A collection of portable workflows, automation actions and reusable artifacts for AI and ML systems in the CK format:
- CK modules with automation actions: [dev] [stable]
- CK program workflows: [dev] [CK platform]
- CK meta packages: [dev] [CK platform]
- CK software detection: [dev] [CK platform]
- CK datasets: [dev] [CK platform]
- CK adaptive containers: [dev] [CK platform]
- CK OS: [dev] [CK platform]
- CK MLPerf system descriptions: [dev] [CK platform]
- CK MLPerf benchmark CMD generators: [dev] [CK platform]
All CK components are available at the CK portal similar to PyPI.
Install the CK framework as described here.
Pull this repository:
ck pull repo:ai
Try portable AI/ML workflows, program pipelines and adaptive CK containers. Note that you do not need to pull other repositories anymore since all the components are aggregated here.
Check public dashboards with reproduced results from research papers.
See real use cases from the community: MLPerf, Arm, General Motors, IBM, Raspberry Pi foundation, ACM, dividiti and others.
Read about the CK concept and format.
We have prepared a CK container with all CK components from this repository: [Docker], [CK meta]
You can start it as follows:
docker run --rm -it ctuning/ck-ai:ubuntu-20.04
You can then prepare and run these portable AI/ML workflows and program pipelines.
BSD 3-clause. We are discussing the possibility to relicense the CK framework to Apache 2.0.
Please contribute as described here and submit your PRs here.
We would like to thank all collaborators for their support, fruitful discussions, and useful feedback! See more acknowledgments in the CK journal article.
Don't hesitate to get in touch with the CK community if you have questions or comments.