Dataset Versions

v4

2023-03-16 11:52am

Generated on Mar 16, 2023

Dataset Split

Train Set 99%
1458Images
Valid Set 1%
10Images
Test Set 0%
3Images

Preprocessing

Auto-Orient: Applied
Resize: Stretch to 640x640
Modify Classes: 11 remapped, 5 dropped

Augmentations

Outputs per training example: 3
Rotation: Between -5° and +5°
Shear: ±5° Horizontal, ±5° Vertical
Grayscale: Apply to 20% of images