RapidOCR 发行版

最新版

v1.2.0

a7acbb2

2023-01-21 21:30

[Python] v1.2.0

最后提交信息为： Update pypi README of rapid orientation

v1.1.0

156b21a

2022-08-17 21:23

[Python] v1.1.0

#### 本次更新要点：
1. 将文本识别部分所需要的字典文件写入到ONNX模型中，这样分发时只有模型文件，避免模型与字典文件不统一的问题，感谢[AutumnSun1996](https://github.com/AutumnSun1996)在[issue 42](https://github.com/RapidAI/RapidOCR/issues/42)提出。
2. 当前推理代码是兼容v1.0.0的，也就是说传入字典文件，会优先加载传入的字典文件。
3. 替换之前可视化的字体（msyh.ttc）为方正姚体(FZYTK.TTF)，主要原因是，后者字体文件更小，便于下载。
4. 统一各个模块的名称，避免引起困惑。
5. ocrweb部分添加API部署调用方式，详情参见[以API方式运行和调用](https://github.com/RapidAI/RapidOCR/tree/main/ocrweb#%E4%BB%A5api%E6%96%B9%E5%BC%8F%E8%BF%90%E8%A1%8C%E5%92%8C%E8%B0%83%E7%94%A8)
6. 添加检测后处理参数score_mode=slow 增加识别率 by @DogeVenci [#37]( https://github.com/RapidAI/RapidOCR/pull/37)

#### 附件各个文件目录结构：
- ocrweb_v1.1.0.zip
  <details>
  
    ```text
    ocrweb_v1.1.0/
    ├── api.py
    ├── config.yaml
    ├── main.py
    ├── rapidocr_onnxruntime
    │   ├── ch_ppocr_v2_cls
    │   ├── ch_ppocr_v3_det
    │   ├── ch_ppocr_v3_rec
    │   ├── __init__.py
    │   └── rapid_ocr_api.py
    ├── README.md
    ├── requirements.txt
    ├── resources
    │   └── models
    │       ├── ch_ppocr_mobile_v2.0_cls_infer.onnx
    │       ├── ch_PP-OCRv3_det_infer.onnx
    │       └── ch_PP-OCRv3_rec_infer.onnx
    ├── static
    │   ├── css
    │   └── js
    ├── task.py
    └── templates
        └── index.html
    ```
  </details>

- rapidocr_onnxruntime_v1.1.0.zip
  <details>
  
    ```text
    rapidocr_onnxruntime_v1.1.0/
    ├── config.yaml
    ├── rapidocr_onnxruntime
    │   ├── ch_ppocr_v2_cls
    │   ├── ch_ppocr_v3_det
    │   ├── ch_ppocr_v3_rec
    │   ├── __init__.py
    │   └── rapid_ocr_api.py
    ├── README.md
    ├── requirements.txt
    ├── resources
    │   ├── fonts
    │   │    └── FZYTK.TTF
    │   └── models
    │       ├── ch_ppocr_mobile_v2.0_cls_infer.onnx
    │       ├── ch_PP-OCRv3_det_infer.onnx
    │       └── ch_PP-OCRv3_rec_infer.onnx
    ├── setup.py
    ├── test_demo.py
    └── test_images
        ├── ch_en_num.jpg
        └── single_line_text.jpg
    ```
  </details>

- rapidocr_openvino_v1.1.0.zip
  <details>
  
    ```text
    rapidocr_openvino_v1.1.0/
    ├── config.yaml
    ├── rapidocr_openvino
    │   ├── ch_ppocr_v2_cls
    │   ├── ch_ppocr_v3_det
    │   ├── ch_ppocr_v3_rec
    │   ├── __init__.py
    │   ├── rapid_ocr_api.py
    │   └── README.md
    ├── README.md
    ├── requirements.txt
    ├── resources
    │   ├── fonts
    │   │    └── FZYTK.TTF
    │   └── models
    │       ├── ch_ppocr_mobile_v2.0_cls_infer.onnx
    │       ├── ch_PP-OCRv3_det_infer.onnx
    │       └── ch_PP-OCRv3_rec_infer.onnx
    ├── setup.py
    ├── test_demo.py
    └── test_images
        ├── ch_en_num.jpg
        └── single_line_text.jpg
    ```
  </details>

