Object Detection
ultralytics
English
German
yolo
yolov8
invoice
document-layout
document-understanding
ocr-prep
invoice-extraction
Eval Results (legacy)
Instructions to use AvoCahDoe/invoice-layout-yolov8n with libraries, inference providers, notebooks, and local apps. Follow these links to get started.
- Libraries
- ultralytics
How to use AvoCahDoe/invoice-layout-yolov8n with ultralytics:
from ultralytics import YOLOvv8 model = YOLOvv8.from_pretrained("AvoCahDoe/invoice-layout-yolov8n") source = 'http://images.cocodataset.org/val2017/000000039769.jpg' model.predict(source=source, save=True) - Notebooks
- Google Colab
- Kaggle
| # Multi-model YOLO comparison training (optimized for RTX 4070 Laptop) | |
| epochs: 100 | |
| patience: 25 | |
| device: 0 | |
| workers: 0 | |
| cache: disk | |
| project: runs/comparison | |
| # Optimizer | |
| optimizer: AdamW | |
| cos_lr: true | |
| lr0: 0.01 | |
| lrf: 0.01 | |
| momentum: 0.937 | |
| weight_decay: 0.0005 | |
| warmup_epochs: 3.0 | |
| # Loss weights | |
| box: 7.5 | |
| cls: 0.5 | |
| dfl: 1.5 | |
| # Geometric augmentation | |
| mosaic: 1.0 | |
| mixup: 0.15 | |
| copy_paste: 0.1 | |
| scale: 0.5 | |
| perspective: 0.0005 | |
| degrees: 3.0 | |
| translate: 0.1 | |
| fliplr: 0.0 | |
| flipud: 0.0 | |
| close_mosaic: 10 | |
| # Visual augmentation | |
| hsv_h: 0.015 | |
| hsv_s: 0.7 | |
| hsv_v: 0.4 | |
| erasing: 0.4 | |
| # Per-model overrides β train smallest first (n β s β m β 11m β v8x β 11x) | |
| models: | |
| yolov8n: | |
| weights: yolov8n.pt | |
| batch: 8 | |
| imgsz: 1024 | |
| yolov8s: | |
| weights: yolov8s.pt | |
| batch: 8 | |
| imgsz: 1024 | |
| yolov8m: | |
| weights: yolov8m.pt | |
| batch: 6 | |
| imgsz: 1024 | |
| yolo11m: | |
| weights: yolo11m.pt | |
| batch: 6 | |
| imgsz: 1024 | |
| yolov8x: | |
| weights: yolov8x.pt | |
| batch: 2 | |
| imgsz: 1024 | |
| yolo11x: | |
| weights: yolo11x.pt | |
| batch: 2 | |
| imgsz: 1024 | |