Clarification on PP-OCRv5 Enhancements and the Role of model_name in Global Config

[ManikSinghSarmaal]

I am curious since the repository has gone through a lot of changes in the last 2-3 weeks and I can also see PP-OCRv5 being added in the configs. I don't see any difference between PP-OCRv4 and PP-OCRv5, from architecture to backbones it's all same. Am I missing on something ? Please clarify.
Also there is a new thing added in most of the configs including older ones which is model_name in the Global
Global:
model_name: PP-OCRv5_mobile_rec # To use static model for inference.
What does this mean, is this the pdiparam model path for inference ?

[liuhongen1234567]

Hello, PP-OCRv5 mainly supports the detection and recognition of handwritten text, traditional Chinese characters, pinyin, and other text by using diverse data. There haven’t been significant changes in the network structure. You can view its introduction in detail by clicking on the Text Recognition Module in the PP-OCRv5 documentation: https://paddlepaddle.github.io/PaddleOCR/latest/en/version3.x/pipeline_usage/OCR.html#1-ocr-pipeline-introduction.

[liuhongen1234567]

Regarding model_name in the Global, it is a new feature in PaddleOCR 3.0. Once set, the exported model will have a Global.model_name parameter, and the inference framework will use this model_name to select the corresponding pre-processing and post-processing methods for inference. Here, model_name must be one of the module names supported by PP-OCRv5. For more details, you can click on the Text Detection Module or Text Recognition Module. The model_name is not the pdiparam model path; the pdiparam model path can be set using the model_dir parameter.