albumentations - A fast and framework agnostic image augmentation library
A fast and flexible Python library for image augmentation in computer vision tasks like classification, segmentation, and object detection.
What is albumentations - A fast and framework agnostic image augmentation library?
Albumentations is a Python library for fast and flexible image augmentation, designed to enhance training datasets for deep learning and computer vision models. It creates new training samples by applying various transformations to existing images, helping improve model robustness and performance across tasks like classification, segmentation, and object detection.
Machine learning engineers, data scientists, and researchers working on computer vision projects who need efficient and comprehensive image augmentation to improve model training and accuracy.
Developers choose Albumentations for its benchmark-leading speed, unified API that supports multiple data types and CV tasks, and extensive library of over 70 high-quality augmentations, all built by experts with competition experience.
Overview
Fast and flexible image augmentation library. Paper about the library: https://www.mdpi.com/2078-2489/11/2/125
Use Cases
Best For
- Augmenting image datasets for deep learning model training
- Applying transformations to images, masks, bounding boxes, and keypoints simultaneously
- Improving model robustness in computer vision competitions
- Enhancing medical imaging datasets with specialized augmentations
- Speeding up data augmentation pipelines in production environments
- Working with 3D volumetric data and masks in medical or scientific imaging
Not Ideal For
- Projects requiring ongoing bug fixes, updates, or compatibility with newer Python/PyTorch versions
- Teams using permissive open-source licenses (e.g., MIT, Apache) who want to avoid AGPL copyleft restrictions
- New initiatives that prioritize long-term support and active community development over legacy performance
Pros & Cons
Pros
Consistently outperforms competitors like imgaug and torchvision in benchmarks, processing thousands of images per second on CPU threads, as shown in the performance comparison table.
Supports all major computer vision tasks—classification, segmentation, object detection, and pose estimation—with a single interface for images, masks, bounding boxes, and keypoints, simplifying pipeline development.
Offers over 70 high-quality transforms, including pixel-level and spatial-level augmentations, validated by Kaggle experts and documented with examples for various use cases.
Works with PyTorch, TensorFlow, and is part of the PyTorch ecosystem, making it easy to integrate into existing deep learning workflows without major refactoring.
Cons
As of June 2025, no further updates, bug fixes, or compatibility patches will be provided, leaving users vulnerable to issues with evolving dependencies and unsupported environments.
The successor, AlbumentationsX, uses AGPL-3.0 licensing, which is incompatible with permissive licenses like MIT or Apache, forcing commercial licensing for many open-source or proprietary projects.
Without active development, the library may fall behind in new features and optimizations, and community support is redirected to AlbumentationsX, leaving legacy users isolated.
Frequently Asked Questions
Related Projects
Tensorflow - Open source software library for numerical computation using data flow graphsAn Open Source Machine Learning Framework for Everyone
PyTorch - Tensors and Dynamic neural networks in Python with strong GPU accelerationTensors and Dynamic neural networks in Python with strong GPU acceleration
OpenCVOpen Source Computer Vision Library
face_recognitionThe world's simplest facial recognition api for Python and the command line
Found a gem we're missing?
Open-Awesome is built by the community, for the community. Submit a project, suggest an awesome list, or help improve the catalog on GitHub.