chore: update docs

Author: SWHL

docs/blog/posts/about_model/adapt_PP-OCRv5_mobile_det.md

@@ -1,5 +1,5 @@
 ---
-title: RapidOCR集成PP-OCRv5_mobile_det模型记录
+title: RapidOCR集成PP-OCRv5_det模型(mobile/server)记录
 date: 2025-05-26
 authors: [SWHL]
 categories:
@@ -10,7 +10,7 @@ hide:
 ---
 
 
-> 该文章主要记录RapidOCR集成PP-OCRv5_mobile_det模型记录的，涉及模型转换，模型精度测试等步骤。
+> 该文章主要记录RapidOCR集成PP-OCRv5_mobile_det和PP-OCRv5_server_det模型记录的，涉及模型转换，模型精度测试等步骤。
 
 <!-- more -->
 
@@ -34,7 +34,7 @@ hide:
 安装`paddlex`:
 
 ```bash linenums="1"
-pip install "paddlex[ocr]==3.0.0rc1"
+pip install "paddlex[ocr]==3.0.0"
 ```
 
 测试PP-OCRv5_mobile_det模型能否正常识别：
@@ -43,158 +43,152 @@ pip install "paddlex[ocr]==3.0.0rc1"
 
     运行以下代码时，模型会自动下载到 **/Users/用户名/.paddlex/official_models** 下。
 
+测试图：[link](https://paddle-model-ecology.bj.bcebos.com/paddlex/imgs/demo_image/general_ocr_001.png)
+
 ```python linenums="1"
 
-from paddleocr import PaddleOCR
-# 初始化 PaddleOCR 实例
+from paddlex import create_model
 
-ocr = PaddleOCR(
-    use_doc_orientation_classify=False,
-    use_doc_unwarping=False,
-    use_textline_orientation=False)
+# mobile
+model = create_model(model_name="PP-OCRv5_mobile_det")
 
-# 对示例图像执行 OCR 推理
-result = ocr.predict(
-    input="https://paddle-model-ecology.bj.bcebos.com/paddlex/imgs/demo_image/general_ocr_002.png")
+# server
+model = create_model(model_name="PP-OCRv5_server_det")
 
-# 可视化结果并保存 json 结果
-for res in result:
+output = model.predict(input="images/general_ocr_001.png", batch_size=1)
+for res in output:
     res.print()
-    res.save_to_img("output")
-    res.save_to_json("output")
+    res.save_to_img(save_path="./output/")
+    res.save_to_json(save_path="./output/res.json")
 ```
 
 预期结果如下，表明成功运行：
 
-![alt text](../images/general_ocr_002_ocr_res_img.png)
+![alt text](../images/general_ocr_001_res.png)
 
 ### 2. 模型转换
 
 该部分主要参考文档： [docs](https://paddlepaddle.github.io/PaddleX/latest/pipeline_deploy/paddle2onnx.html?h=paddle2onnx#22)
 
-PaddleX官方集成了paddle2onnx的转换代码：
-
-```bash linenums="1"
-paddle2onnx --model_dir models/official_models/PP-OCRv5_mobile_det --model_filename inference.json --params_filename inference.pdiparams --save_file models/PP-OCRv5_mobile_det/inference.onnx
-```
+=== "转换PP-OCRv5_mobile_det"
 
-输出日志如下，日志中存在报错信息，但是最终ONNX模型仍然生成了：
-
-```bash linenums="1" hl_lines="11 16"
-/Users/xxxx/miniconda3/envs/py310/lib/python3.10/site-packages/paddle/utils/cpp_extension/extension_utils.py:711: UserWarning: No ccache found. Please be aware that recompiling all source files may be required. You can download and install ccache from: https://github.com/ccache/ccache/blob/master/doc/INSTALL.md
-  warnings.warn(warning_message)
-[Paddle2ONNX] Start parsing the Paddle model file...
-[Paddle2ONNX] Use opset_version = 14 for ONNX export.
-[Paddle2ONNX] PaddlePaddle model is exported as ONNX format now.
-2025-05-26 11:20:46 [INFO]      Try to perform constant folding on the ONNX model with Polygraphy.
-[W] 'colored' module is not installed, will not use colors when logging. To enable colors, please install the 'colored' module: python3 -m pip install colored
-[I] Folding Constants | Pass 1
-[W] colored module is not installed, will not use colors when logging. To enable colors, please install the colored module: python3 -m pip install colored
-[W] Inference failed. You may want to try enabling partitioning to see better results. Note: Error was:
-[ONNXRuntimeError] : 1 : FAIL : /Users/runner/work/1/s/onnxruntime/core/graph/model.cc:182 onnxruntime::Model::Model(ModelProto &&, const PathString &, const IOnnxRuntimeOpSchemaRegistryList *, const logging::Logger &, const ModelOptions &) Unsupported model IR version: 11, max supported IR version: 10
-[I]     Total Nodes | Original:   925, After Folding:   612 |   313 Nodes Folded
-[I] Folding Constants | Pass 2
-[W] colored module is not installed, will not use colors when logging. To enable colors, please install the colored module: python3 -m pip install colored
-[W] Inference failed. You may want to try enabling partitioning to see better results. Note: Error was:
-[ONNXRuntimeError] : 1 : FAIL : /Users/runner/work/1/s/onnxruntime/core/graph/model.cc:182 onnxruntime::Model::Model(ModelProto &&, const PathString &, const IOnnxRuntimeOpSchemaRegistryList *, const logging::Logger &, const ModelOptions &) Unsupported model IR version: 11, max supported IR version: 10
-[I]     Total Nodes | Original:   612, After Folding:   612 |     0 Nodes Folded
-2025-05-26 11:20:52 [INFO]      ONNX model saved in models/PP-OCRv5_mobile_det/inference.onnx.
-```
+    PaddleX官方集成了paddle2onnx的转换代码：
 
