Intel Nauta: for Deep Learning on Kubernetes
Enterprises are still exploring use cases to augment their business models with Artificial intelligence (AI)… this is a market that is very much still-nascent.
Magical analyst house Gartner has conjoured up figures to suggest that real world AI deployments could reach nearly $4TN by 2022… and Deep Learning (DL) is key to the growth.
But, while DL in the enterprise is palpable, it is still a complex, risky and time-consuming proposition because it is tough to integrate, validate and optimise DL software solutions.
In an attempt to answer these challenges, we can look to Nauta as a new open source platform for distributed DL using Kubernetes.
What is Nauta?
Nauta (from Intel) provides a multi-user, distributed computing environment for running DL model training experiments on Intel Xeon Scalable processor-based systems.
Results can be viewed and monitored using a command line interface, web UI and/or TensorBoard.
Developers can use existing data sets, proprietary data, or downloaded data from online sources and create public or private folders to make collaboration among teams easier.
For scalability and management, Nauta uses components from the Kubernetes orchestration system, using Kubeflow and Docker for containerized machine learning at scale.
DL model templates are available (and customisable) on the platform — for model testing, Nauta also supports both batch and streaming inference.
Intel has said that it created Nauta with the workflow of developers and data scientists in mind.
“Nauta is an enterprise-grade stack for teams who need to run DL workloads to train models that will be deployed in production. With Nauta, users can define and schedule containerised deep learning experiments using Kubernetes on single or multiple worker nodes… and check the status and results of those experiments to further adjust and run additional experiments, or prepare the trained model for deployment,” said Intel, in a press statement.
The promise here is that Nauta gives users the ability to use shared best practices from seasoned machine learning developers and operators.
At every level of abstraction, developers still have the opportunity to fall back to Kubernetes and use primitives directly.
Essentially, Nauta gives newcomers to Kubernetes the ability to experiment – while maintaining guard rails.