Albumentations is a fast and flexible library for image augmentations that provides a wide range of image transformation operations. It is also an easy-to-use wrapper around other augmentation libraries. The library is designed to be efficient and flexible, allowing users to apply various transformations to images for different computer vision tasks. The paper presents Albumentations as a solution that is faster than other commonly used image augmentation tools on most commonly used image transformations. The paper discusses the importance of image augmentation in computer vision, particularly in combating overfitting in deep convolutional neural networks. It highlights the need for flexible and rich image augmentation tools that can handle a wide range of transformations for different tasks. The paper also discusses the performance of Albumentations on various computer vision tasks, including street view image detection, satellite and aerial imagery analysis, and biomedical image analysis. In the case of biomedical image analysis, image augmentations are particularly useful due to the typically limited amount of available labeled data. Albumentations provides a variety of transformations that can be applied to medical images, including color transformations, grid distortion, and elastic transforms, which are useful for handling non-rigid structures in medical imaging. The paper also presents benchmark results comparing the performance of Albumentations with other image augmentation tools. The results show that Albumentations is consistently faster than other tools for most image transformations. The source code for Albumentations is publicly available online at https://github.com/albu/albumentations.Albumentations is a fast and flexible library for image augmentations that provides a wide range of image transformation operations. It is also an easy-to-use wrapper around other augmentation libraries. The library is designed to be efficient and flexible, allowing users to apply various transformations to images for different computer vision tasks. The paper presents Albumentations as a solution that is faster than other commonly used image augmentation tools on most commonly used image transformations. The paper discusses the importance of image augmentation in computer vision, particularly in combating overfitting in deep convolutional neural networks. It highlights the need for flexible and rich image augmentation tools that can handle a wide range of transformations for different tasks. The paper also discusses the performance of Albumentations on various computer vision tasks, including street view image detection, satellite and aerial imagery analysis, and biomedical image analysis. In the case of biomedical image analysis, image augmentations are particularly useful due to the typically limited amount of available labeled data. Albumentations provides a variety of transformations that can be applied to medical images, including color transformations, grid distortion, and elastic transforms, which are useful for handling non-rigid structures in medical imaging. The paper also presents benchmark results comparing the performance of Albumentations with other image augmentation tools. The results show that Albumentations is consistently faster than other tools for most image transformations. The source code for Albumentations is publicly available online at https://github.com/albu/albumentations.