PaddleOCR/main/en/version3.x/algorithm/PP-OCRv5/PP-OCRv5_multi_languages

[giscus[bot]]

PaddleOCR/main/en/version3.x/algorithm/PP-OCRv5/PP-OCRv5_multi_languages

Awesome multilingual OCR toolkits based on PaddlePaddle (practical ultra lightweight OCR system, support 80+ languages recognition, provide data annotation and synthesis tools, support training and deployment among server, mobile, embedded and IoT devices)

https://paddlepaddle.github.io/PaddleOCR/main/en/version3.x/algorithm/PP-OCRv5/PP-OCRv5_multi_languages.html

[stormA4ka]

Hello, lads. How can i use ru, model for recognition? Russian text is not recognized.
how cai i fix it?
from paddleocr import PaddleOCR

ocr = PaddleOCR(
lang="ru", # Specify French recognition model with the lang parameter
use_doc_orientation_classify=False, # Disable document orientation classification model
use_doc_unwarping=False, # Disable text image unwarping model
use_textline_orientation=False, # Disable text line orientation classification model
# rec_model_dir="/data/shared_folder/NLP/image2text/paddle_inference/cyrillic_PP-OCRv3_mobile_rec_infer"
)
result = ocr.predict("/data/shared_folder/NLP/image2text/data/gimn.png")
for res in result:
res.print()
res.save_to_img("output")
res.save_to_json("output")

[adenm1]

I'm good with these code, but accuracy is low.

[zhangyubo0722]

Could you send me the original image？

[adenm1]

Could you send me the original image？

我已经迭代好几个版本了，有问题留下邮箱交流？

[ezatgz]

我用以下配置只能OCR出中文和日文，无法OCR出繁体中文和英文。如果用PP-OCRv5_server_rec则可以识别出这四种语言。

ocr_engine = PaddleOCR(
    use_doc_orientation_classify=False,
    use_doc_unwarping=False,
    use_textline_orientation=False,
    text_detection_model_name="PP-OCRv5_mobile_det",
    text_recognition_model_name="PP-OCRv5_mobile_rec",
    lang="cn",
    #lang="chinese_cht",
    #lang="en"
    cpu_threads=8,
    text_det_limit_side_len=1280,
    text_detection_model_dir="./ocrmodels/PP-OCRv5_mobile_det",
    text_recognition_model_dir="./ocrmodels/PP-OCRv5_mobile_rec", 
    text_det_limit_type="max"
    # 移除enable_mkldnn=False以测试是否能正常运行
)

虽然lang设置为cn，但根据提示：UserWarning: lang and ocr_version will be ignored when model names or model directories are not None. 这个lang设置也会被忽略。

[Yugx93]

请问PaddleOCR是否有功能实现根据图像识别出语种，然后在调用相应模型的功能么？

[one1simus]

Hello,

I’d like to add support for my native language, Tigrinya. Tigrinya uses the [Ethiopic Unicode script[ (https://www.unicode.org/charts/PDF/U1200.pdf), which is also shared by Amharic, Tigre, Ge’ez, and other Eritrean and Ethiopian languages. Could you please point me to any documentation or resources on how to train PaddleOCR for new languages?

Thank you very much!

[one1simus]

[Ethiopic Unicode script] (https://www.unicode.org/charts/PDF/U1200.pdf)

[DesBw]

I have a lot of data for the training (multiple GBs of text, and images; prepared for training). But, I don't have the resources (the CPU and GPU) to train Ethiopic. If anyone has the resources, we can put our hands together for this.

[Danicortex]

have you considered using modal it provides cloud resources for training whatever you desire and you get free credits monthly, should be enough for your use case.