tar && cd ..
```
wget -nc -P ./inference https://paddleocr.bj.bcebos.com/PP-OCRv3/english/en_PP-OCRv3_rec_infer.tar
cd ./inference && tar xf en_PP-OCRv3_rec_infer.tar && cd ..
### Convert model
wget -nc -P ./inference https://paddleocr.bj.bcebos.com/dygraph_v2.0/ch/ch_ppocr_mobile_v2.0_cls_infer.tar
cd ./inference && tar xf ch_ppocr_mobile_v2.0_cls_infer.tar && cd ..
```
Convert Paddle inference model to ONNX model format using Paddle2ONNX:
=== "PP-OCRv4"
```bash linenums="1"
paddle2onnx --model_dir ./inference/en_PP-OCRv3_det_infer \
--model_filename inference.pdmodel \
--params_filename inference.pdiparams \
--save_file ./inference/det_onnx/model.onnx \
--opset_version 10 \
--input_shape_dict="{'x':[-1,3,-1,-1]}" \
--enable_onnx_checker True
```bash linenums="1"
wget -nc -P ./inference https://paddleocr.bj.bcebos.com/PP-OCRv3/english/en_PP-OCRv3_det_infer.tar
cd ./inference && tar xf en_PP-OCRv3_det_infer.tar && cd ..
paddle2onnx --model_dir ./inference/en_PP-OCRv3_rec_infer \
--model_filename inference.pdmodel \
--params_filename inference.pdiparams \
--save_file ./inference/rec_onnx/model.onnx \
--opset_version 10 \
--input_shape_dict="{'x':[-1,3,-1,-1]}" \
--enable_onnx_checker True
wget -nc -P ./inference https://paddleocr.bj.bcebos.com/PP-OCRv4/english/en_PP-OCRv4_rec_infer.tar
cd ./inference && tar xf en_PP-OCRv4_rec_infer.tar && cd ..
paddle2onnx --model_dir ./inference/ch_ppocr_mobile_v2.0_cls_infer \
--model_filename ch_ppocr_mobile_v2.0_cls_infer/inference.pdmodel \
--params_filename ch_ppocr_mobile_v2.0_cls_infer/inference.pdiparams \
--save_file ./inferencecls_onnx/model.onnx \
--opset_version 10 \
--input_shape_dict="{'x':[-1,3,-1,-1]}" \
--enable_onnx_checker True
```
wget -nc -P ./inference https://paddleocr.bj.bcebos.com/dygraph_v2.0/ch/ch_ppocr_mobile_v2.0_cls_infer.tar
cd ./inference && tar xf ch_ppocr_mobile_v2.0_cls_infer.tar && cd ..
```
After execution, the ONNX model will be saved in `./inference/det_onnx/`, `./inference/rec_onnx/`, `./inference/cls_onnx/` paths respectively
### Model Conversion
- Note: For the OCR model, the conversion process must be in the form of dynamic shape, that is, add the option --input_shape_dict="{'x': [-1, 3, -1, -1]}", otherwise the prediction result may be the same as Predicting directly with Paddle is slightly different.
Use Paddle2ONNX to convert Paddle static graph models to the ONNX model format:
In addition, the following models do not currently support conversion to ONNX models: NRTR, SAR, RARE, SRN.
=== "PP-OCRv3"
If you have optimization needs for the exported ONNX model, we recommend using `onnxslim`.
