Instructions to use PaddlePaddle/PP-DocLayout_plus-L_safetensors with libraries, inference providers, notebooks, and local apps. Follow these links to get started.
- Libraries
- PaddleOCR
How to use PaddlePaddle/PP-DocLayout_plus-L_safetensors with PaddleOCR:
# 1. See https://www.paddlepaddle.org.cn/en/install to install paddlepaddle # 2. pip install paddleocr from paddleocr import LayoutDetection model = LayoutDetection(model_name="PP-DocLayout_plus-L_safetensors") 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
Upload config.json with huggingface_hub
Browse files- config.json +17 -2
config.json
CHANGED
|
@@ -106,7 +106,22 @@
|
|
| 106 |
2,
|
| 107 |
1
|
| 108 |
],
|
| 109 |
-
"use_learnable_affine_block": false
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
| 110 |
},
|
| 111 |
"batch_norm_eps": 1e-05,
|
| 112 |
"box_noise_scale": 1.0,
|
|
@@ -214,4 +229,4 @@
|
|
| 214 |
"weight_loss_giou": 2.0,
|
| 215 |
"weight_loss_vfl": 1.0,
|
| 216 |
"with_box_refine": true
|
| 217 |
-
}
|
|
|
|
| 106 |
2,
|
| 107 |
1
|
| 108 |
],
|
| 109 |
+
"use_learnable_affine_block": false,
|
| 110 |
+
"return_idx": [
|
| 111 |
+
1,
|
| 112 |
+
2,
|
| 113 |
+
3
|
| 114 |
+
],
|
| 115 |
+
"freeze_stem_only": true,
|
| 116 |
+
"freeze_at": 0,
|
| 117 |
+
"freeze_norm": false,
|
| 118 |
+
"lr_mult_list": [
|
| 119 |
+
0.05,
|
| 120 |
+
0.05,
|
| 121 |
+
0.1,
|
| 122 |
+
0.15,
|
| 123 |
+
0.2
|
| 124 |
+
]
|
| 125 |
},
|
| 126 |
"batch_norm_eps": 1e-05,
|
| 127 |
"box_noise_scale": 1.0,
|
|
|
|
| 229 |
"weight_loss_giou": 2.0,
|
| 230 |
"weight_loss_vfl": 1.0,
|
| 231 |
"with_box_refine": true
|
| 232 |
+
}
|