- required_for_whl_v1.1.0.zip
  <details>
  
    ```text
    required_for_whl_v1.1.0/
    ├── config.yaml
    ├── README.md
    ├── resources
    │   └── models
    │       ├── ch_ppocr_mobile_v2.0_cls_infer.onnx
    │       ├── ch_PP-OCRv3_det_infer.onnx
    │       └── ch_PP-OCRv3_rec_infer.onnx
    ├── test_demo.py
    └── test_images
        ├── ch_en_num.jpg
        └── single_line_text.jpg
    ```
  </details>

- resources.zip
  <details>
  
    ```text
    resources/
    ├── fonts
    │   └── FZYTK.TTF
    └── models
        ├── ch_ppocr_mobile_v2.0_cls_infer.onnx
        ├── ch_PP-OCRv3_det_infer.onnx
        └── ch_PP-OCRv3_rec_infer.onnx
    ```
  </details>

最后提交信息为： Merge pull request #44 from RapidAI/develop

v1.0.0

9445cca

2022-07-10 15:56

[Python] Inference code base ONNXRuntime, OpenVINO and ocrweb

## v1.0.0

- Considering that it is tedious to directly use the inference code under the existing repository, we hereby release a version that packages the resources directory with each other inference scenario independently.
- The contents of the resources directory are shown below, and the models are the same as those in the download [link](https://drive.google.com/drive/folders/1x_a9KpCo_1blxH1xFOfgKVkw1HYRVywY?usp=sharing) given in the repository, which is the current The best model combination.
  ```text
  resources
  ├── fonts
  │   └── msyh.ttc
  ├── models
  │   ├── ch_ppocr_mobile_v2.0_cls_infer.onnx
  │   ├── ch_PP-OCRv3_det_infer.onnx
  │   └── ch_PP-OCRv3_rec_infer.onnx
  └── rec_dict
      └── ppocr_keys_v1.txt
  ```
- You can go to [BaiduNetDisk](https://pan.baidu.com/s/1PTcgXG2zEgQU6A_A3kGJ3Q?pwd=jhai) | [Google Drive](https://drive.google.com/drive/folders/1x_a9KpCo_1blxH1xFOfgKVkw1HYRVywY?usp=sharing) to download other models according to your needs to download other models.
-  The following attached zip file contains the complete runtime code and model, which can be downloaded directly and refer to the README to run the sample demo.
---
中文版：

- 考虑到直接使用现有仓库下的推理代码步骤繁琐，特此release一版，将resources目录与其他各个推理场景独立打包。
- 其中，resources目录下内容如下所示，模型与仓库中所给下载[链接](https://drive.google.com/drive/folders/1x_a9KpCo_1blxH1xFOfgKVkw1HYRVywY?usp=sharing)中的一致，为目前最优模型组合。
  ```text
  resources
  ├── fonts
  │   └── msyh.ttc
  ├── models
  │   ├── ch_ppocr_mobile_v2.0_cls_infer.onnx
  │   ├── ch_PP-OCRv3_det_infer.onnx
  │   └── ch_PP-OCRv3_rec_infer.onnx
  └── rec_dict
      └── ppocr_keys_v1.txt
  ```
- 小伙伴可根据需要，自行去 [百度网盘](https://pan.baidu.com/s/1PTcgXG2zEgQU6A_A3kGJ3Q?pwd=jhai) | [Google Drive](https://drive.google.com/drive/folders/1x_a9KpCo_1blxH1xFOfgKVkw1HYRVywY?usp=sharing)下载其他模型。
- 以下附件中的zip文件中含有完整的运行代码和模型，可直接下载，参照README来运行示例Demo。

最后提交信息为： Test CI

RapidAI/RapidOCR .gitee-modal { width: 500px !important; }

搜索帮助

RapidAI/RapidOCR