```bash linenums="1"
paddle2onnx --model_dir ./inference/en_PP-OCRv3_det_infer \
--model_filename inference.pdmodel \
--params_filename inference.pdiparams \
--save_file ./inference/det_onnx/model.onnx \
--opset_version 11 \
--enable_onnx_checker True
paddle2onnx --model_dir ./inference/en_PP-OCRv3_rec_infer \
--model_filename inference.pdmodel \
--params_filename inference.pdiparams \
--save_file ./inference/rec_onnx/model.onnx \
--opset_version 11 \
--enable_onnx_checker True
paddle2onnx --model_dir ./inference/ch_ppocr_mobile_v2.0_cls_infer \
--model_filename inference.pdmodel \
--params_filename inference.pdiparams \
--save_file ./inference/cls_onnx/model.onnx \
--opset_version 11 \
--enable_onnx_checker True
```
=== "PP-OCRv4"
```bash linenums="1"
paddle2onnx --model_dir ./inference/en_PP-OCRv3_det_infer \
--model_filename inference.pdmodel \
--params_filename inference.pdiparams \
--save_file ./inference/det_onnx/model.onnx \
--opset_version 11 \
--enable_onnx_checker True
paddle2onnx --model_dir ./inference/en_PP-OCRv4_rec_infer \
--model_filename inference.pdmodel \
--params_filename inference.pdiparams \
--save_file ./inference/rec_onnx/model.onnx \
--opset_version 11 \
--enable_onnx_checker True
paddle2onnx --model_dir ./inference/ch_ppocr_mobile_v2.0_cls_infer \
--model_filename inference.pdmodel \
--params_filename inference.pdiparams \
--save_file ./inference/cls_onnx/model.onnx \
--opset_version 11 \
--enable_onnx_checker True
```
After execution, the ONNX models will be saved respectively under `./inference/det_onnx/`, `./inference/rec_onnx/`, and `./inference/cls_onnx/`.
- **Note**: For OCR models, dynamic shapes must be used during conversion; otherwise, the prediction results may slightly differ from directly using Paddle for prediction. Additionally, the following models currently do not support conversion to ONNX models: NRTR, SAR, RARE, SRN.
- **Note**: After [Paddle2ONNX version v1.2.3](https://github.com/PaddlePaddle/Paddle2ONNX/releases/tag/v1.2.3), dynamic shapes are supported by default, i.e., `float32[p2o.DynamicDimension.0,3,p2o.DynamicDimension.1,p2o.DynamicDimension.2]`. The option `--input_shape_dict` has been deprecated. If you need to adjust shapes, you can use the following command to adjust the input shape of the Paddle model.
```bash linenums="1"
python3 -m paddle2onnx.optimize --input_model inference/det_onnx/model.onnx \
--output_model inference/det_onnx/model.onnx \
--input_shape_dict "{'x': [-1,3,-1,-1]}"
```
If you have optimization requirements for the exported ONNX model, it is recommended to use [OnnxSlim](https://github.com/inisis/OnnxSlim) to optimize the model:
```bash linenums="1"
pip install onnxslim
onnxslim model.onnx slim.onnx
```
## 3. prediction
## 3. Inference and Prediction
Take the English OCR model as an example, use **ONNXRuntime** to predict and execute the following commands:
Using the Chinese OCR model as an example, you can perform prediction using ONNXRuntime by executing the following command:
```bash linenums="1"
python3.7 tools/infer/predict_system.py --use_gpu=False --use_onnx=True \
python3 tools/infer/predict_system.py --use_gpu=False --use_onnx=True \
--det_model_dir=./inference/det_onnx/model.onnx \
--rec_model_dir=./inference/rec_onnx/model.onnx \
--cls_model_dir=./inference/cls_onnx/model.onnx \
--image_dir=doc/imgs_en/img_12.jpg \
--image_dir=./docs/ppocr/infer_deploy/images/img_12.jpg \
--rec_char_dict_path=ppocr/utils/en_dict.txt
```
Taking the English OCR model as an example, use **Paddle Inference** to predict and execute the following commands:
Taking the English OCR model as an example, you can perform prediction using Paddle Inference by executing the following command:
```bash linenums="1"
python3.7 tools/infer/predict_system.py --use_gpu=False \
--cls_model_dir=./inference/ch_ppocr_mobile_v2.0_cls_infer \
--rec_model_dir=./inference/en_PP-OCRv3_rec_infer \
--det_model_dir=./inference/en_PP-OCRv3_det_infer \
--image_dir=doc/imgs_en/img_12.jpg \
--rec_char_dict_path=ppocr/utils/en_dict.txt
```
=== "PP-OCRv3"
After executing the command, the predicted identification information will be printed out in the terminal, and the visualization results will be saved under `./inference_results/`.
