Greek Support

[gipt66]

Will there be any greek support in the new model ?

[Aidenwu0209]

There is already an official Greek recognition model in the current PaddleOCR 3.x model list: el_PP-OCRv5_mobile_rec. It supports Greek, English, and digits.

You can use it as the recognition component of the OCR pipeline:

from paddleocr import PaddleOCR

ocr = PaddleOCR(
    text_recognition_model_name="el_PP-OCRv5_mobile_rec",
    use_doc_orientation_classify=False,
    use_doc_unwarping=False,
    use_textline_orientation=False,
)

for result in ocr.predict("greek_document.jpg"):
    result.print()

Or test the recognizer alone on already-cropped text lines:

from paddleocr import TextRecognition

rec = TextRecognition(model_name="el_PP-OCRv5_mobile_rec")
for result in rec.predict("greek_text_line.jpg"):
    result.print()

The official documentation lists the model and its inference/training downloads here:
https://github.com/PaddlePaddle/PaddleOCR/blob/main/docs/version3.x/algorithm/PP-OCRv5/PP-OCRv5_multi_languages.en.md#4-model-list

For the new unified PP-OCRv6 model, the current documentation describes Chinese, English, Japanese, and 46 Latin-script languages; it does not document Greek-script support. Only the maintainers can confirm a future PP-OCRv6 roadmap, but Greek OCR is usable today through the dedicated PP-OCRv5 model above. If your pages mix Greek with mathematical notation or unusual polytonic characters, fine-tuning this model with representative text-line samples would still be worth evaluating.