Object Detection
Transformers
ONNX
PaddleOCR
English
Chinese
multilingual
pp_doclayout_v3
PaddlePaddle
image-segmentation
ocr
layout
layout_detection
Instructions to use beclab/PP-DocLayoutV3_onnx with libraries, inference providers, notebooks, and local apps. Follow these links to get started.
- Libraries
- Transformers
How to use beclab/PP-DocLayoutV3_onnx with Transformers:
# Use a pipeline as a high-level helper from transformers import pipeline pipe = pipeline("object-detection", model="beclab/PP-DocLayoutV3_onnx")# Load model directly from transformers import AutoImageProcessor, AutoModelForObjectDetection processor = AutoImageProcessor.from_pretrained("beclab/PP-DocLayoutV3_onnx") model = AutoModelForObjectDetection.from_pretrained("beclab/PP-DocLayoutV3_onnx", device_map="auto") - PaddleOCR
How to use beclab/PP-DocLayoutV3_onnx 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-DocLayoutV3_onnx") 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 PP-DocLayoutV3 ONNX for layout-service (ORT CPU)
Browse files- PP-DocLayoutV3.onnx +3 -0
- README.md +130 -0
- config.json +107 -0
- inference.yml +100 -0
- preprocessor_config.json +36 -0
PP-DocLayoutV3.onnx
ADDED
|
@@ -0,0 +1,3 @@
|
|
|
|
|
|
|
|
|
|
|
|
|
| 1 |
+
version https://git-lfs.github.com/spec/v1
|
| 2 |
+
oid sha256:b0deee066f8b71e6f8ae3a645c242f1985c4b66e6e3332e7d72cae774a7f70ac
|
| 3 |
+
size 142486972
|
README.md
ADDED
|
@@ -0,0 +1,130 @@
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
| 1 |
+
---
|
| 2 |
+
license: apache-2.0
|
| 3 |
+
pipeline_tag: object-detection
|
| 4 |
+
tags:
|
| 5 |
+
- PaddleOCR
|
| 6 |
+
- PaddlePaddle
|
| 7 |
+
- image-segmentation
|
| 8 |
+
- ocr
|
| 9 |
+
- layout
|
| 10 |
+
- layout_detection
|
| 11 |
+
language:
|
| 12 |
+
- en
|
| 13 |
+
- zh
|
| 14 |
+
- multilingual
|
| 15 |
+
library_name: transformers
|
| 16 |
+
base_model:
|
| 17 |
+
- PaddlePaddle/PP-DocLayoutV3
|
| 18 |
+
---
|
| 19 |
+
<div align="center">
|
| 20 |
+
|
| 21 |
+
|
| 22 |
+
<h1 align="center">
|
| 23 |
+
|
| 24 |
+
Unified Layout Module for PaddleOCR-VL 1.5/1.6 & GLM-OCR
|
| 25 |
+
|
| 26 |
+
</h1>
|
| 27 |
+
|
| 28 |
+
[](https://github.com/PaddlePaddle/PaddleOCR)
|
| 29 |
+
[](https://huggingface.co/PaddlePaddle/PP-DocLayoutV3)
|
| 30 |
+
[](https://modelscope.cn/models/PaddlePaddle/PP-DocLayoutV3)
|
| 31 |
+
[](https://huggingface.co/spaces/PaddlePaddle/PaddleOCR-VL-1.5_Online_Demo)
|
| 32 |
+
[](https://modelscope.cn/studios/PaddlePaddle/PaddleOCR-VL-1.5_Online_Demo/summary)
|
| 33 |
+
[](https://discord.gg/JPmZXDsEEK)
|
| 34 |
+
[](https://x.com/PaddlePaddle)
|
| 35 |
+
[](./LICENSE)
|
| 36 |
+
|
| 37 |
+
**🔥 [Official Website](https://www.paddleocr.com)** |
|
| 38 |
+
**📝 [Technical Report](https://arxiv.org/abs/2606.23344)**
|
| 39 |
+
|
| 40 |
+
</div>
|
| 41 |
+
|
| 42 |
+
|
| 43 |
+
|
| 44 |
+
|
| 45 |
+
## Introduction
|
| 46 |
+
|
| 47 |
+
This is the model weights for PP-DocLayoutv3 in safetensors format. Get PaddlePaddle weights at [PP-DocLayoutV3](https://huggingface.co/PaddlePaddle/PP-DocLayoutV3)
|
| 48 |
+
|
| 49 |
+
**PP-DocLayoutV3 is specifically engineered to handle non-planar document images. It can directly predict multi-point bounding boxes for layout elements—as opposed to standard two-point boxes—and determine logical reading orders for skewed and curved surfaces within a single forward pass, significantly reducing cascading errors.** This model is an essential component of PaddleOCR-VL-1.5, providing crucial layout analysis for the high-precision parsing of various real-world documents in PaddleOCR-VL.
