What is KubeDL used for?

KubeDL is a Kubernetes-native controller for running deep learning training and inference workloads, and a CNCF sandbox project. It is a good fit for ML platforms on Kubernetes and managing distributed training jobs.

Is KubeDL open source?

Yes. KubeDL is open source under the Apache-2.0 licence. Its source code is on GitHub at kubedl-io/kubedl and is written mainly in Go.

Yes. KubeDL is open source, so the software itself is free to use.

Can I self-host KubeDL?

Yes. KubeDL can be self-hosted on your own server or infrastructure.

What is KubeDL an alternative to?

KubeDL is an open-source alternative to Amazon SageMaker, Google Vertex AI, Azure Machine Learning and Domino Data Lab. Other open-source alternatives to Amazon SageMaker include MLflow, Backend.AI and vLLM.

Is KubeDL actively maintained?

The most recent commit to KubeDL was on 4 March 2024, and the latest release is v0.5.0, published on 5 December 2022. The project has 534 stars on GitHub.

KubeDL: open-source alternative to Amazon SageMaker

About KubeDL

KubeDL is a Kubernetes-based system that makes it easier and more efficient to run deep learning workloads. It is a CNCF sandbox project and handles both training and inference jobs for frameworks such as TensorFlow, PyTorch and Mars through a single unified controller.

Features include advanced scheduling, cache-based acceleration, metadata persistence, file synchronization and service discovery for training jobs in host networking. It can automatically tune the best configuration for ML model deployment through the related Morphling project, and it can package and deploy models in containers while tracking model lineage natively using Kubernetes custom resources.

KubeDL is written in Go and released under Apache-2.0. A related research paper on Morphling, an auto-configuration approach for cloud-native model serving, is cited in the README. It suits platform teams that run machine learning jobs on shared Kubernetes clusters.

Key features

Unified controller for training and inference
Supports TensorFlow, PyTorch and Mars jobs
Advanced job scheduling
Model packaging with lineage tracking
Auto-tuning of model deployment configs
Cache acceleration and file sync

Good fit for

→ML platforms on Kubernetes
→Managing distributed training jobs

Built with: Go; Kubernetes
Tags: kubernetes; deep-learning; machine-learning; scheduling; mlops; model-serving; golang; cncf

KubeDL: questions and answers

What is KubeDL used for?: KubeDL is a Kubernetes-native controller for running deep learning training and inference workloads, and a CNCF sandbox project. It is a good fit for ML platforms on Kubernetes and managing distributed training jobs.
Is KubeDL open source?: Yes. KubeDL is open source under the Apache-2.0 licence. Its source code is on GitHub at kubedl-io/kubedl and is written mainly in Go.
Is KubeDL free?: Yes. KubeDL is open source, so the software itself is free to use.
Can I self-host KubeDL?: Yes. KubeDL can be self-hosted on your own server or infrastructure.
What is KubeDL an alternative to?: KubeDL is an open-source alternative to Amazon SageMaker, Google Vertex AI, Azure Machine Learning and Domino Data Lab. Other open-source alternatives to Amazon SageMaker include MLflow, Backend.AI and vLLM.
Is KubeDL actively maintained?: The most recent commit to KubeDL was on 4 March 2024, and the latest release is v0.5.0, published on 5 December 2022. The project has 534 stars on GitHub.