```bash linenums="1"
python3 tools/infer/predict_system.py --use_gpu=False \
--cls_model_dir=./inference/ch_ppocr_mobile_v2.0_cls_infer \
--rec_model_dir=./inference/en_PP-OCRv3_rec_infer \
--det_model_dir=./inference/en_PP-OCRv3_det_infer \
--image_dir=./docs/ppocr/infer_deploy/images/img_12.jpg\
--rec_char_dict_path=ppocr/utils/en_dict.txt
```
ONNXRuntime result：
=== "PP-OCRv4"
![](./images/lite_demo_onnx.png)
```bash linenums="1"
python3 tools/infer/predict_system.py --use_gpu=False \
--cls_model_dir=./inference/ch_ppocr_mobile_v2.0_cls_infer \
--rec_model_dir=./inference/en_PP-OCRv4_rec_infer \
--det_model_dir=./inference/en_PP-OCRv3_det_infer \
--image_dir=./docs/ppocr/infer_deploy/images/img_12.jpg \
--rec_char_dict_path=ppocr/utils/en_dict.txt
```
Paddle Inference result：
After executing the command, the terminal will print out the predicted recognition information, and the visualization results will be saved under `./inference_results/`.
![](./images/lite_demo_paddle.png)
**ONNXRuntime Execution Result:**
Using ONNXRuntime to predict, terminal output:
![](./images/img_12_result.jpg)
**Paddle Inference Execution Result:**
![](./images/img_12_result.jpg)
Using ONNXRuntime for prediction, terminal output:
```bash linenums="1"
[2022/10/10 12:06:28] ppocr DEBUG: dt_boxes num : 11, elapse : 0.3568880558013916
@ -154,7 +211,7 @@ Using ONNXRuntime to predict, terminal output:
[2022/10/10 12:06:31] ppocr INFO: The predict total time is 3.2482550144195557
```
Using Paddle Inference to predict, terminal output:
Using Paddle Inference for prediction, terminal output:
```bash linenums="1"
[2022/10/10 12:06:28] ppocr DEBUG: dt_boxes num : 11, elapse : 0.3568880558013916

134
docs/ppocr/infer_deploy/paddle2onnx.md
View File

@ -41,47 +41,89 @@ Paddle2ONNX 支持将 PaddlePaddle 模型格式转化到 ONNX 模型格式，算
### Paddle 模型下载
有两种方式获取Paddle静态图模型：在 [model_list](../model_list.md) 中下载PaddleOCR提供的预测模型；
有两种方式获取Paddle静态图模型：在 [model_list](../model_list.md) 中下载PaddleOCR提供的预测模型；参考[模型导出说明](https://paddlepaddle.github.io/PaddleOCR/latest/ppocr/infer_deploy/python_infer.html#inference)把训练好的权重转为推理模型。
以 PP-OCRv3 中文检测、识别、分类模型为例：
以 PP-OCR 系列中文检测、识别、分类模型为例：
```bash linenums="1"
wget -nc -P ./inference https://paddleocr.bj.bcebos.com/PP-OCRv3/chinese/ch_PP-OCRv3_det_infer.tar
cd ./inference && tar xf ch_PP-OCRv3_det_infer.tar && cd ..
=== "PP-OCRv3"
wget -nc -P ./inference https://paddleocr.bj.bcebos.com/PP-OCRv3/chinese/ch_PP-OCRv3_rec_infer.tar
cd ./inference && tar xf ch_PP-OCRv3_rec_infer.tar && cd ..
```bash linenums="1"
wget -nc -P ./inference https://paddleocr.bj.bcebos.com/PP-OCRv3/chinese/ch_PP-OCRv3_det_infer.tar
cd ./inference && tar xf ch_PP-OCRv3_det_infer.tar && cd ..
wget -nc -P ./inference https://paddleocr.bj.bcebos.com/dygraph_v2.0/ch/ch_ppocr_mobile_v2.0_cls_infer.tar
cd ./inference && tar xf ch_ppocr_mobile_v2.0_cls_infer.tar && cd ..