|
| 50 |
+
|
| 51 |
+
This work has been accepted to ECCV 2026! 🎉
|
| 52 |
+
|
| 53 |
+
|
| 54 |
+
### **Model Architecture**
|
| 55 |
+
|
| 56 |
+
<div align="center">
|
| 57 |
+
<img src="https://raw.githubusercontent.com/cuicheng01/PaddleX_doc_images/refs/heads/main/images/paddleocr_vl_1_5/PP-DocLayoutV3.png" width="800"/>
|
| 58 |
+
</div>
|
| 59 |
+
|
| 60 |
+
## Model Usage
|
| 61 |
+
|
| 62 |
+
```python
|
| 63 |
+
import requests
|
| 64 |
+
from PIL import Image
|
| 65 |
+
from transformers import AutoImageProcessor, AutoModelForObjectDetection
|
| 66 |
+
|
| 67 |
+
model_path = "PaddlePaddle/PP-DocLayoutV3_safetensors"
|
| 68 |
+
model = AutoModelForObjectDetection.from_pretrained(model_path)
|
| 69 |
+
image_processor = AutoImageProcessor.from_pretrained(model_path)
|
| 70 |
+
|
| 71 |
+
image = Image.open(requests.get("https://paddle-model-ecology.bj.bcebos.com/paddlex/imgs/demo_image/layout_demo.jpg", stream=True).raw)
|
| 72 |
+
inputs = image_processor(images=image, return_tensors="pt")
|
| 73 |
+
|
| 74 |
+
outputs = model(**inputs)
|
| 75 |
+
results = image_processor.post_process_object_detection(outputs, target_sizes=[image.size[::-1]])
|
| 76 |
+
for result in results:
|
| 77 |
+
for idx, (score, label_id, box, polygon_points) in enumerate(zip(result["scores"], result["labels"], result["boxes"], result["polygon_points"])):
|
| 78 |
+
score, label = score.item(), label_id.item()
|
| 79 |
+
box = [round(i, 2) for i in box.tolist()]
|
| 80 |
+
print(f"Order {idx + 1}: {model.config.id2label[label]}, score: {score:.2f}, box: {box}, polygon_points: {polygon_points}")
|
| 81 |
+
```
|
| 82 |
+
|
| 83 |
+
## Visualization
|
| 84 |
+
|
| 85 |
+
|
| 86 |
+
### Light Variation
|
| 87 |
+
|
| 88 |
+
<div align="center">
|
| 89 |
+
<img src="https://raw.githubusercontent.com/cuicheng01/PaddleX_doc_images/refs/heads/main/images/paddleocr_vl_1_5/layout_lighting.jpg" width="800"/>
|
| 90 |
+
</div>
|
| 91 |
+
|
| 92 |
+
|
| 93 |
+
### Skewing
|
| 94 |
+
|
| 95 |
+
<div align="center">
|
| 96 |
+
<img src="https://raw.githubusercontent.com/cuicheng01/PaddleX_doc_images/refs/heads/main/images/paddleocr_vl_1_5/layout_skew.jpg" width="800"/>
|
| 97 |
+
</div>
|
| 98 |
+
|
| 99 |
+
|
| 100 |
+
### Screen-photo
|
| 101 |
+
|
| 102 |
+
<div align="center">
|
| 103 |
+
<img src="https://raw.githubusercontent.com/cuicheng01/PaddleX_doc_images/refs/heads/main/images/paddleocr_vl_1_5/layout_screen.jpg" width="800"/>
|
| 104 |
+
</div>
|
| 105 |
+
|
| 106 |
+
|
| 107 |
+
### Curving
|
| 108 |
+
|
| 109 |
+
<div align="center">
|
| 110 |
+
<img src="https://raw.githubusercontent.com/cuicheng01/PaddleX_doc_images/refs/heads/main/images/paddleocr_vl_1_5/layout_curv.jpg" width="800"/>
|
| 111 |
+
</div>
|
| 112 |
+
|
| 113 |
+
|
| 114 |
+
## Citation
|
| 115 |
+
|
| 116 |
+
If you find PP-DocLayoutV3 helpful, feel free to give us a star and citation.
