Benchmarks

Albumentations is designed to be fast and efficient. We regularly benchmark it against other augmentation libraries.

Available Benchmarks

• Image Benchmarks - Performance comparison for standard 3-channel image augmentations
• Multichannel Benchmarks - Performance comparison for 9-channel image augmentations (satellite, medical, scientific)
• Video Benchmarks - Performance comparison for video augmentations (CPU vs GPU)
• Albumentations vs PIL/Pillow - Comparison guide from the docs library
• AlbumentationsX vs PIL (RGB table) - Interactive RGB throughput table and migration snippets
• Albumentations vs Torchvision - Comparison guide from the docs library
• AlbumentationsX vs Torchvision (tables) - RGB, multichannel, and video throughput tables plus migration snippets
• Albumentations vs Kornia - Comparison guide from the docs library
• AlbumentationsX vs Kornia (tables) - RGB, multichannel, and video throughput tables plus migration snippets