```
wget -nc -P ./inference https://paddleocr.bj.bcebos.com/PP-OCRv3/chinese/ch_PP-OCRv3_rec_infer.tar
cd ./inference && tar xf ch_PP-OCRv3_rec_infer.tar && cd ..
wget -nc -P ./inference https://paddleocr.bj.bcebos.com/dygraph_v2.0/ch/ch_ppocr_mobile_v2.0_cls_infer.tar
cd ./inference && tar xf ch_ppocr_mobile_v2.0_cls_infer.tar && cd ..
```
=== "PP-OCRv4"
```bash linenums="1"
wget -nc -P ./inference https://paddleocr.bj.bcebos.com/PP-OCRv4/chinese/ch_PP-OCRv4_det_infer.tar
cd ./inference && tar xf ch_PP-OCRv4_det_infer.tar && cd ..
wget -nc -P ./inference https://paddleocr.bj.bcebos.com/PP-OCRv4/chinese/ch_PP-OCRv4_rec_infer.tar
cd ./inference && tar xf ch_PP-OCRv4_rec_infer.tar && cd ..
wget -nc -P ./inference https://paddleocr.bj.bcebos.com/dygraph_v2.0/ch/ch_ppocr_mobile_v2.0_cls_infer.tar
cd ./inference && tar xf ch_ppocr_mobile_v2.0_cls_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：
```bash linenums="1"
paddle2onnx --model_dir ./inference/ch_PP-OCRv3_det_infer \
--model_filename inference.pdmodel \
--params_filename inference.pdiparams \
--save_file ./inference/det_onnx/model.onnx \
--opset_version 11 \
--enable_onnx_checker True
=== "PP-OCRv3"
paddle2onnx --model_dir ./inference/ch_PP-OCRv3_rec_infer \
--model_filename inference.pdmodel \
--params_filename inference.pdiparams \
--save_file ./inference/rec_onnx/model.onnx \
--opset_version 11 \
--enable_onnx_checker True
```bash linenums="1"
paddle2onnx --model_dir ./inference/ch_PP-OCRv3_det_infer \
--model_filename inference.pdmodel \
--params_filename inference.pdiparams \
--save_file ./inference/det_onnx/model.onnx \
--opset_version 11 \
--enable_onnx_checker True
paddle2onnx --model_dir ./inference/ch_ppocr_mobile_v2.0_cls_infer \
--model_filename inference.pdmodel \
--params_filename inference.pdiparams \
--save_file ./inference/cls_onnx/model.onnx \
--opset_version 11 \
--enable_onnx_checker True
```
paddle2onnx --model_dir ./inference/ch_PP-OCRv3_rec_infer \
--model_filename inference.pdmodel \
--params_filename inference.pdiparams \
--save_file ./inference/rec_onnx/model.onnx \
--opset_version 11 \
--enable_onnx_checker True
paddle2onnx --model_dir ./inference/ch_ppocr_mobile_v2.0_cls_infer \
--model_filename inference.pdmodel \
--params_filename inference.pdiparams \
--save_file ./inference/cls_onnx/model.onnx \
--opset_version 11 \
--enable_onnx_checker True
```
=== "PP-OCRv4"
```bash linenums="1"
paddle2onnx --model_dir ./inference/ch_PP-OCRv4_det_infer \
--model_filename inference.pdmodel \
--params_filename inference.pdiparams \
--save_file ./inference/det_onnx/model.onnx \
--opset_version 11 \
--enable_onnx_checker True
paddle2onnx --model_dir ./inference/ch_PP-OCRv4_rec_infer \
--model_filename inference.pdmodel \
--params_filename inference.pdiparams \
--save_file ./inference/rec_onnx/model.onnx \