|
| 117 |
+
|
| 118 |
+
```bibtex
|
| 119 |
+
@misc{cui2026rtdoclayoutrealtimeendtoenddocument,
|
| 120 |
+
title={RT-DocLayout: Real-Time End-to-End Document Layout Analysis with Reading Order in the Wild},
|
| 121 |
+
author={Cheng Cui and Tingquan Gao and Xueqing Wang and Changda Zhou and Hongen Liu and Ting Sun and Yubo Zhang and Zelun Zhang and Jiaxuan Liu and Manhui Lin and Yue Zhang and Suyin Liang and Yiqing Xiang and Yi Liu},
|
| 122 |
+
year={2026},
|
| 123 |
+
eprint={2606.23344},
|
| 124 |
+
archivePrefix={arXiv},
|
| 125 |
+
primaryClass={cs.CV},
|
| 126 |
+
url={https://arxiv.org/abs/2606.23344},
|
| 127 |
+
}
|
| 128 |
+
}
|
| 129 |
+
|
| 130 |
+
```
|
config.json
ADDED
|
@@ -0,0 +1,107 @@
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
| 1 |
+
{
|
| 2 |
+
"activation_dropout": 0.0,
|
| 3 |
+
"activation_function": "silu",
|
| 4 |
+
"anchor_image_size": null,
|
| 5 |
+
"architectures": [
|
| 6 |
+
"PPDocLayoutV3ForObjectDetection"
|
| 7 |
+
],
|
| 8 |
+
"attention_dropout": 0.0,
|
| 9 |
+
"backbone": null,
|
| 10 |
+
"backbone_config": {
|
| 11 |
+
"model_type": "hgnet_v2",
|
| 12 |
+
"arch": "L",
|
| 13 |
+
"return_idx": [0, 1, 2, 3],
|
| 14 |
+
"freeze_stem_only": true,
|
| 15 |
+
"freeze_at": 0,
|
| 16 |
+
"freeze_norm": true,
|
| 17 |
+
"lr_mult_list": [0, 0.05, 0.05, 0.05, 0.05],
|
| 18 |
+
"out_features": ["stage1", "stage2", "stage3", "stage4"]
|
| 19 |
+
},
|
| 20 |
+
"backbone_kwargs": null,
|
| 21 |
+
"batch_norm_eps": 1e-05,
|
| 22 |
+
"box_noise_scale": 1.0,
|
| 23 |
+
"d_model": 256,
|
| 24 |
+
"decoder_activation_function": "relu",
|
| 25 |
+
"decoder_attention_heads": 8,
|
| 26 |
+
"decoder_ffn_dim": 1024,
|
| 27 |
+
"decoder_in_channels": [
|
| 28 |
+
256,
|
| 29 |
+
256,
|
| 30 |
+
256
|
| 31 |
+
],
|
| 32 |
+
"decoder_layers": 6,
|
| 33 |
+
"decoder_n_points": 4,
|
| 34 |
+
"disable_custom_kernels": true,
|
| 35 |
+
"dropout": 0.0,
|
| 36 |
+
"encode_proj_layers": [
|
| 37 |
+
2
|
| 38 |
+
],
|
| 39 |
+
"encoder_activation_function": "gelu",
|
| 40 |
+
"encoder_attention_heads": 8,
|
| 41 |
+
"encoder_ffn_dim": 1024,
|
| 42 |
+
"encoder_hidden_dim": 256,
|
| 43 |
+
"encoder_in_channels": [
|
| 44 |
+
512,
|
| 45 |
+
1024,
|
| 46 |
+
2048
|
| 47 |
+
],
|
| 48 |
+
"encoder_layers": 1,
|
| 49 |
+
"eos_coefficient": 0.0001,
|
| 50 |
+
"eval_size": null,
|
| 51 |
+
"feature_strides": [
|
| 52 |
+
8,
|
| 53 |
+
16,
|
| 54 |
+
32
|
| 55 |
+
],
|
| 56 |
+
