We are pleased to announce that DKube and VMware teams are working closely together to bring Kubeflow and MLFlow based MLOps to clients needing on-prem and multi-cloud implementations of their AI projects. In addition, the two companies have further enhanced customer experience by adding plug-ins for clients to use their HPC clusters for compute and GPU capacity, as well for federated learning on Kubeflow (FATE). The solution and use cases will be showcased at VMware Explore 2023 in Las Vegas from Aug 21-24. Please visit VMware Booth# 305, AI Showcase to learn more about these solutions.
DKube is an MLOps platform from One Convergence Inc. based on the popular Kubeflow architecture of AI model lifecycle management. DKube is further augmented with best of MLFlow, bringing together powerful components and enhancing them with best-in-class capabilities, such as:
All of this functionality is integrated into a flexible, UI-based workflow that is intuitive enough to allow team members to collaborate on common AI/ML projects within hours of starting the installation. Given the open source foundations of Kubeflow and MLFlow, customers get best-in-class innovations at much lower cost than other closed source MLOps products in the industry.
DKube has been optimized for VMware vSphere and its Kubernetes distribution Tanzu or other distributions like Rancher or Openshift.
Together the two companies have also brought forth the federated learning with Kubeflow with VMware’s FATE contribution to Kubeflow which is now integrated into DKube: https://octo.vmware.com/streamlining-federated-learning-workflows-with-mlops-platform/
The two companies are bringing the IT and data science communities together at enterprise clients so that their AI/ML projects can successfully be deployed and maintained in production. As an example, the two companies have partnered to showcase an enterprise-wide deployment of AI/ML services on vSphere for a major client in the financial services industry.
“Many enterprise clients want to bring AI to where their data is. Data security is paramount today, and we understand this priority. DKube is a unique MLOps solution that can operate on-premises, via any cloud, or across multi-clouds. DKube brings your favorite open-source AI tool sets to where your data is. With DKube customers have a choice to be independent from a specific cloud.” said Prasad Vellanki, CEO of One Convergence Inc.
"We have been collaborating with the DKube team for over a year to make Kubeflow enterprise grade with Tanzu while augmenting it with federated learning (Kubeflow FATE) or integrating with HPC clusters." said Alan Ren, senior director, engineering & ecosystems, VMware AI Labs.
VMware, Explore, Tanzu, and vSphere are registered trademarks or trademarks of VMware, Inc. and its subsidiaries in the United States and other jurisdictions.
How to launch a single model training or data preprocessing run through the DKube UI while passing new parameters to pass to the training program via the UI and designated environment variables. Learn how to track the performance characteristics and the lineage of the run with the dataset and model versions used or created in the process.
How to set-up a Jupyter Notebook or R-Studio IDE to import or write your program code including Kubeflow pipeline DSL. Learn about the supported ML frameworks or custom image import into your notebook.
Machine learning has been traditionally focused on the important initial training phase. This ensures that a model will provide the best outcomes for the expected live data stream - at least for a while. However, it is well known within the machine learning community that however good a job you do at training a model, your results will