Hugging Peers

anydoclayout

☆ Save
by anyforge ocr-recognition Pytorch Apache License 2.0
389 downloads 0 likes on ModelScope updated

Model card

anydoclayout

docs layout detection

Static Badge 

    _                ____             _                            _
   / \   _ __  _   _|  _ \  ___   ___| |    __ _ _   _  ___  _   _| |_
  / _ \ | '_ \| | | | | | |/ _ \ / __| |   / _` | | | |/ _ \| | | | __|
 / ___ \| | | | |_| | |_| | (_) | (__| |__| (_| | |_| | (_) | |_| | |_
/_/   \_\_| |_|\__, |____/ \___/ \___|_____\__,_|\__, |\___/ \__,_|\__|
               |___/                             |___/

训练数据

1. 标签

{0: 'header',
 1: 'title',
 2: 'text',
 3: 'table',
 4: 'figure',
 5: 'formula',
 6: 'footer',
 7: 'pagenum'}

2. 数据集

  • train: 841862 (opendata: 667426, selfgen: 174436)
  • eval: 5500
  • imgsize：1280

3. 评估结果

Class     Images  Instances      Box(P          R)
   all       5500      52274      0.921      0.897
header       1461       2337       0.92      0.878
 title       2308       5473      0.896      0.893 
  text       4149      34156      0.937      0.927
 table       1476       1913      0.946      0.958
figure       1842       3343       0.94       0.94
formula        735       1506      0.881      0.876
footer        745       1157      0.909      0.781
pagenum       2164       2389      0.938      0.919

如果您想要数据集

  • 联系邮箱：christnowx@qq.com

SDK下载

#安装ModelScope
pip install modelscope



#SDK模型下载
from pathlib import Path
from ultralytics import YOLO
from modelscope import snapshot_download
model_dir = snapshot_download('anyforge/anydoclayout')

modelfile = Path(model_dir).joinpath('models/anydoclayout-yolo11s-imgsz1280.pt')
model = YOLO(modelfile)
res = model.predict('your img file', imgsz = 1280)

Buy me a coffee

  • 微信(WeChat)

如果您是本模型的贡献者，我们邀请您根据模型贡献文档，及时完善模型卡片内容。