[Question]: When recognizing PDF files, layout analysis was used, and OCR often recognizes text. Is there any method to avoid text loss? What parameters can be adjusted to make OCR more accurate?

[wanyuks]

Describe your problem

some examples

[KevinHuSh]

Is it locally deployed? We are gona upgrade the open source model soon.

[wanyuks]

Is it locally deployed? We are gona upgrade the open source model soon.

Yes, this is a locally deployed onnx model downloaded from HF.

[chenhanbiao]

Is it locally deployed? We are gona upgrade the open source model soon.

hope to see the new model soon!