YOLO11n DocLayNet
compact document layout analysis model
2 min readupdated
YOLO11n DocLayNet is a lightweight AI model built on Ultralytics YOLO11 (packing just 2.6M parameters) designed to detect structural elements across document pages. It identifies text, tables, headings, formulas, images, and other layout blocks in PDFs and scans. Despite its tiny footprint, it outperforms the larger standard YOLO11s model on the DocLayNet benchmark. Combining high accuracy with modest hardware requirements, it is a solid fit for developers building automated data extraction and document parsing pipelines.
Key features
- Compact YOLO11n architecture with 2.6M parameters that outperforms YOLO11s.
- Supports 11 layout categories, including text, tables, headings, formulas, and footers.
- High accuracy of 0.760 mAP50-95 on the DocLayNet test dataset.
- Optimized for processing document pages at 1024px resolution.
What you can do
- Automatically parse layouts of research papers, financial reports, and patents.
- Integrate the model into OCR and PDF content extraction pipelines.
- Export weights to ONNX format for faster inference.
Example
from huggingface_hub import hf_hub_download
from ultralytics import YOLO
model = YOLO(hf_hub_download("avbiswas/yolo11n-doclaynet", "yolo11n-doclaynet.pt"))
results = model.predict("page.png", imgsz=1024, conf=0.25)
for box in results[0].boxes:
print(model.names[int(box.cls)], float(box.conf), box.xyxy.tolist())
Read more
DetailsBenchmark: 0.760 mAP50-95 across 2,500 DocLayNet pages.
Training: 22 epochs on a single A100 using a 69k-page dataset.
Limitations: slides, forms, and handwritten text are not supported.
License: AGPL-3.0 (inherited from Ultralytics).










Comments