DocLayout YOLO¶

Here we use yolo layout to detect analyze, its a example, and how it will ideally run.

In [2]:

from omnidocs.tasks.layout_analysis.extractors.doc_layout_yolo import YOLOLayoutDetector

from omnidocs.tasks.layout_analysis.extractors.doc_layout_yolo import YOLOLayoutDetector

c:\Users\laxma\OneDrive\Desktop\CogLab\11-07-2025\Omnidocs\new\Lib\site-packages\transformers\utils\hub.py:111: FutureWarning: Using `TRANSFORMERS_CACHE` is deprecated and will be removed in v5 of Transformers. Use `HF_HOME` instead.
  warnings.warn(

In [5]:

detector = YOLOLayoutDetector(show_log=True)
image_path = "assets/news_paper.png"
annotated_image, layout_output = detector.detect(image_path)
print(f"Detected {len(layout_output.bboxes)} elements")

detector = YOLOLayoutDetector(show_log=True) image_path = "assets/news_paper.png" annotated_image, layout_output = detector.detect(image_path) print(f"Detected {len(layout_output.bboxes)} elements")

INFO     [timestamp]2025-07-29 20:19:30[/] |                                                                       
         [logger.name]omnidocs.tasks.layout_analysis.extractors.doc_layout_yolo[/] |                               
         [function]doc_layout_yolo.py:53[/] | [info]Initializing YOLOLayoutDetector[/]                             

INFO     [timestamp]2025-07-29 20:19:30[/] |                                                                       
         [logger.name]omnidocs.tasks.layout_analysis.extractors.doc_layout_yolo[/] |                               
         [function]doc_layout_yolo.py:53[/] | [info]Initializing YOLOLayoutDetector[/]                             

INFO     [timestamp]2025-07-29 20:19:30[/] |                                                                       
         [logger.name]omnidocs.tasks.layout_analysis.extractors.doc_layout_yolo[/] |                               
         [function]doc_layout_yolo.py:58[/] | [info]Using device: cuda[/]                                          

INFO     [timestamp]2025-07-29 20:19:30[/] |                                                                       
         [logger.name]omnidocs.tasks.layout_analysis.extractors.doc_layout_yolo[/] |                               
         [function]doc_layout_yolo.py:58[/] | [info]Using device: cuda[/]                                          

INFO     [timestamp]2025-07-29 20:19:30[/] |                                                                       
         [logger.name]omnidocs.tasks.layout_analysis.extractors.doc_layout_yolo[/] |                               
         [function]doc_layout_yolo.py:66[/] | [info]Model directory:                                               
         C:\Users\laxma\OneDrive\Desktop\CogLab\11-07-2025\Omnidocs\omnidocs\models\yolo_layout\juliozhao_DocLayout
         -YOLO-DocStructBench[/]                                                                                   

INFO     [timestamp]2025-07-29 20:19:30[/] |                                                                       
         [logger.name]omnidocs.tasks.layout_analysis.extractors.doc_layout_yolo[/] |                               
         [function]doc_layout_yolo.py:66[/] | [info]Model directory:                                               
         C:\Users\laxma\OneDrive\Desktop\CogLab\11-07-2025\Omnidocs\omnidocs\models\yolo_layout\juliozhao_DocLayout
         -YOLO-DocStructBench[/]                                                                                   

INFO     [timestamp]2025-07-29 20:19:30[/] |                                                                       
         [logger.name]omnidocs.tasks.layout_analysis.extractors.doc_layout_yolo[/] |                               
         [function]doc_layout_yolo.py:141[/] | [info]Loading YOLO model from                                       
         C:\Users\laxma\OneDrive\Desktop\CogLab\11-07-2025\Omnidocs\omnidocs\models\yolo_layout\juliozhao_DocLayout
         -YOLO-DocStructBench\doclayout_yolo_docstructbench_imgsz1024.pt[/]                                        

INFO     [timestamp]2025-07-29 20:19:30[/] |                                                                       
         [logger.name]omnidocs.tasks.layout_analysis.extractors.doc_layout_yolo[/] |                               
         [function]doc_layout_yolo.py:141[/] | [info]Loading YOLO model from                                       
         C:\Users\laxma\OneDrive\Desktop\CogLab\11-07-2025\Omnidocs\omnidocs\models\yolo_layout\juliozhao_DocLayout
         -YOLO-DocStructBench\doclayout_yolo_docstructbench_imgsz1024.pt[/]                                        

SUCCESS  [timestamp]2025-07-29 20:19:30[/] |                                                                       
         [logger.name]omnidocs.tasks.layout_analysis.extractors.doc_layout_yolo[/] | [function]logging.py:196[/] | 
         [success]YOLO model loaded successfully on cuda[/]                                                        

SUCCESS  [timestamp]2025-07-29 20:19:30[/] |                                                                       
         [logger.name]omnidocs.tasks.layout_analysis.extractors.doc_layout_yolo[/] | [function]logging.py:196[/] | 
         [success]YOLO model loaded successfully on cuda[/]                                                        

SUCCESS  [timestamp]2025-07-29 20:19:30[/] |                                                                       
         [logger.name]omnidocs.tasks.layout_analysis.extractors.doc_layout_yolo[/] | [function]logging.py:196[/] | 
         [success]Model initialized successfully[/]                                                                

SUCCESS  [timestamp]2025-07-29 20:19:30[/] |                                                                       
         [logger.name]omnidocs.tasks.layout_analysis.extractors.doc_layout_yolo[/] | [function]logging.py:196[/] | 
         [success]Model initialized successfully[/]                                                                

0: 1024x672 17 titles, 59 plain texts, 2 abandons, 7 figures, 1 figure_caption, 274.8ms
Speed: 12.0ms preprocess, 274.8ms inference, 219.8ms postprocess per image at shape (1, 3, 1024, 672)

INFO     [timestamp]2025-07-29 20:19:32[/] |                                                                       
         [logger.name]omnidocs.tasks.layout_analysis.extractors.doc_layout_yolo[/] | [function]logging.py:150[/] | 
         [info]detect completed in 2.25s[/]                                                                        

INFO     [timestamp]2025-07-29 20:19:32[/] |                                                                       
         [logger.name]omnidocs.tasks.layout_analysis.extractors.doc_layout_yolo[/] | [function]logging.py:150[/] | 
         [info]detect completed in 2.25s[/]                                                                        

Detected 84 elements

In [ ]:

output_path = "output/yolo_result.png"
detector.visualize((annotated_image, layout_output), output_path)
print(f"Saved visualization to {output_path}")

output_path = "output/yolo_result.png" detector.visualize((annotated_image, layout_output), output_path) print(f"Saved visualization to {output_path}")

Saved visualization & json to output/yolo_result.png

In [7]:

from IPython.display import Image, display
# Display in notebook
display(Image(output_path))

from IPython.display import Image, display # Display in notebook display(Image(output_path))

In [ ]: