Help Required – Model Download URLs Not Working for PaddleOCR Inference (StructureV3/Table/Rec)

[RajeshGit001]

Hello PaddleOCR Community,

I have been attempting to set up the StructureV3 + Table + Chinese PP-OCRv3 models for running layout and table extraction using the model_download.sh script and demo code provided in the GitHub README.

However, the .tar file links such as:

• https://paddleocr.bj.bcebos.com/structure/layout/layout_cdla_infer.tar
• https://paddleocr.bj.bcebos.com/structure/table_structure_mobile_v2_0_SLANet_infer.tar
• https://paddleocr.bj.bcebos.com/PP-OCRv3/ch_PP-OCRv3_rec_infer.tar

…are returning 404 Not Found. It appears the URLs or model hosting has been changed since the recent PaddleOCR 3.1.0 release.

My environment:

• Python 3.10.14
• PaddleOCR latest from GitHub
• Linux (Azure VM)
• Dependencies: paddlepaddle-gpu==2.5.0

Can someone kindly share:

1. Updated links to download these models
2. Whether the model download has shifted to HuggingFace or another CDN?
3. Any working model_download.sh with the new paths?

I’m in the middle of a production-grade document analysis pipeline and would appreciate your help.

Thanks in advance 🙏

— Rajesh

[zhangyubo0722]

It is recommended to check the latest documentation at https://github.com/PaddlePaddle/PaddleOCR/blob/main/docs/version3.x/pipeline_usage/PP-StructureV3.en.md