Experimental functionality for the Albumentaions library.
Project description
Albumentations Experimental
The Albumentations Experimental library provides experimental and cutting edge augmentation techniques on top Albumentations.
Why a separate library
Albumentations provides stable and well-tested interfaces for performing augmentations. We don't want to pollute the library with features that may be prone to rapid changes in interfaces and behavior since they could break users' pipelines. But we also want to implement new, experimental features and see whether they will be useful.
So we created Albumentations Experimental, a library that will help us to iterate faster and remove the need for striving for backward compatibility and rigorous testing.
Beware, that each new version of Albumentations Experimental may contain backward-incompatible changes both in interfaces and behavior.
When features in Albumentations Experimental are mature enough, we will port them to the main library with all our usual policies such as rigorous testing, extensive documentation, and stable behavior.
Installation
Albumentations Experimental requires Python 3.5 or higher.
Install the latest stable version from PyPI
pip install -U albumentations_experimental
Install the latest version from the master's branch on GitHub
pip install -U git+https://github.com/albumentations-team/albumentations-experimental
Usage
Import augmentations from the library:
from albumentations_experimental import FlipSymmetricKeypoints
Documentation
Documentation is available at https://albumentations.ai/docs/experimental/overview/
List of augmentations and their supported targets
Spatial-level transforms
Transform | Image | Masks | BBoxes | Keypoints |
---|---|---|---|---|
FlipSymmetricKeypoints | ✓ | ✓ | ✓ | ✓ |
HorizontalFlipSymmetricKeypoints | ✓ | ✓ | ✓ | ✓ |
TransposeSymmetricKeypoints | ✓ | ✓ | ✓ | ✓ |
VerticalFlipSymmetricKeypoints | ✓ | ✓ | ✓ | ✓ |
Project details
Release history Release notifications | RSS feed
Download files
Download the file for your platform. If you're not sure which to choose, learn more about installing packages.
Source Distribution
Built Distribution
Hashes for albumentations_experimental-0.0.1.tar.gz
Algorithm | Hash digest | |
---|---|---|
SHA256 | 4093aa7df6c20d7b42806af9e962fccb83ad3f64dbb9c2b7c176e30ff8f2933c |
|
MD5 | 76004ae045a2cf45ea08b796d7f31472 |
|
BLAKE2b-256 | 8f94e7de86cee07f1360df1a014d9b480be833911d75bdee543cf213d2580aba |
Hashes for albumentations_experimental-0.0.1-py3-none-any.whl
Algorithm | Hash digest | |
---|---|---|
SHA256 | 639a2222af4941435eebc1c0350af6eb70c10b4efa16c5c09db69492c69dfbac |
|
MD5 | a731502c3d5bab36599759c948efa6b0 |
|
BLAKE2b-256 | 64315b71e25c44e90b3fcdc99e31f99e5e7af575482a8ad95c496b8b25726709 |