Back to Gap Map Infrastructure / Classic ML & computer vision

Albumentations

Albumentations
open source / Overall score: 1.6

Albumentations is a Python library of image augmentations for training computer-vision models, with more than 70 transforms that apply consistently to images, masks, bounding boxes and keypoints. Its current line, AlbumentationsX, continues the MIT-licensed albumentations package, which is no longer maintained, under AGPL-3.0, and Albumentations LLC also sells a commercial license to the same code.

Openness

5 high confidence
5.0
license
AGPL-3.0 (OSI; AGPL-3.0-only)
source
public (github.com/albumentations-team/AlbumentationsX)
core features withheld
no — the commercial license grants other terms for the same library; nothing is withheld

AlbumentationsX is AGPL-3.0 and builds from the public repository. Albumentations LLC also sells a commercial license, but it only lifts the AGPL's source-sharing terms for the same library; no feature is kept back for paying customers.

Adoption

2 high confidence
2.0

Measured on monthly PyPI downloads of the albumentationsx package. The unmaintained MIT-licensed albumentations package still draws its own downloads, which are not counted here.

Capability

1 high confidence
1.0

Albumentations supplies the augmentation step of a vision training pipeline and fits no model of its own, as OpenCV supplies image operations and leaves its learners to a separate contrib repository. Training is left to the framework that consumes its output.

Verified 2026-11-12, except some scores, still awaiting re-confirmation