Nvidia launches tools to accelerate training of humanoid robots

Nvidia announced its new efforts to develop humanoid robots during SIGGRAPH today (29). The company introduced a suite of tools aimed at accelerating industry projects, including virtual simulations from its Omniverse metaverse.

One of the new features is the Nvidia NIM update. A set of microservices and frameworks for simulating and learning robots and AI. Consists of ready-made containers to significantly shorten launch time from weeks to minutes.

New AI microservices could improve work on Nvidia’s Isaac Sim e platform, a simulation streaming tool connected to Nvidia Omniverse, a kind of advanced metadata platform where the company’s solutions replace (or supplement) real-life training with virtual experiments in the company’s virtual universe.

Inside Another development front is Nvidia OSMO. A cloud-native service for managing complex workflows in robotics development.

Simplifying model training

Nvidia’s solutions “dramatically simplify” robot training and simulation workflows, according to the brand itself.

“Developing humanoid robots is extremely complex and requires incredible amounts of real data that is tediously collected in the real world,” said Alex Gu, CEO of Fourier, a company focused on developing general-purpose robots.

“Nvidia’s new simulation and generative AI development tools will help accelerate the development of our models,” he added.

Although the new features do not directly benefit the general public, products DNvidia catalyzes the development of vending machines. It will take some time for the fruits of this investment to come to market, but it is promising for companies focusing on the segment.

According to Nvidia, this news is available to developers enrolled in the Nvidia Humanoid Robot Developer Program. Nvidia OSMO and Isaac Lab will be released first, followed by Nvidia NIM news.