-此时得到的模型，直接用`rapidocr`推理会报错：
+    ```bash linenums="1"
+    paddlex --install paddle2onnx
+    pip install onnx==1.16.0
 
-```python linenums="1"
-from rapidocr import RapidOCR
+    paddlex --paddle2onnx --paddle_model_dir models/official_models/PP-OCRv5_mobile_det --onnx_model_dir models/PP-OCRv5_mobile_det
+    ```
 
-model_path = "models/PP-OCRv5_mobile_det/inference.onnx"
-engine = RapidOCR(params={"Det.model_path": model_path})
+    输出日志如下，表明转换成功：
+
+    ```bash linenums="1"
+    Input dir: models/official_models/PP-OCRv5_mobile_det
+    Output dir: models/PP-OCRv5_mobile_det
+    Paddle2ONNX conversion starting...
+    [Paddle2ONNX] Start parsing the Paddle model file...
+    [Paddle2ONNX] Use opset_version = 14 for ONNX export.
+    [Paddle2ONNX] PaddlePaddle model is exported as ONNX format now.
+    2025-05-26 21:53:00 [INFO]      Try to perform constant folding on the ONNX model with Polygraphy.
+    [W] 'colored' module is not installed, will not use colors when logging. To enable colors, please install the 'colored' module: python3 -m pip install colored
+    [I] Folding Constants | Pass 1
+    [I]     Total Nodes | Original:   925, After Folding:   502 |   423 Nodes Folded
+    [I] Folding Constants | Pass 2
+    [I]     Total Nodes | Original:   502, After Folding:   502 |     0 Nodes Folded
+    2025-05-26 21:53:08 [INFO]      ONNX model saved in models/PP-OCRv5_mobile_det/inference.onnx.
+    Paddle2ONNX conversion succeeded
+    Copied models/official_models/PP-OCRv5_mobile_det/inference.yml to models/PP-OCRv5_mobile_det/inference.yml
+    Done
+    ```
 
-img_url = "https://img1.baidu.com/it/u=3619974146,1266987475&fm=253&fmt=auto&app=138&f=JPEG?w=500&h=516"
-result = engine(img_url)
-print(result)
+=== "转换PP-OCRv5_server_det"
 
-result.vis("vis_result.jpg")
-```
+    PaddleX官方集成了paddle2onnx的转换代码：
 
-报错信息如下：
+    ```bash linenums="1"
+    paddlex --install paddle2onnx
+    pip install onnx==1.16.0
 
-```bash linenums="1" hl_lines="15"
-[INFO] 2025-05-26 11:21:27,698 [RapidOCR] base.py:41: Using engine_name: onnxruntime
-Traceback (most recent call last):
-  File "/Users/xxxx/projects/RapidOCR/python/demo.py", line 9, in <module>
-    engine = RapidOCR(params={"Det.model_path": model_path})
-  File "/Users/xxxx/projects/RapidOCR/python/rapidocr/main.py", line 60, in __init__
-    self.text_det = TextDetector(config.Det)
-  File "/Users/xxxx/projects/RapidOCR/python/rapidocr/ch_ppocr_det/main.py", line 45, in __init__
-    self.session = get_engine(config.engine_name)(config)
-  File "/Users/xxxx/projects/RapidOCR/python/rapidocr/inference_engine/onnxruntime.py", line 60, in __init__
-    self.session = InferenceSession(
-  File "/Users/xxxx/miniconda3/envs/py310/lib/python3.10/site-packages/onnxruntime/capi/onnxruntime_inference_collection.py", line 472, in __init__
-    self._create_inference_session(providers, provider_options, disabled_optimizers)
-  File "/Users/xxxx/miniconda3/envs/py310/lib/python3.10/site-packages/onnxruntime/capi/onnxruntime_inference_collection.py", line 550, in _create_inference_session
-    sess = C.InferenceSession(session_options, self._model_path, True, self._read_config_from_model)
-onnxruntime.capi.onnxruntime_pybind11_state.Fail: [ONNXRuntimeError] : 1 : FAIL : Load model from /Users/xxxx/projects/LittleCode/models/PP-OCRv5_mobile_det/inference.onnx failed:/Users/runner/work/1/s/onnxruntime/core/graph/model.cc:182 onnxruntime::Model::Model(ModelProto &&, const PathString &, const IOnnxRuntimeOpSchemaRegistryList *, const logging::Logger &, const ModelOptions &) Unsupported model IR version: 11, max supported IR version: 10
-```
+    paddlex --paddle2onnx --paddle_model_dir models/official_models/PP-OCRv5_server_det --onnx_model_dir models/PP-OCRv5_server_det
+    ```
 