--opset_version 11 \
--enable_onnx_checker True
paddle2onnx --model_dir ./inference/ch_ppocr_mobile_v2.0_cls_infer \
--model_filename inference.pdmodel \
--params_filename inference.pdiparams \
--save_file ./inference/cls_onnx/model.onnx \
--opset_version 11 \
--enable_onnx_checker True
```
执行完毕后，ONNX 模型会被分别保存在 `./inference/det_onnx/`，`./inference/rec_onnx/`，`./inference/cls_onnx/`路径下
@ -89,7 +131,7 @@ paddle2onnx --model_dir ./inference/ch_ppocr_mobile_v2.0_cls_infer \
另外，以下几个模型暂不支持转换为 ONNX 模型：
NRTR、SAR、RARE、SRN
- 注意：[当前Paddle2ONNX版本(v1.2.3)](https://github.com/PaddlePaddle/Paddle2ONNX/releases/tag/v1.2.3)现已默认支持动态shape，即 `float32[p2o.DynamicDimension.0,3,p2o.DynamicDimension.1,p2o.DynamicDimension.2]`，选项 `--input_shape_dict` 已废弃。如果有shape调整需求可使用如下命令进行Paddle模型输入shape调整。
- 注意：[Paddle2ONNX 版本 v1.2.3](https://github.com/PaddlePaddle/Paddle2ONNX/releases/tag/v1.2.3)后已默认支持动态shape，即 `float32[p2o.DynamicDimension.0,3,p2o.DynamicDimension.1,p2o.DynamicDimension.2]`，选项 `--input_shape_dict` 已废弃。如果有shape调整需求可使用如下命令进行Paddle模型输入shape调整。
```bash linenums="1"
python3 -m paddle2onnx.optimize --input_model inference/det_onnx/model.onnx \
@ -97,7 +139,7 @@ paddle2onnx --model_dir ./inference/ch_ppocr_mobile_v2.0_cls_infer \
--input_shape_dict "{'x': [-1,3,-1,-1]}"
```
如你对导出的 ONNX 模型有优化的需求，推荐使用 `onnxslim` 对模型进行优化:
如你对导出的 ONNX 模型有优化的需求，推荐使用 [onnxslim](https://github.com/inisis/OnnxSlim) 对模型进行优化:
```bash linenums="1"
pip install onnxslim
@ -113,18 +155,30 @@ python3 tools/infer/predict_system.py --use_gpu=False --use_onnx=True \
--det_model_dir=./inference/det_onnx/model.onnx \
--rec_model_dir=./inference/rec_onnx/model.onnx \
--cls_model_dir=./inference/cls_onnx/model.onnx \
--image_dir=./deploy/lite/imgs/lite_demo.png
--image_dir=./docs/ppocr/infer_deploy/images/lite_demo.png
```
以中文OCR模型为例，使用 Paddle Inference 预测可执行如下命令：
```bash linenums="1"
python3 tools/infer/predict_system.py --use_gpu=False \
--cls_model_dir=./inference/ch_ppocr_mobile_v2.0_cls_infer \
--rec_model_dir=./inference/ch_PP-OCRv3_rec_infer \
--det_model_dir=./inference/ch_PP-OCRv3_det_infer \
--image_dir=./deploy/lite/imgs/lite_demo.png
```
=== "PP-OCRv3"
```bash linenums="1"
python3 tools/infer/predict_system.py --use_gpu=False \
--cls_model_dir=./inference/ch_ppocr_mobile_v2.0_cls_infer \
--rec_model_dir=./inference/ch_PP-OCRv3_rec_infer \
--det_model_dir=./inference/ch_PP-OCRv3_det_infer \
--image_dir=./docs/ppocr/infer_deploy/images/lite_demo.png
```
=== "PP-OCRv4"
```bash linenums="1"
python3 tools/infer/predict_system.py --use_gpu=False \
--cls_model_dir=./inference/ch_ppocr_mobile_v2.0_cls_infer \
--rec_model_dir=./inference/ch_PP-OCRv4_rec_infer \
--det_model_dir=./inference/ch_PP-OCRv4_det_infer \
--image_dir=./docs/ppocr/infer_deploy/images/lite_demo.png
```
执行命令后在终端会打印出预测的识别信息，并在 `./inference_results/` 下保存可视化结果。