Nvidia Opens Access to Physics ML Platform for ScientistsNvidia Opens Access to Physics ML Platform for Scientists
Modulus users can build digital twins of physical environments to improve analytics
March 31, 2023
At a Glance
- Nvidia has announced it's providing open source access to its Modulus platform.
- Users can customize pre-trained models and visualize them as physics-based digital twins.
Nvidia is open-sourcing its Modulus physics-ML platform, giving wider access to scientists to train data to build AI models.
Modulus provides access to pre-trained AI models from the Nvidia NGC catalog, which users can download and customize for specific use cases.
Modulus users can also build virtual representations of physical environments that are governed by partial differential equations, like planetary atmospheres. Users can then use the digital twins to virtually analyze everything from electromagnetics to climate science.
According to Nvidia, users have already used Modulus for weather forecasting, reducing power plant greenhouse gasses and accelerating clean energy transitions.
The tech giant said open sourcing Modulus would provide scientists and engineers with improved opportunities to collaborate, improve the transparency of physics-ML work and provide wider accessibility to other stakeholders like policymakers and national lab directors.
“An open-source workflow can help AI developers and engineering and science domain experts to work more collaboratively, transparently, innovatively, and with greater accessibility to enhance the impact and relevance of their research,” a company blog post reads.
Modulus is also integrated into Nvidia’s metaverse collaboration platform, Omniverse via an extension. The extension enables users to import results into a visualization pipeline for common output scenarios. It also provides an interface that enables interactive exploration of design variables and parameters to visualize results.
The PyTorch-based deep-learning toolkit for Modulus can be downloaded via GitHub.
About the Author(s)
You May Also Like