"hidden_expansion": 1.0,
|
| 57 |
+
"id2label": {
|
| 58 |
+
"0": "abstract",
|
| 59 |
+
"1": "algorithm",
|
| 60 |
+
"2": "aside_text",
|
| 61 |
+
"3": "chart",
|
| 62 |
+
"4": "content",
|
| 63 |
+
"5": "formula",
|
| 64 |
+
"6": "doc_title",
|
| 65 |
+
"7": "figure_title",
|
| 66 |
+
"8": "footer",
|
| 67 |
+
"9": "footer",
|
| 68 |
+
"10": "footnote",
|
| 69 |
+
"11": "formula_number",
|
| 70 |
+
"12": "header",
|
| 71 |
+
"13": "header",
|
| 72 |
+
"14": "image",
|
| 73 |
+
"15": "formula",
|
| 74 |
+
"16": "number",
|
| 75 |
+
"17": "paragraph_title",
|
| 76 |
+
"18": "reference",
|
| 77 |
+
"19": "reference_content",
|
| 78 |
+
"20": "seal",
|
| 79 |
+
"21": "table",
|
| 80 |
+
"22": "text",
|
| 81 |
+
"23": "text",
|
| 82 |
+
"24": "vision_footnote"
|
| 83 |
+
},
|
| 84 |
+
"initializer_range": 0.01,
|
| 85 |
+
"is_encoder_decoder": true,
|
| 86 |
+
"label2id": {},
|
| 87 |
+
"label_noise_ratio": 0.5,
|
| 88 |
+
"layer_norm_eps": 1e-05,
|
| 89 |
+
"learn_initial_query": false,
|
| 90 |
+
"matcher_alpha": 0.25,
|
| 91 |
+
"matcher_bbox_cost": 5.0,
|
| 92 |
+
"matcher_class_cost": 2.0,
|
| 93 |
+
"matcher_gamma": 2.0,
|
| 94 |
+
"matcher_giou_cost": 2.0,
|
| 95 |
+
"model_type": "pp_doclayout_v3",
|
| 96 |
+
"normalize_before": false,
|
| 97 |
+
"num_denoising": 100,
|
| 98 |
+
"num_feature_levels": 3,
|
| 99 |
+
"num_queries": 300,
|
| 100 |
+
"positional_encoding_temperature": 10000,
|
| 101 |
+
"torch_dtype": "float32",
|
| 102 |
+
"use_pretrained_backbone": false,
|
| 103 |
+
"use_timm_backbone": false,
|
| 104 |
+
"global_pointer_head_size": 64,
|
| 105 |
+
"mask_feature_channels": [64, 64],
|
| 106 |
+
"x4_feat_dim": 128
|
| 107 |
+
}
|
inference.yml
ADDED
|
@@ -0,0 +1,100 @@
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
| 1 |
+
mode: paddle
|
| 2 |
+
draw_threshold: 0.5
|
| 3 |
+
metric: COCO
|
| 4 |
+
use_dynamic_shape: false
|
| 5 |
+
Global:
|
| 6 |
+
model_name: PP-DocLayoutV3
|
| 7 |
+
arch: DETR
|
| 8 |
+
min_subgraph_size: 3
|
| 9 |
+
Preprocess:
|
| 10 |
+
- interp: 2
|
| 11 |
+
keep_ratio: false
|
| 12 |
+
target_size:
|
| 13 |
+
- 800
|
| 14 |
+
- 800
|
| 15 |
+
type: Resize
|
| 16 |
+
- mean:
|
| 17 |
+
- 0.0
|
| 18 |
+
- 0.0
|
| 19 |
+
- 0.0
|
| 20 |
+
norm_type: none
|
| 21 |
+
std:
|
| 22 |
+
- 1.0
|
| 23 |
+
- 1.0
|
| 24 |
+
- 1.0
|
| 25 |
+
type: NormalizeImage
|
| 26 |
+
- type: Permute
|
| 27 |
+
label_list:
|
| 28 |
+
- abstract
|
| 29 |
+
- algorithm
