| title: PP-OCRv6 Medium Text Detection | |
| emoji: 🔍 | |
| colorFrom: pink | |
| colorTo: yellow | |
| sdk: gradio | |
| sdk_version: 6.15.1 | |
| app_file: app.py | |
| short_description: Detect text regions in images with PP-OCRv6 medium | |
| python_version: "3.12" | |
| startup_duration_timeout: 30m | |
| # PP-OCRv6 Medium Text Detection | |
| This Space runs [PaddlePaddle/PP-OCRv6_medium_det](https://huggingface.co/PaddlePaddle/PP-OCRv6_medium_det) — a lightweight 15.5M-parameter text detection model from the PaddleOCR team. | |
| ## What it does | |
| Upload an image and the model will detect text regions, drawing bounding boxes around each detected text line. The model performs **detection only** (bounding boxes), not recognition (reading the characters). | |
| ## Details | |
| - **Architecture**: LCNetV4 backbone + RepLKFPN neck | |
| - **Parameters**: 15.5M | |
| - **Detection Hmean**: 86.2% | |
| - **License**: Apache-2.0 |