How to load non-default PaddleOCR detection model in a python script?

Question

PaddleOCR seems to support different algorithms/models for text detection, as written in their algorithms overview. However, when trying to load a different model from "DB" (for example "SAST" or "EAST") an error is encountered (I get the following error: ppocr ERROR: det_algorithm must in ['DB']). From the code it seems that only the "DB" model is supported - see here. I have downloaded the models from the links posted in the documentation, is there a way to use them in a python script? I don't intend to use them as illustrated in the tutorial, for example running python3 tools/infer/predict_det.py --image_dir="./doc/imgs_en/img_10.jpg" --det_model_dir="./inference/det_r50_east/" --det_algorithm="EAST".

I have tried something like the following:

from paddleocr import PaddleOCR

ocr = PaddleOCR(use_angle_cls=True, lang="en", det_algorithm="DB")
result = ocr.ocr(img, cls=True)

Anidh Singh · Accepted Answer · 2023-05-30 10:25:47Z

2

This looks like you have installed the paddleOCR using the whl package. The whl package currently only supports the DB algorithm. For predictions of other models, it is recommended to clone the entire project. You can view this discussion here

answered May 30, 2023 at 10:25

Anidh Singh

3232 gold badges7 silver badges19 bronze badges

Sign up to request clarification or add additional context in comments.

1 Comment

N4ppeL Over a year ago

wouldn't you need to re-train the detection model for another algorithm anyway?

Collectives™ on Stack Overflow

How to load non-default PaddleOCR detection model in a python script?

1 Answer 1

1 Comment

Your Answer

Hot Network Questions

Collectives™ on Stack Overflow

1 Answer 1

1 Comment

Your Answer

Sign up or log in

Post as a guest

Related