Dataset Versions

v3

2024-10-29 10:49am

Generated on Oct 29, 2024

Dataset Split

Train Set 88%
2688Images
Valid Set 8%
256Images
Test Set 4%
128Images

Preprocessing

Auto-Orient: Applied

Augmentations

Outputs per training example: 3
Flip: Horizontal, Vertical
Hue: Between -180° and +180°
Saturation: Between -34% and +34%
Brightness: Between -62% and +62%
Exposure: Between -7% and +7%
Blur: Up to 1px