|
| 30 |
+
- aside_text
|
| 31 |
+
- chart
|
| 32 |
+
- content
|
| 33 |
+
- display_formula
|
| 34 |
+
- doc_title
|
| 35 |
+
- figure_title
|
| 36 |
+
- footer
|
| 37 |
+
- footer_image
|
| 38 |
+
- footnote
|
| 39 |
+
- formula_number
|
| 40 |
+
- header
|
| 41 |
+
- header_image
|
| 42 |
+
- image
|
| 43 |
+
- inline_formula
|
| 44 |
+
- number
|
| 45 |
+
- paragraph_title
|
| 46 |
+
- reference
|
| 47 |
+
- reference_content
|
| 48 |
+
- seal
|
| 49 |
+
- table
|
| 50 |
+
- text
|
| 51 |
+
- vertical_text
|
| 52 |
+
- vision_footnote
|
| 53 |
+
Hpi:
|
| 54 |
+
backend_configs:
|
| 55 |
+
paddle_infer:
|
| 56 |
+
trt_dynamic_shapes: &id001
|
| 57 |
+
image:
|
| 58 |
+
- - 1
|
| 59 |
+
- 3
|
| 60 |
+
- 800
|
| 61 |
+
- 800
|
| 62 |
+
- - 1
|
| 63 |
+
- 3
|
| 64 |
+
- 800
|
| 65 |
+
- 800
|
| 66 |
+
- - 8
|
| 67 |
+
- 3
|
| 68 |
+
- 800
|
| 69 |
+
- 800
|
| 70 |
+
scale_factor:
|
| 71 |
+
- - 1
|
| 72 |
+
- 2
|
| 73 |
+
- - 1
|
| 74 |
+
- 2
|
| 75 |
+
- - 8
|
| 76 |
+
- 2
|
| 77 |
+
trt_dynamic_shape_input_data:
|
| 78 |
+
scale_factor:
|
| 79 |
+
- - 2
|
| 80 |
+
- 2
|
| 81 |
+
- - 1
|
| 82 |
+
- 1
|
| 83 |
+
- - 0.67
|
| 84 |
+
- 0.67
|
| 85 |
+
- 0.67
|
| 86 |
+
- 0.67
|
| 87 |
+
- 0.67
|
| 88 |
+
- 0.67
|
| 89 |
+
- 0.67
|
| 90 |
+
- 0.67
|
| 91 |
+
- 0.67
|
| 92 |
+
- 0.67
|
| 93 |
+
- 0.67
|
| 94 |
+
- 0.67
|
| 95 |
+
- 0.67
|
| 96 |
+
- 0.67
|
| 97 |
+
- 0.67
|
| 98 |
+
- 0.67
|
| 99 |
+
tensorrt:
|
| 100 |
+
dynamic_shapes: *id001
|
preprocessor_config.json
ADDED
|
@@ -0,0 +1,36 @@
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
| 1 |
+
{
|
| 2 |
+
"_valid_processor_keys": [
|
| 3 |
+
"images",
|
| 4 |
+
"do_resize",
|
| 5 |
+
"size",
|
| 6 |
+
"resample",
|
| 7 |
+
"do_rescale",
|
| 8 |
+
"rescale_factor",
|
| 9 |
+
"do_normalize",
|
| 10 |
+
"image_mean",
|
| 11 |
+
"image_std",
|
| 12 |
+
"return_tensors",
|
| 13 |
+
"data_format",
|
| 14 |
+
"input_data_format"
|
| 15 |
+
],
|
| 16 |
+
"do_normalize": true,
|
| 17 |
+
"do_rescale": true,
|
| 18 |
+
"do_resize": true,
|
| 19 |
+
"image_mean": [
|
| 20 |
+
0,
|
| 21 |
+
0,
|
| 22 |
+
0
|
| 23 |
+
],
|
| 24 |
+
"image_processor_type": "PPDocLayoutV3ImageProcessor",
|
| 25 |
+
"image_std": [
|
| 26 |
+
1,
|
| 27 |
+
1,
|
| 28 |
+
1
|
| 29 |
+
],
|
| 30 |
+
"resample": 3,
|
| 31 |
+
"rescale_factor": 0.00392156862745098,
|
| 32 |
+
"size": {
|
| 33 |
+
"height": 800,
|
| 34 |
+
"width": 800
|
| 35 |
+
}
|
| 36 |
+
}
|