Nvidia updates AI Enterprise software to offer low-code solutions

Version 2.1 focuses on increasing accessibility across multi-cloud environments

Nvidia has published an updated version of its AI Enterprise software, which includes a low-code toolkit and support for industry tools from Red Hat and Microsoft.

Nvidia AI Enterprise 2.1, now in general availability, can enable enterprises to “deploy and scale AI applications across bare metal, virtual, container and cloud environments,” an announcement reads.

The software update also aims to increase accessibility across hybrid or multi-cloud environments.

As such, Nvidia has opened support for its software to Red Hat’s OpenShift hybrid cloud platform and Microsoft’s cloud-based Azure NVads A10 v5 virtual machines. The added support will enable standardized AI workflows to scale and maximize the cost efficiency of deployments, according to Nvidia.

NVads and OpenShift are the first Nvidia virtual GPU instances offered from the public cloud, which product marketing manager Phoebe Lee said “enables affordable GPU sharing.”

Also included in the update is the Nvidia TAO Toolkit, a low-code version of the company’s namesake framework that’s designed to give developers the ability to create custom models for speech and vision-focused AI applications.

Now supported through AI Enterprise, the TAO Toolkit includes new pre-trained models that users can fine-tune using their own data without AI expertise or large training datasets.

The update also saw the release of the latest iteration of Nvidia RAPIDS, a collection of open-source software APIs for data science applications that run on GPUs.

Nvidia’s Enterprise upgrade follows the company partnering with Pure Storage to launch AIRI//S, a new solution designed to more easily deploy and manage AI infrastructure.

Systems from the likes of Asus, Dell Technologies, Gigabyte and Lenovo are all certified to run Nvidia’s AI Enterprise software. To become achieve certification, systems must undergo testing and adhere to the company’s design best practices for performance, security and scalability.

One company that is not certified, however, is Arm, the chipmaker Nvidia tried to buy before its deal dramatically fell apart earlier this year. In a nod to the merger during previous updates to its AI Enterprise software last June, Nvidia said it was going to extend its certification efforts to Arm CPU servers in 2022. Since the deal’s collapse, this has yet to occur, with Arm nowhere to be seen on the company’s list of certified brands.