Image-to-Text
PaddleOCR
Safetensors
English
Chinese
pp_ocrv5_mobile_det
OCR
PaddlePaddle
textline_detection
Instructions to use tomsanbear/pp-ocrv5-mobile-det with libraries, inference providers, notebooks, and local apps. Follow these links to get started.
- Libraries
- PaddleOCR
How to use tomsanbear/pp-ocrv5-mobile-det with PaddleOCR:
# 1. See https://www.paddlepaddle.org.cn/en/install to install paddlepaddle # 2. pip install paddleocr from paddleocr import TextDetection model = TextDetection(model_name="pp-ocrv5-mobile-det") output = model.predict(input="path/to/image.png", batch_size=1) for res in output: res.print() res.save_to_img(save_path="./output/") res.save_to_json(save_path="./output/res.json") - Notebooks
- Google Colab
- Kaggle
| { | |
| "conversion_tool": "tools/models/candle/convert_ppocr_dbnet.py@schema-v1", | |
| "dtype_device_results": [ | |
| { | |
| "device": "cpu", | |
| "dtype": "f32", | |
| "note": "native safetensors metadata and probability-map shape contract smoke; Candle forward parity lands in impl-candle-ocr-detectors", | |
| "status": "passed" | |
| }, | |
| { | |
| "device": "cuda", | |
| "dtype": "f16", | |
| "note": "accelerator validation deferred until native DBNet forward exists", | |
| "status": "unvalidated" | |
| }, | |
| { | |
| "device": "metal", | |
| "dtype": "f16", | |
| "note": "accelerator validation deferred until native DBNet forward exists", | |
| "status": "unvalidated" | |
| } | |
| ], | |
| "generated_files": [ | |
| { | |
| "path": "model.safetensors", | |
| "sha256": "c49eea87d579f931c79bf29ac1ffb1b3166f86f6811b8e46b982bb66e3f0f287", | |
| "size_bytes": 14330820 | |
| }, | |
| { | |
| "path": "config.json", | |
| "sha256": "f4f66922bfb21ce8be118fee59f8f7b0fddbfed950041d45dd234f92c8862acd", | |
| "size_bytes": 423 | |
| }, | |
| { | |
| "path": "preprocessor_config.json", | |
| "sha256": "0246cb8d114eec1447a9c358a1db636dcb8f315718d9781e0937f70a89a56900", | |
| "size_bytes": 719 | |
| }, | |
| { | |
| "path": "inference.yml", | |
| "sha256": "98069072e1b6b37d727fd9d9f11725faa46d6ea0de012f2ed26caea011c37699", | |
| "size_bytes": 903 | |
| }, | |
| { | |
| "path": "README.md", | |
| "sha256": "2e6a3112ae23e53bb2d9d653f84d15eda859bed0b9a3accb5d1c4d23607b2792", | |
| "size_bytes": 2041 | |
| }, | |
| { | |
| "path": "ocr_pipeline.py", | |
| "sha256": "50cc6b790afc3c2073d45a94dd57843fa971d1dfa45bece8ccbf77830d5c27e9", | |
| "size_bytes": 5477 | |
| }, | |
| { | |
| "path": "tensor_map.json", | |
| "sha256": "581ffa08dbe2546717b07f650e547fc099187ca84ce74c712dc26d2eee5bcb06", | |
| "size_bytes": 326748 | |
| }, | |
| { | |
| "path": "bundle.json", | |
| "sha256": "3db78e357008fe2d45642588738f56443cd86d28228a80ed5bd54a0d3061dd11", | |
| "size_bytes": 4326 | |
| } | |
| ], | |
| "parity_fixture_ids": [ | |
| "axera-ppocrv5-dbnet-synthetic-gray-probability-map", | |
| "axera-ppocrv5-dbnet-omnidocbench-line-polygons" | |
| ], | |
| "schema_version": 1, | |
| "source_files": [ | |
| { | |
| "path": "README.md", | |
| "sha256": "2e6a3112ae23e53bb2d9d653f84d15eda859bed0b9a3accb5d1c4d23607b2792", | |
| "size_bytes": 2041 | |
| }, | |
| { | |
| "path": "config.json", | |
| "sha256": "f4f66922bfb21ce8be118fee59f8f7b0fddbfed950041d45dd234f92c8862acd", | |
| "size_bytes": 423 | |
| }, | |
| { | |
| "path": "inference.yml", | |
| "sha256": "98069072e1b6b37d727fd9d9f11725faa46d6ea0de012f2ed26caea011c37699", | |
| "size_bytes": 903 | |
| }, | |
| { | |
| "path": "model.safetensors", | |
| "sha256": "c49eea87d579f931c79bf29ac1ffb1b3166f86f6811b8e46b982bb66e3f0f287", | |
| "size_bytes": 14330820 | |
| }, | |
| { | |
| "path": "ocr_pipeline.py", | |
| "sha256": "50cc6b790afc3c2073d45a94dd57843fa971d1dfa45bece8ccbf77830d5c27e9", | |
| "size_bytes": 5477 | |
| }, | |
| { | |
| "path": "preprocessor_config.json", | |
| "sha256": "0246cb8d114eec1447a9c358a1db636dcb8f315718d9781e0937f70a89a56900", | |
| "size_bytes": 719 | |
| } | |
| ], | |
| "tested_io": [ | |
| { | |
| "id": "ppocr-dbnet-native-960-probability-map-shape-smoke", | |
| "inputs": [ | |
| { | |
| "dtype": "f32", | |
| "name": "pixel_values", | |
| "shape": [ | |
| 1, | |
| 3, | |
| 960, | |
| 960 | |
| ] | |
| } | |
| ], | |
| "outputs": [ | |
| { | |
| "dtype": "f32", | |
| "name": "probability_map", | |
| "shape": [ | |
| 1, | |
| 1, | |
| 960, | |
| 960 | |
| ] | |
| } | |
| ] | |
| }, | |
| { | |
| "id": "ppocr-dbnet-source-dynamic-contract", | |
| "inputs": [ | |
| { | |
| "dtype": "f32", | |
| "name": "pixel_values", | |
| "shape": [ | |
| "batch", | |
| 3, | |
| "height_32_aligned", | |
| "width_32_aligned" | |
| ] | |
| } | |
| ], | |
| "outputs": [ | |
| { | |
| "dtype": "f32", | |
| "name": "probability_map", | |
| "shape": [ | |
| "batch", | |
| 1, | |
| "height_32_aligned", | |
| "width_32_aligned" | |
| ] | |
| } | |
| ], | |
| "source_model_input_names": [ | |
| "pixel_values", | |
| "original_image_size" | |
| ], | |
| "source_resize_long": 960 | |
| } | |
| ] | |
| } | |