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-yolov8m with libraries, inference providers, notebooks, and local apps. Follow these links to get started.
- Libraries
- ultralytics
How to use AvoCahDoe/invoice-layout-yolov8m with ultralytics:
from ultralytics import YOLOvv8 model = YOLOvv8.from_pretrained("AvoCahDoe/invoice-layout-yolov8m") source = 'http://images.cocodataset.org/val2017/000000039769.jpg' model.predict(source=source, save=True) - Notebooks
- Google Colab
- Kaggle
File size: 1,099 Bytes
45179c8 | 1 2 3 4 5 6 7 8 9 10 11 12 13 14 15 16 17 18 19 20 21 22 23 24 25 26 27 28 29 30 31 32 33 34 35 36 37 38 39 40 41 42 43 44 45 46 47 48 49 50 51 52 53 54 55 56 57 58 59 60 61 62 63 64 65 66 67 | # 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 — train smallest first (n → s → m → 11m → v8x → 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
|