Dataset Versions

v1

2024-07-22 6:13am

Generated on Jul 22, 2024

Dataset Split

Train Set 87%
2541Images
Valid Set 8%
246Images
Test Set 4%
120Images

Preprocessing

Auto-Orient: Applied
Resize: Stretch to 640x640
Modify Classes: 34 remapped, 2 dropped

Augmentations

Outputs per training example: 3
Crop: 0% Minimum Zoom, 20% Maximum Zoom
Rotation: Between -15° and +15°
Grayscale: Apply to 15% of images
Hue: Between -15° and +15°
Blur: Up to 2.5px