Error

[thebest132]

Here is my code:

from paddleocr import PaddleOCR, draw_ocr

ocr = PaddleOCR(lang="japan", det_db_unclip_ratio=2, det_algorithm="SAST", use_angle_cls=False)

Error:
ppocr ERROR: det_algorithm must in ['DB']

[GreatV]

The error message:

ppocr ERROR: det_algorithm must in ['DB']

indicates that the det_algorithm parameter in your PaddleOCR initialization is invalid. Currently, PaddleOCR supports only the "DB" (Differentiable Binarization) detection algorithm, whereas you have specified "SAST".

Solution:

Modify your code to use an appropriate value for det_algorithm. Change the following line:

ocr = PaddleOCR(lang="japan", det_db_unclip_ratio=2, det_algorithm="SAST", use_angle_cls=False)

to:

ocr = PaddleOCR(lang="japan", det_db_unclip_ratio=2, det_algorithm="DB", use_angle_cls=False)

Explanation:

• The "SAST" algorithm is not supported in the current PaddleOCR implementation for text detection.
• "DB" is the only supported option for det_algorithm, as indicated in the error message.

After making this change, your code should work correctly without the det_algorithm error.

Let me know if you need further assistance! 🚀

---

Response generated by 🤖 feifei-bot | chatgpt-4o-latest

⚠️ This is an AI-generated response. Please verify any critical information.

[thebest132]

@GreatV In this link

PaddleOCR/docs/ppocr/blog/inference_args.en.md

Line 47 in 1d52a2b

| det_algorithm | str | "DB" | Text detection algorithm name, currently supports `DB`, `EAST`, `SAST`, `PSE`, `DB++`, `FCE` |

docs it says SAST and EAST are supported. What's the problem?

[GreatV]

This applies to our Python inference scripts, not the Python API interface. See https://paddlepaddle.github.io/PaddleOCR/latest/en/infer_deploy/python_infer.html for more information

[thebest132]

@GreatV It seems to all be through command line, which has limited usability. I would like to use it from code. Is this possible?