-经过一系列的查阅资料，发现了两种解决方案，经过测试，精度一样。
+    输出日志如下，表明转换成功：
+
+    ```bash linenums="1"
+    Input dir: models/official_models/PP-OCRv5_server_det
+    Output dir: models/PP-OCRv5_server_det
+    Paddle2ONNX conversion starting...
+    [Paddle2ONNX] Start parsing the Paddle model file...
+    [Paddle2ONNX] Use opset_version = 14 for ONNX export.
+    [Paddle2ONNX] PaddlePaddle model is exported as ONNX format now.
+    2025-05-26 21:43:10 [INFO]      Try to perform constant folding on the ONNX model with Polygraphy.
+    [W] 'colored' module is not installed, will not use colors when logging. To enable colors, please install the 'colored' module: python3 -m pip install colored
+    [I] Folding Constants | Pass 1
+    [I]     Total Nodes | Original:  1306, After Folding:   596 |   710 Nodes Folded
+    [I] Folding Constants | Pass 2
+    [I]     Total Nodes | Original:   596, After Folding:   596 |     0 Nodes Folded
+    2025-05-26 21:43:21 [INFO]      ONNX model saved in models/PP-OCRv5_server_det/inference.onnx.
+    Paddle2ONNX conversion succeeded
+    Copied models/official_models/PP-OCRv5_server_det/inference.yml to models/PP-OCRv5_server_det/inference.yml
+    Done
+    ```
 
-=== 方案一
+### 3. 模型推理验证
 
-    安装`onnx==1.16.0`，然后再次尝试转换。我这里实际测过，的确可以成功转换。
+=== "验证PP-OCRv5_mobile_det模型"
 
-=== 方案二
+    该部分主要是在RapidOCR项目中测试能否直接使用onnx模型。要点主要是确定模型前后处理是否兼容。从PaddleOCR config文件中比较[PP-OCRv4](https://github.com/PaddlePaddle/PaddleOCR/blob/549d83a88b7c75144120e6ec03de80d3eb9e48a5/configs/det/PP-OCRv4/PP-OCRv4_mobile_det.yml)和[PP-OCRv5 mobile det](https://github.com/PaddlePaddle/PaddleOCR/blob/549d83a88b7c75144120e6ec03de80d3eb9e48a5/configs/det/PP-OCRv5/PP-OCRv5_mobile_det.yml)文件差异：
 
-    方案来自：onnxruntime issue [#23602](https://github.com/microsoft/onnxruntime/issues/23602#issuecomment-2642348849)
+    ![alt text](../images/v4_v5_mobile_det.png)
 
-    运行下面代码，将上一步所得模型重新指定一下**IR_VERSION**，就可以用`rapidocr`加载推理了。
+    从上图中可以看出，配置基本一模一样，因此现有`rapidocr`前后推理代码可以直接使用。
 
     ```python linenums="1"
-    import onnx
-    from onnx import version_converter
+    from rapidocr import RapidOCR
 
-    OPT_VERSION = 14
-    IR_VERSION = 10
+    model_path = "models/PP-OCRv5_mobile_det/inference.onnx"
+    engine = RapidOCR(params={"Det.model_path": model_path})
 
-    source_path = "models/PP-OCRv5_mobile_det/inference.onnx"
-    dist_path = "models/PP-OCRv5_mobile_det/inference_v2.onnx"
+    img_url = "https://img1.baidu.com/it/u=3619974146,1266987475&fm=253&fmt=auto&app=138&f=JPEG?w=500&h=516"
+    result = engine(img_url)
+    print(result)
 
-    model = onnx.load(source_path)
-    model.ir_version = IR_VERSION
-    model = version_converter.convert_version(model, OPT_VERSION)
-    onnx.save(model, dist_path)
+    result.vis("vis_result.jpg")
     ```
 
