/kubedl

Run your deep learning workloads on Kubernetes more easily and efficiently.

Primary LanguageGoApache License 2.0Apache-2.0

License KubeDL Action Status FOSSA Status CII Best Practices


KubeDL enables deep learning workloads to run on Kubernetes more easily and efficiently.

KubeDL is a CNCF sandbox project.


Features

  • Support training and inferences workloads (Tensorflow, Pytorch. Mars etc.)in a single unified controller. Features include advanced scheduling, acceleration using cache, metadata persistentcy, file sync, enable service discovery for training in host network etc.
  • Automatically tunes the best configurations for ML model deployment. - Morphling Github
  • Package and deploy ML Model in container and track the model lineage natively with Kubernentes CRD.

Check the website: https://kubedl.io


Getting Involved

Platform Purpose Estimated Response Time
DingTalk For discussions about development and questions about usage. < 1 day
Github Issues For reporting bugs and filing feature requests. < 2 days
E-Mail(cncf-kubedl-maintainers@lists.cncf.io) For discussing specific topics or ask for help from community members/maintainers. < 3 days

Publications

Morphling: Fast, Near-Optimal Auto-Configuration for Cloud-Native Model Serving. ACM Socc 2021link

License

FOSSA Status