A demonstration of albumentations library.
The whole demo is a jupyter notebook at notebooks/demo.ipynb
.
Define environmental variables:
$USER_ID
and$GROUP_ID
, equal to your userid -u
andid -g
respectively,$BASE_IMAGE
equal toubuntu:20.04
,$PYTHON_VERSION
equal to3.9
.
You may store these variables in .env
file in your project root along with any secrets and tokens
that your project needs. Just make sure it is not tracked by git.
Build an image and start a container, it will start a jupyter server automatically:
docker compose up --build alb_demo
- docs: https://albumentations.ai/docs/
- benchmarks: https://github.com/albumentations-team/albumentations/blob/master/README.md#benchmarking-results
- more examples: https://github.com/albumentations-team/albumentations_examples
- AutoAlbument: https://github.com/albumentations-team/autoalbument
- Faster AutoAugment (base paper of AutoAlbument): https://arxiv.org/pdf/1911.06987.pdf