Dataset Versions

v13

best so far_640_3x_allmaped to trash

Generated on Jan 10, 2024

Dataset Split

Train Set 87%
3147Images
Valid Set 8%
299Images
Test Set 4%
152Images

Preprocessing

Auto-Orient: Applied
Resize: Stretch to 640x640
Modify Classes: 7 remapped, 52 dropped

Augmentations

Outputs per training example: 3
Flip: Horizontal
90° Rotate: Clockwise, Counter-Clockwise, Upside Down
Crop: 0% Minimum Zoom, 15% Maximum Zoom
Rotation: Between -8° and +8°
Shear: ±3° Horizontal, ±3° Vertical
Grayscale: Apply to 5% of images
Hue: Between -15° and +15°
Saturation: Between -20% and +20%
Brightness: Between -20% and +20%
Exposure: Between -10% and +10%
Blur: Up to 0.75px