-### 3. 模型推理验证
+    ![alt text](../images/v5_mobile_det_vis_result.jpg)
 
-该部分主要是在RapidOCR项目中测试能否直接使用onnx模型。要点主要是确定模型前后处理是否兼容。从PaddleOCR config文件中比较[PP-OCRv4](https://github.com/PaddlePaddle/PaddleOCR/blob/549d83a88b7c75144120e6ec03de80d3eb9e48a5/configs/det/PP-OCRv4/PP-OCRv4_mobile_det.yml)和[PP-OCRv5 mobile det](https://github.com/PaddlePaddle/PaddleOCR/blob/549d83a88b7c75144120e6ec03de80d3eb9e48a5/configs/det/PP-OCRv5/PP-OCRv5_mobile_det.yml)文件差异：
+==== "验证PP-OCRv5_server_det模型"
 
-![alt text](../images/v4_v5_mobile_det.png)
+    该部分主要是在RapidOCR项目中测试能否直接使用onnx模型。要点主要是确定模型前后处理是否兼容。从PaddleOCR config文件中比较[PP-OCRv4_server_det](https://github.com/PaddlePaddle/PaddleOCR/blob/b0b31c38aef135617a98fbf89c92efd8b2eebd73/configs/det/PP-OCRv4/PP-OCRv4_server_det.yml)和[PP-OCRv5_server_det](https://github.com/PaddlePaddle/PaddleOCR/blob/b0b31c38aef135617a98fbf89c92efd8b2eebd73/configs/det/PP-OCRv5/PP-OCRv5_server_det.yml)文件差异：
 
-从上图中可以看出，配置基本一模一样，因为现有`rapidocr`前后推理代码可以直接使用。
+    ![alt text](../images/v4_v5_server_det.png)
 
-```python linenums="1"
-from rapidocr import RapidOCR
+    从上图中可以看出，配置基本一模一样，backbone换了，但是前后处理配置是一样的。因此现有`rapidocr`前后推理代码可以直接使用。
 
-model_path = "models/PP-OCRv5_mobile_det/inference.onnx"
-engine = RapidOCR(params={"Det.model_path": model_path})
+    ```python linenums="1"
+    from rapidocr import RapidOCR
 
-img_url = "https://img1.baidu.com/it/u=3619974146,1266987475&fm=253&fmt=auto&app=138&f=JPEG?w=500&h=516"
-result = engine(img_url)
-print(result)
+    model_path = "models/PP-OCRv5_server_det/inference.onnx"
+    engine = RapidOCR(params={"Det.model_path": model_path})
 
-result.vis("vis_result.jpg")
-```
+    img_url = "https://img1.baidu.com/it/u=3619974146,1266987475&fm=253&fmt=auto&app=138&f=JPEG?w=500&h=516"
+    result = engine(img_url)
+    print(result)
 
-![alt text](../images/v5_mobile_det_vis_result.jpg)
+    result.vis("vis_result.jpg")
+    ```
+
+    ![alt text](../images/v5_server_det_vis_result.jpg)
 
 ### 4. 模型精度测试
 
 !!! warning
 
     测试集[text_det_test_dataset](https://huggingface.co/datasets/SWHL/text_det_test_dataset)包括卡证类、文档类和自然场景三大类。其中卡证类有82张，文档类有75张，自然场景类有55张。缺少手写体、繁体、日文、古籍文本、拼音、艺术字等数据。因此，该基于该测评集的结果仅供参考。
 
-    欢迎有兴趣的小伙伴，可以和我们一起共建更加完整的测评集。
-
+    欢迎有兴趣的小伙伴，可以和我们一起共建更加全面的测评集。
 
 该部分主要使用[TextDetMetric](https://github.com/SWHL/TextDetMetric)和测试集[text_det_test_dataset](https://huggingface.co/datasets/SWHL/text_det_test_dataset)来评测。
 
@@ -208,7 +202,7 @@ result.vis("vis_result.jpg")
 
 ### 5. 集成到rapidocr中
 
-该部分主要包括将字典文件写入到ONNX模型中、托管模型到魔搭、更改rapidocr代码适配等。
+该部分主要包括将托管模型到魔搭、更改rapidocr代码适配等。
 
 #### 托管模型到魔搭
 

docs/blog/posts/about_model/model_summary.md

@@ -2,11 +2,13 @@
 title: 开源OCR模型对比
 date:
   created: 2022-04-16
-  updated: 2025-05-17
+  updated: 2025-05-26
 authors: [SWHL]
 categories:
   - 模型相关
 comments: true
+hide:
+  - toc
 ---
 
 > 本文主要给出了常见开源文本检测和文本识别模型的对比和评测，给大家一个使用参考。