New NVIDIA TAO Toolkit Capabilities Ease AI Deployments

The News: An updated NVIDIA TAO Toolkit is now generally available from NVIDIA, incorporating a host of improvements and new components that are designed to make it easier for developers to build AI models for speech and vision AI applications. The toolkit is used with the NVIDIA Train, Adapt, and Optimize (TAO) framework, unveiled by NVIDIA in April of 2021. TAO allows developers to use transfer learning to create production-ready models customized and optimized for a wide range of use cases, including detecting defects, translating languages, or managing traffic, without requiring massive amounts of data. Read the full NVIDIA blog post.

New NVIDIA TAO Toolkit Capabilities Ease AI Deployments

Analyst Take: The nascent NVIDIA TAO Toolkit was a useful addition when NVIDIA unveiled it in April 2021, intended to make it simpler and faster for developers to create new production ready AI applications for speech and vision uses. These latest improvements in the latest version of the NVIDIA TAO Toolkit bring even deeper capabilities for AI developers, including:

  • The addition of REST APIs and the inclusion of pretrained models to speed up application customization and fine-tuning.
  • The ability to now import pretrained weights from ONNX to allow developers to prune and perform quantization on their own models for image classification and segmentation tasks.
  • The ability to understand model training performance by visualizing scalars such as training and validation loss, model weights, and predicted images using TensorFlow’s TensorBoard visualization toolkit.

For developers, these broad new features add great tools which help the updated NVIDIA TAO Toolkit shine.

Creating AI applications is not an easy task and any tools that can help developers simplify and streamline the complex processes that are involved are sure to help create new opportunities for broader AI innovation.

With the addition of REST APIs to the NVIDIA TAO Toolkit, developers can now build new AI services or update existing AI applications to allow the creation and delivery of scalable services using industry-standard APIs. This is huge for AI developers and for the companies that are building and using these applications.

Making the NVIDIA TAO Toolkit even more valuable and useful is that it is available with enterprise support with the NVIDIA AI Enterprise software suite, which is built for AI development and deployment.

I’m excited about the bold new features aimed at enterprise AI developers in the latest NVIDIA TAO Toolkit release and for the broad new capabilities it will give to developers to continue their innovations with this technology. It will be interesting to watch NVIDIA as the company continues to drive new innovations, products, and tools to bolster enterprise AI use in a wildly competitive marketplace.

Disclosure: Futurum Research is a research and advisory firm that engages or has engaged in research, analysis, and advisory services with many technology companies, including those mentioned in this article. The author does not hold any equity positions with any company mentioned in this article.

Analysis and opinions expressed herein are specific to the analyst individually and data and other information that might have been provided for validation, not those of Futurum Research as a whole.

Other insights from Futurum Research:

Computex: NVIDIA Grace CPU-Powered Servers Coming 1H 2023

NVIDIA Innovations, Enhancements in NVIDIA Omniverse, Digital Twins and Industrial Robotics Technologies Are Driving New Possibilities for Enterprises

NVIDIA Delivers Another Record Quarter

Image Credit: NVIDIA

Author Information

Todd joined The Futurum Group as an Analyst after over 20 years as a technology journalist covering such topic areas as artificial intelligence (AI), deep learning (DL), machine learning (ML), open source and Linux, high-performance computing, supercomputers, cloud computing, virtualization, containers and microservices, IT security and more.

Prior to his work with The Futurum Group, Todd previously served as managing editor of from 2020 through 2022 where he worked to drive coverage of AI use and innovation in the enterprise. He also served in the past as a staff writer for Computerworld and eWEEK and freelanced for a wide range of tech websites, including TechRepublic, Channel Futures and Channel Partners, Computerworld, PC World, Data Center Knowledge, IT Pro Today, and The Linux Foundation.

Todd holds a Bachelor’s Degree in Journalism from the University of Wisconsin-Madison. A Philadelphia native, he lives in Lancaster County, Pa., and spends his spare time tinkering with his vintage Mazda Miata convertible and collecting toy taxis from around the world.


Latest Insights:

On this episode of The Six Five – On The Road, hosts Daniel Newman and Patrick Moorhead welcome Microsoft’s Steven Bathiche, Technical Fellow, Windows & Devices, for a conversation on bringing AI to life for the Microsoft Event in NYC.
On this episode of The Six Five – On The Road, hosts Daniel Newman and Patrick Moorhead welcome Microsoft’s Matt Barlow for a conversation on Microsoft's latest Copilot announcements, Generative AI and Microsoft's vision for the future.
On this episode of The Six Five – On The Road, hosts Daniel Newman and Patrick Moorhead welcome Colette Stallbaume for a conversation on the GA release of Microsoft 365 Copilot, harnessing the power of Generative AI responsibly, and unlocking a new wave of productivity.
Oracle, In Partnership With Cohere, Has Launched a Generative AI Service on Its OCI Platform, Offering Users API-Based AI Integration and Dedicated Infrastructure
The Futurum Group’s Steven Dickens and Guy Currier discuss the latest announcements from Oracle, including limited availability of OCI Generative AI service, developed in partnership with Cohere.