Pablo

Refrescos

Classification

Refrescos Image Dataset

v31

2024-04-23 12:44pm

Generated on Apr 23, 2024

Dataset Split

Train Set 96%
14082Images
Valid Set %
0Images
Test Set 4%
649Images

Preprocessing

Auto-Orient: Applied
Resize: Stretch to 640x640

Augmentations

Outputs per training example: 3
Crop: 0% Minimum Zoom, 5% Maximum Zoom
Rotation: Between -3° and +3°
Shear: ±14° Horizontal, ±10° Vertical
Grayscale: Apply to 4% of images
Saturation: Between -10% and +10%
Brightness: Between -9% and +9%
Exposure: Between -10% and +10%
Blur: Up to 1.3px
Noise: Up to 0.58% of pixels