Dataset Versions

v4

2023-01-17 9-33am - Even Image Number for Vision Transformer

Generated on Jan 17, 2023

Dataset Split

Train Set 86%
1650Images
Valid Set 4%
80Images
Test Set 10%
190Images

Preprocessing

Auto-Orient: Applied
Resize: Stretch to 720x1280

Augmentations

Outputs per training example: 3
Rotation: Between -10° and +10°
Brightness: Between -25% and +25%
Blur: Up to 1px
Noise: Up to 8% of pixels

Similar Projects

See More