PP-OCRv5 Arabic Mobile - ONNX export

NOTICE: these are derivative ONNX exports of the PaddlePaddle PP-OCRv5 models below, converted for use with ONNX Runtime. They are not original PaddlePaddle releases.

Sources

Conversion

  • Tool: paddle2onnx==2.1.0
  • Opset: 17
  • Recognizer classes: 749 (blank + 747 dictionary characters + space, PaddleX use_space_char=True)

Artifacts

file size (bytes) sha256
PP-OCRv5_mobile_det.onnx 4766440 c8d9b07063420ce5365c74e42532de48238feeeedcdb7a330b195708bc38a93f
arabic_PP-OCRv5_mobile_rec.onnx 7994035 4e2f4ae42104e1b272463966c56ddafa3c6ad98ce9d8c7ed765ce66666ea13e1
arabic_PP-OCRv5_mobile_rec_dict.txt 2369 7f92f7dbb9b75a4787a83bfb4f6d14a8ab515525130c9d40a9036f61cf6999e9

Intended use

Selective on-device OCR in the RealTyme Android SDK.

Downloads last month

-

Downloads are not tracked for this model. How to track
Inference Providers NEW
This model isn't deployed by any Inference Provider. 🙋 Ask for provider support