Kreuzberg/paddleocr-onnx-models
By Kreuzberg · ↓ 0 · ♥ 0 · Updated 2026-03-15
What's in this repo
This page lists the files in Kreuzberg/paddleocr-onnx-models, including PP-OCRv5_server_det_infer.onnx, ch_ppocr_mobile_v2.0_cls_infer.onnx, model.onnx and more.
- 28 ONNX models (.onnx files)
- 44 files, including 28 .onnx, 14 .txt, 1 .md
- Last updated 2026-03-15
ONNX files
| File | Size | Download |
|---|---|---|
| PP-OCRv5_server_det_infer.onnx | — | Download |
| ch_ppocr_mobile_v2.0_cls_infer.onnx | — | Download |
| rec/arabic/model.onnx | — | Download |
| rec/chinese/model.onnx | — | Download |
| rec/devanagari/model.onnx | — | Download |
| rec/english/model.onnx | — | Download |
| rec/eslav/model.onnx | — | Download |
| rec/greek/model.onnx | — | Download |
| rec/korean/model.onnx | — | Download |
| rec/latin/model.onnx | — | Download |
Showing 10 of 28 .onnx files — see the full file list on the right.
How to use
These ONNX files can be loaded with ONNX Runtime. To inspect a model graph, upload the file to the Netron viewer. Download links point to the original files on Hugging Face.
README
PaddleOCR ONNX Models
Pre-converted PaddleOCR models in ONNX format for use with Kreuzberg.
Models
| File | Type | Version | Size | Description |
|---|---|---|---|---|
ch_PP-OCRv4_det_infer.onnx |
Detection | PP-OCRv4 | 4.5 MB | Text region detection (language-agnostic) |
ch_ppocr_mobile_v2.0_cls_infer.onnx |
Classification | PPOCRv2 | 572 KB | Text angle classification (0° vs 180°) |
en_PP-OCRv4_rec_infer.onnx |
Recognition | PP-OCRv4 | 7.3 MB | English text recognition |
Pipeline
PaddleOCR uses a three-stage pipeline:
- Detection (
det): Locates text regions in the image using a DB (Differentiable Binarization) network. The detection model is language-agnostic — it identifies bounding boxes around text regardless of script. - Classification (
cls): Determines text orientation (upright vs upside-down) to correct rotated text before recognition. - Recognition (
rec): Reads characters from each detected text region. This is the only language-specific model.
Sources & Credits
These models are derived from PaddlePaddle/PaddleOCR, licensed under Apache 2.0.
- Detection model (
ch_PP-OCRv4_det_infer.onnx): Pre-converted ONNX model sourced from SWHL/RapidOCR on HuggingFace (PP-OCRv4 folder). Originally trained by PaddlePaddle asch_PP-OCRv4_det_infer. - Classification model (
ch_ppocr_mobile_v2.0_cls_infer.onnx): Pre-converted ONNX model sourced from SWHL/RapidOCR on HuggingFace (PP-OCRv1 folder). Originally trained by PaddlePaddle asch_ppocr_mobile_v2.0_cls_infer. - Recognition model (
en_PP-OCRv4_rec_infer.onnx): Converted from PaddlePaddle format to ONNX using paddle2onnx (opset 14). Original PaddlePaddle model downloaded from PaddlePaddle's official distribution (en_PP-OCRv4_rec_infer).
Acknowledgments
- PaddlePaddle for creating and training the PP-OCR model series
- SWHL/RapidOCR for maintaining pre-converted ONNX models
- Paddle2ONNX for the PaddlePaddle-to-ONNX conversion tool
License
The original PaddleOCR models are licensed under Apache License 2.0.
Model info
Created: 2026-02-06
Updated: 2026-03-15
View on Hugging Face