Albumentations is a fast and flexible library for image augmentations, designed to increase the size and diversity of labeled training sets in computer vision tasks. It offers a wide range of image transformation operations, including color, contrast, brightness, and more, which are essential for various tasks such as image classification, segmentation, and detection. The library is also a wrapper around other augmentation libraries, making it easy to use. Albumentations is faster than commonly used tools for most image transformations, as demonstrated through benchmarks. The source code is publicly available on GitHub. The paper provides examples of image augmentations for different computer vision tasks, highlighting its versatility and efficiency.Albumentations is a fast and flexible library for image augmentations, designed to increase the size and diversity of labeled training sets in computer vision tasks. It offers a wide range of image transformation operations, including color, contrast, brightness, and more, which are essential for various tasks such as image classification, segmentation, and detection. The library is also a wrapper around other augmentation libraries, making it easy to use. Albumentations is faster than commonly used tools for most image transformations, as demonstrated through benchmarks. The source code is publicly available on GitHub. The paper provides examples of image augmentations for different computer vision tasks, highlighting its versatility and efficiency.